White Paper One
Talent Buying & Event Management Since 1975
Fenced In
Putting Location Data to Work
Which platforms will take the audience you have identified, which will not, and what to run once you know the difference.
The second of three white papers. For fair, festival, and event organizers.
Prepared By
TSE Entertainment, LLC
Austin, Texas 2026 Edition
About This Series
This is the second of three white papers on location data for live events.
The first covers what location data exists, what each kind can tell you, and where to buy it. This white paper covers how to activate it: the advertising platforms, the campaigns that work, and how to budget them. The third covers what happens inside your gate, including event apps, proximity technology, and sponsor activation.
Each stands on its own. The first is the better starting point if you have not bought anything yet, because most of the money wasted in this field is spent before anyone reads a report.
Executive Summary
White Paper One ended on a boundary. An organizer who has read a visitation report, added a segmentation layer, and learned who actually walks through the gate arrives at an obvious next question: now what do I do with it?
The answer is less straightforward than the vendor deck suggests, because the advertising market is split down the middle, and the split is not where most organizers assume it is.
What this paper argues
- White Paper One’s conclusion stands, and this paper is downstream of it. Location data is a stronger intelligence tool than an advertising tool. Knowing who your audience is and where they come from remains the higher-value use, and nothing in this paper displaces it. Reaching them is the cheaper problem. This is the paper about the cheaper problem, and it is cheaper precisely because the first one was done first.
- There are three routes to market, not one, and they accept different things. Neither Facebook nor Google offers an ordinary way to upload a vendor-built audience of devices seen at a place, and the two differ on everything else. TikTok, the third closed platform, follows the same pattern for an organizer’s own data. Google’s customer list tool is explicitly limited to data you collected yourself. Meta permits audiences shared through its approved tools, subject to its rights, permissions, and partner requirements. The independent buying platforms will take both kinds of audience. The profile route is a chain rather than a single company: the visitation platform identifies the profile, and a segmentation or onboarding provider turns it into an audience and delivers it into Meta, into the independent platforms, and into a mail file. On Google, the mechanism varies by provider and has to be asked about: an organizer cannot upload a third-party segment file as its own customer list, and “exports to Google” can mean several materially different things. Everything else in this paper follows from that, and knowing it in advance saves a season of arguing with a sales representative who cannot do what he is describing.
- White Paper One’s preference for profiles over device lists survives contact with the delivery platforms. That argument was made in the first paper and is not re-argued here. What this paper adds is the mechanism: segment audiences can travel into Meta through approved partner tools where geoframe audiences cannot, do not decay on the twelve month device clock, and are not what the state legislatures have been writing bills about. For an organizer who will fund exactly one purchased audience, that is the one.
- The cheapest reach you own is the reach you already have. Your ticket buyer list, matched into a platform, is the one audience that has already done what you want, costs nothing to acquire, and does not decay the way a device audience does, provided it is refreshed each season. Most organizers activate it last, if at all.
- Four campaigns do most of the work. Reaching your own past crowd, reaching the crowd of a comparable event, reaching the audience a shared headliner brings, and developing the nearby markets where your own ticket data shows you are underselling. Everything else is variation. For most standalone events, the device-based versions of the second and third are out of reach for the reason White Paper One gave: the audience is too small for a vendor to accept. The profile versions are generally much more broadly available. The fourth runs on your own ticket addresses and free Census tools, and it turns geofencing from capturing an audience into developing a market.
- Streaming television and audio are the quiet beneficiaries of the restrictions. Because the same audience that cannot enter Facebook can enter a demand side platform, and that platform can put it on a living room screen, streaming inventory has become the natural home for location-derived targeting rather than a premium extra.
- Text messaging can be among the highest-response, and highest-risk, channels in this paper. Consent rules tightened again in 2025 and 2026, Texas added a private right of action, and the damages are per message. It belongs in the plan. It does not belong in the hands of whoever has the ticketing password.
- Measurement is where these programs are oversold. Location data measures days and places, not performers, and an advertising platform reporting a lift figure it calculated on its own behalf is not evidence. What can be proven honestly is worth more than what can be claimed loosely, and Section 12 draws that line.
What this paper is not
It is not a media plan. Rates, minimums, and state availability move faster than a printed page, and any figure quoted here would be wrong by spring. What it gives instead is the shape of the market, the questions that separate a real capability from a described one, and a sequence for spending that starts with what you already own.
1. What White Paper One Established
Eight conclusions from the first paper carry directly into this one. An organizer who read White Paper One can move to Section 2. An organizer who has not can start here without being lost.
Location data is a stronger intelligence tool than an advertising tool. This is the first paper’s central finding and it governs this one. Its most reliable value is telling you who your audience is and where they come from. Reaching them is the cheaper problem, and it becomes cheaper still once you know who they are. White Paper One deferred the advertising platforms, the campaigns, and the budgets to White Paper Two, and this is that paper. It is not a revision of the hierarchy. It is the second half of it, and an organizer who arrives here without having done the first half will pay more for worse results.
There are five distinct kinds of location data, not one. They differ in how far back they reach, how much of an audience they capture, how quickly they decay, and how exposed they are legally. Aggregated visitation data and household segmentation describe people. Device identifiers and household linkage reach people. Carrier data, for marketing purposes, is effectively off the table.
Audience profiles outlast device lists. Most event-focused device products reach back about twelve months and are being restricted state by state. A segment-based profile does not decay on that clock and can compound year over year, although the segment systems themselves are revised periodically and audiences built on them must be refreshed when they are. For an annual event, that difference decides the strategy, and it decides this paper too: the campaigns in Part Two are built to run off a profile wherever a profile will do the job.
For most standalone events, device-based targeting is unavailable rather than expensive. Vendors will not run a device campaign below a floor, and one vendor serving the event market publishes a floor of fifty thousand unique devices, counted in devices rather than attendance. A standalone county fair, a community festival, or a single-day event will usually be declined. Large single events, portfolios bundled inside one lookback window, or nobody. Household linkage is built from device identifiers, so it falls with the same floor. Every device-based recommendation in this paper is conditional on clearing it.
Segmentation is an add-on inside the platform you already bought, not a second subscription. The leading visitation platforms run a marketplace, and the household classification systems are licensed through it. One vendor relationship, one login, one invoice with line items. The exception is activation, which is a separate purchase from the segmentation provider, and Section 4 is where that exception stops being a footnote and becomes the most durable route in the paper.
The legal ground shifted again in 2026. White Paper One reported three states that prohibit the sale of precise geolocation data: Maryland, whose law took effect October 1, 2025, Oregon on January 1, 2026, and Virginia on July 1, 2026. Connecticut has since become the fourth. Its prohibition, enacted as Public Act 26-64, took effect October 1, 2026. Texas, often described in this market as an open state, is not one: precise geolocation is sensitive data under Texas law, processing it requires consent, and universal opt-out signals have had to be honored since January 2025. Any organizer working across state lines needs an execution map maintained by the vendor, reviewed before each campaign.
Some of what you will be told is state law is vendor policy. A geoframing provider that applies a three thousand foot minimum radius in several states is describing its own risk posture, not four identical statutes. The distinction matters when you shop, because a vendor rule may not apply to the next company you call.
Location data measures days and places, not performers. On a bill with six acts on one stage, the day’s draw belongs to the whole bill. Attribution becomes possible only where an act can be isolated: a single headliner night, the same act across two years, a day whose lineup is otherwise unchanged. This limit does not soften in White Paper Two. It gets stricter, because advertising adds a second layer of things that want to take credit.
2. The Platform Map
White Paper One stopped at the door of this section deliberately. It established which of the four capabilities in its own Figure 1 each platform can perform, then handed campaign construction, creative, sequencing, and budget to this paper. What follows is that handoff, and it starts with the thing organizers get wrong before any of it.
Every conversation about activating location data goes wrong in the same place. The organizer describes an audience. The vendor says yes. Nobody establishes whether the platform where the money will actually be spent can accept that audience, and the answer turns out to be no somewhere around the second invoice.
The market divides into three routes that behave completely differently.
The closed platforms. Facebook, Instagram, Google, and TikTok, along with their video and connected television inventory. These are the platforms your board has heard of and the ones your marketing coordinator already knows. They accept first-party customer data and, in Meta’s case, certain audiences shared through approved partner tools, under their own policies and consent requirements. They do not accept an audience a location vendor assembled from past presence at a place. There is no tier, no budget level, and no representative who unlocks it. The capability does not exist. Figure 1 charts Meta and Google; TikTok follows the same pattern for an organizer’s own data and is covered in Section 3.
The open platforms. The independent demand side platforms, the geoframing vendors that resell them, and the managed service agencies that sit on top. These accept both. A device audience built from your grounds last August, a household list matched to physical addresses, a segment audience pushed out of a segmentation provider: all of it enters here, and from here it can be routed to display, to streaming television, and to streaming audio.
The profile platforms. The third route is the one most organizers do not know exists, and it is the one White Paper One’s argument points at. It is a chain rather than a single product. The visitation platform identifies the profile: which household segments over-index at your event. A segmentation provider or an onboarder then turns those segments into a targetable audience and sends it onward, into Meta and the open platforms. An ordinary visitation subscription does not do the second step by itself unless the specific product and contract say so, and Section 4 explains how to tell.
That is the map. Three routes, not two, and the third behaves differently enough from the other two to deserve its own section.
Why the third route matters more than its share of the market suggests
A segment audience is not a set of devices seen at a place. It is a description of households that resemble the households your report identified, assembled from aggregated and modeled sources. That distinction is technical, and it has three practical consequences.
It travels further, though not everywhere. A segment audience can enter Meta, which a vendor-built geoframe audience cannot, with the advertiser representing that it holds the rights to use it. Google is stricter. An organizer cannot upload a third-party segment audience as its own Customer Match list; Customer Match is based on first-party customer data. Some segmentation and onboarding providers advertise supported Google integrations, but the mechanism matters: it may involve an appended version of the advertiser’s own customer list, a provider-shared audience available through an approved integration, or a third-party audience purchased through Display & Video 360. Do not assume that “exports to Google” means an ordinary Google Ads customer-list upload. Ask the provider which Google product, audience type, and integration it actually uses. The closed side of the map is therefore two walls, not one, and they are built in different places.
It does not decay on the same clock. Device audiences reach back about twelve months and thin from the moment they are built. A segment profile does not decay because people replace phones, and it can compound year over year, which for an annual event is the difference between renting an audience every season and owning a description of one. The qualification is that the segment systems are commercial products that get rebuilt. Experian has rebuilt Mosaic, and the previous version’s activation audiences retire in February 2027, which falls inside the 2027 season. A profile persists; the segment names and destination mappings under it have to be refreshed when the provider revises them.
It is the less exposed of the two legally. The state prohibitions in White Paper One are aimed at precise geolocation, meaning data that places an individual inside a small radius. Segment-based profiling built on aggregated sources is generally not the direct target of those statutes. That is not legal advice and not a guarantee, and Section 4 says where the caution still applies. But an organizer building a program that has to survive several more legislative sessions should notice which of the two routes the legislatures keep writing bills about.
The three consequences worth reading twice
One. Your own data travels every route. A ticket buyer list works in both closed platforms, works in the open ones, and is the seed a profile platform reads to tell you which segments you actually have. Only two rows in Figure 1 run clean across every column, and this is the one that costs nothing to acquire. Most organizers have never exported it. Section 8 is built entirely on it.
Two. Restriction pushes budget toward the open and profile routes, and that is where the screens are anyway. An organizer who cannot put a geoframe audience into Facebook has not lost the audience. He has lost one destination for it. The audience still runs on streaming television, streaming audio, and display through a demand side platform, and those environments suit an event advertisement better than a feed does anyway. Section 6 makes that case in full.
Three. The wall is doing part of your compliance work. A closed platform that refuses third-party location audiences is refusing the category most affected by the state prohibitions in White Paper One. That is not a reason to relax. It is a reason to understand that the compliance question lives almost entirely on the open side of the map, where the vendor question checklist at the back of this paper earns its keep.
What the map does not tell you
It does not tell you which door to use. That depends on what you are trying to accomplish, and the honest answer for most organizers is both, in a specific order, for different jobs. The closed platforms are where your own audience is cheapest to reach at scale. The open platforms are where the audiences you cannot build yourself become reachable at all. Part Three sequences the spending. Part One establishes what each door will actually do once you knock on it.
3. Facebook and Google, and Tiktok
These two get their own section for the same reason White Paper One gave them one: they are the first place every organizer looks, and the conversation about them is usually conducted at cross purposes. A vendor says the platforms cannot do location targeting. A marketing coordinator says of course they can, she does it every week. Both are describing something real, and they are describing different things. TikTok, the third closed platform, is covered at the end of the section. It follows the same rules, with a job of its own.
What they will not do
Neither platform will build an audience from past presence at a place. You cannot ask either one who was on your grounds last August, and neither provides an ordinary customer list workflow for uploading a vendor’s historical geofence audience. This is the geoframe row of Figure 1, and it stops at this column on both platforms. The two are not identical beyond that point. Google’s Customer Match is explicitly first-party. Meta permits audiences shared by service providers and data partners through its approved tools, subject to its rights, permissions, lawful basis, and partner requirements, which is why Section 4 treats segment audiences reaching Meta as routine and segment audiences reaching Google as a question to put to the provider.
Three further limits are worth knowing before a sales representative characterizes them for you.
- Radius floors that exceed most venues. Both platforms set a minimum on how tight a boundary you can draw. The figures differ between the two, differ by whether you are anchoring to a city or dropping a pin, and are not always the same in the buying interface as in the platform’s own developer documentation. Confirm the current number in the platform’s documentation rather than from a vendor deck or a blog post, because the widely repeated figures do not match the published ones.
- No separation of visitors from residents. A boundary drawn around a venue reaches the people who live nearby alongside anyone who visited, and there is no setting that tells them apart. For a fairgrounds surrounded by housing, that distinction is the entire question.
- Delivery that can reach past the boundary. Meta generally treats location as an audience control, but some campaign setups offer a “reach more people likely to respond” option that can broaden geographic delivery. If geographic containment matters, verify the current location and expansion settings in every ad set rather than once at account level.
So if the question is whether Facebook can fence your grounds and reach the people who attended, the answer is still no, and no amount of budget changes it.
What they will do, and why it matters more
Both platforms accept a customer list you collected yourself, matched against their user base, subject to their policies and your own consent obligations. For an event organizer this is the single most valuable capability either platform offers, and it is routinely left unused while the organizer argues about geofencing.
One condition on the Google side decides more fair campaigns than any other, and it is almost never mentioned. Every policy-compliant Google Ads account can use a customer list for exclusion and for observation. Using it to target requires at least ninety days of account history and more than fifty thousand dollars in lifetime spend on the account. Many fair offices will not clear that threshold on their own account. Those that do not can still suppress their buyers on Google, and can still target them on Meta, where no equivalent spend threshold applies; an agency account that does clear it is the other common route. Confirm the current terms in Google’s own policy pages before planning around either.
Three things follow from it.
- Your ticket buyers become reachable directly. A list of people who bought a ticket in the last three seasons, uploaded and matched, is an audience of proven attenders. No location vendor can sell you a better one, because no location vendor has it. You do.
- Similar-audience modeling turns a small list into a large one. Meta builds lookalike audiences from a seed list you provide. Google retired Similar Audiences in 2023. In Google Ads, Lookalike segments are now available in Demand Gen and Video brand campaigns. Demand Gen spans YouTube, Discover, Gmail, Maps, and the Google Display Network. During 2026, Google began transitioning Demand Gen Lookalikes toward a suggestion or signal model rather than a hard audience boundary. Google has also been expanding lookalike functionality in Display & Video 360, which in 2026 extended lookalike segments to YouTube line items bought there. Seed either with your best customers rather than all of them, and it becomes the cheapest prospecting audience available to an event. This is also the most dependable way to make White Paper One’s segmentation work pay off on Google, where you cannot upload a third-party segment as your own list: seed the model with the households on your own list that match the segment you learned matters.
- Suppression is as valuable as targeting. The same matched list that lets you reach past buyers lets you exclude them, which is how you stop spending prospecting budget on people who already have wristbands. Most organizers never build the exclusion.
The conditions attached
The list has to be genuinely yours, collected with notice and the appropriate consent, and cleaned of anyone who has opted out before it is uploaded rather than after. That last point has teeth now. Texas has required recognition of universal opt-out signals since January 2025, and an organizer who ignores a documented opt-out and then uploads the record has created a problem that no platform setting fixes.
Match rates vary widely and no platform guarantees one. A list of ticket buyers with personal email addresses will match better than a list of corporate contacts, for the ordinary reason that people sign up for consumer services with personal addresses. Expect attrition, plan the budget on matched size rather than list size, and ask the platform what it matched rather than assuming. Lists also expire inside the platforms. Google drops any list membership not added or refreshed within 540 days, so a list uploaded once and left alone quietly empties. Refresh every list each season.
One difference between the two worth noting
Google’s strength for an event is intent. Somebody typing the name of your fair, or the name of the headliner plus a city, is telling you something a location audience never can, which is that they are thinking about attending right now. Search is not location targeting and does not pretend to be, and it belongs in the plan for exactly that reason.
Meta’s strength is the opposite. Nobody opens Instagram looking for your event. What Meta offers is a matched audience and a modeled extension of it, delivered to people who were not looking, which is what an event needs in the months before anyone is searching.
An organizer who understands that division stops asking which platform is better and starts asking which one is doing which job.
TikTok, the third closed platform
TikTok belongs in the closed group, and for an organizer’s purposes it behaves much like Meta. It accepts a customer file you collected yourself, matched against its users by email, phone number, or mobile advertising ID. It builds lookalike audiences from that file, lets you exclude the same list from prospecting, and retargets visitors to your website through its own tracking pixel. Its location settings run from states and metro areas down to five-digit ZIP codes, with up to 3,000 locations in a single ad group, so a long list of target ZIP codes from a penetration analysis fits in one campaign.
Three limits apply. It does not support excluding locations or ZIP codes, so an organizer cannot carve its strongest areas out of a broader target and has to list the target areas instead. Its finest geography is the five-digit ZIP code, so block-group precision is not available there, any more than on Meta or Google. And like the other two, it reaches people it judges to be located in or regularly in an area, with no setting that separates residents from visitors.
Because its customer file accepts mobile advertising IDs, whether a vendor’s list of devices seen at a place could be uploaded is a question of rights and policy rather than mechanics. The file is meant for identifiers you hold the rights to, and a purchased device list is the category the state sale bans reach. Treat it as unavailable unless TikTok confirms otherwise in writing.
Its strength is discovery. TikTok has become one of the main places younger listeners find music, country included, and nobody opens it looking for your event any more than they open Instagram that way. For the younger concert audience, a lookalike seeded with concert-only buyers and pointed at target ZIP codes is a natural campaign there. It is a less obvious first choice for the family audience, which Meta reaches well.
One business note. TikTok’s U.S. operation moved into a majority-American joint venture in January 2026, with ByteDance retaining a 19.9 percent minority stake, which removed the near-term risk of a shutdown. Its recommendation system is being retrained on U.S. data, so performance may shift for a time. Run it as one channel among several rather than the only one.
What to hold on to from this section
The closed platforms are not a location product and will not become one. They are the cheapest way to reach an audience you already own and the cheapest way to model an audience that resembles it. Treat them that way, spend the geofencing argument somewhere else, and the two hours a week your coordinator already spends there starts producing considerably more.
4. Aggregated Audience Profile Platforms
White Paper One treated the visitation platform and its segmentation layer as instruments of analysis. You draw a boundary, you read who visited, you learn that your Saturday crowd and your Sunday crowd are not the same people. That was the whole job.
It is not the whole job any more. The household classification systems licensed inside these platforms will now take the segments you identified and turn them into an audience you can advertise to. The visitation platform identifies the profile; the segmentation or onboarding provider activates it. The report can become a campaign, but through a second company and usually a second purchase.
Who is in this category
The companies named below are examples chosen to illustrate the category. They are not recommendations, TSE has not evaluated their service, and inclusion or omission should not be read as a judgment. This is the same list White Paper One introduced, read from the activation end rather than the analysis end.
The visitation platforms are where the profile is read. Placer.ai is the category leader for trade area and visitation analysis and the one most event organizers meet first. PassBy offers tiered plans aimed at teams that do not need an enterprise deployment. Unacast, Foursquare, Azira, and Advan offer products ranging from data feeds and APIs to analytical platforms, and most of their offerings suit an organization with an analyst better than a fair office. Azira, for example, launched a conversational analysis platform in September 2026, initially in beta for destination marketers. What matters for this section is that these platforms are built to tell you who visited. They deliver that answer through a dashboard, a data feed, or an API. Reading a report is not the same as pushing an audience, and an organizer who assumes his visitation subscription includes activation will find out at the wrong moment.
The household classification systems are where the audience is built. Spatial.ai’s PersonaLive sorts households into eighty segments across seventeen families, and its own platform includes an audience builder that exports the chosen segments to named advertising destinations including the major social platforms, programmatic buying platforms, and direct mail. Experian Mosaic, the long-established classification system, reaches activation differently: its audiences are carried as third-party segments inside demand side platforms, so the buying happens where the media is bought rather than inside the segmentation tool. Environics Analytics and its PRIZM system are the Canadian counterpart and the place to start for organizers working north of the border. AGS, STI PopStats, and Niche add expenditure and population layers that describe rather than activate.
The identity resolution providers are the connective tissue. LiveRamp is the dominant one. It resolves an audience to a durable household identifier and distributes it across a large network of media and advertising platforms, which is how a segment audience reaches connected television and streaming audio inventory it could not reach directly. Most organizers will never contract with an onboarder themselves. They will encounter it as the mechanism inside somebody else’s quote, and it is worth recognizing the word when it appears.
What the mechanism actually is
A household classification system sorts every household in the country into a limited number of behavioral types, built from a mix of aggregated, modeled, and behavioral inputs. What comes out, and what you target, is a segment rather than a list of identified people who attended your event. Your visitation report tells you which of those types over-index at your event. The activation product then assembles the households nationally, or inside a geography you specify, that belong to those types, and delivers that audience to an advertising platform you name.
The important word is delivers. You are not buying a list of people who were at your fairgrounds. You are buying reach against a description of the kind of household that comes to your fairgrounds, which is a different product with different properties, and White Paper One argued at length that it is usually the more useful one.
The division of labor, stated plainly
White Paper One made the point that segmentation is an add-on inside your visitation platform rather than a second subscription, and that the exception is activation. This section is that exception in full.
Analysis and activation are usually two different companies doing two different jobs, even when the segments carry the same name in both places. You read Mosaic or PersonaLive segments inside Placer.ai because they are licensed there through the marketplace. You activate those segments through the segmentation provider’s own product, or through a demand side platform that carries them, or through an onboarder. The segment name travels. The subscription does not.
The practical consequence is a question, and it is the question to ask before renewing anything: does my current subscription let me advertise to these segments, or only read them? A great many organizers have been paying for the reading half for two seasons under the impression that the advertising half came with it.
Where it goes
This is the part that surprises organizers. A segment audience of this kind can be delivered into Meta as well as into the open platforms, because it arrives as a household audience rather than as a set of devices observed at a location. It enters Meta through the provider’s own approved integration rather than as a file the organizer uploads, and Meta’s rights and partner requirements still apply, so the question to put to the segmentation provider is what it warrants about that. Where a provider lists Google among its destinations, ask exactly what that means. An organizer cannot independently upload a third-party segment file as Customer Match. Depending on the provider, Google activation may instead involve an appended first-party customer list, a provider-shared audience through an approved integration, or a third-party segment in Display & Video 360. Those are materially different products with different eligibility and inventory. Google is different, for the reason given in Section 2. In practice the same segment audience can run on social, through a demand side platform on display and streaming inventory, and into a direct mail file, from one build. Whether and how it runs on Google depends on the provider.
Where the audience is delivered through an identity resolution provider rather than platform to platform, the reach extends further still, into connected television and streaming audio inventory that would otherwise require a separate buy. Section 6 picks that up.
Three questions that separate the real capability from the described one
Vendors in this category describe activation in similar language regardless of what they actually operate. These three questions sort them quickly.
- Is activation included, or is it a separate product and a separate purchase? White Paper One established that reading segments inside your visitation dashboard is a marketplace add-on, one login and one invoice. Pushing those segments out to an advertising platform is generally a capability of the segmentation provider’s own product, and that is a different purchase at a different price. Establish which one you are being quoted before the renewal, not after.
- Which destinations are live today, and which are on a roadmap? Ask for the current list in writing, by platform name. Destination lists in this category change quarterly, and a slide showing eleven logos is not a statement that all eleven are available to your account at your spend level.
- What is the minimum audience size, and what happens to a rural county below it? This is where profile activation most often fails an event organizer quietly. A national segment audience is enormous. The same segment inside a three-county trade area may be too small to deliver against, and the system’s response to that is usually to widen the geography rather than to tell you it could not fill the order.
Where the caution still applies
Two cautions, and they are not the ones organizers expect.
The first is that aggregated does not mean unregulated. These products are built from data with its own upstream consent chain, they are subject to the same universal opt-out signals as everything else, and a state that has not written a location statute may still reach the practice through general consumer protection law. The relevant question for a vendor is not whether the product is legal. It is what the vendor does when a household opts out, and how quickly that propagates to an audience already delivered.
The second is that a modeled audience is a modeled audience. It describes households that resemble your attenders. It does not know that any particular household has ever heard of your event. That is precisely why it works for prospecting and precisely why it should never be measured as though it were a list of past customers. Section 12 is unambiguous about this.
What this route is for
Prospecting, and specifically prospecting that has to keep working for several years. It reaches households that look like the ones you already draw, in geographies you have not saturated, without depending on a data category that four states have now restricted and more are considering.
White Paper One argued that audience profiles outlast device lists and called that the strategic argument of the paper. This section is what that argument looks like when it is executed rather than asserted. The profile you already bought to answer a question turns out to be the most durable thing you can also advertise against, which is a convenient result rather than a coincidence: it is durable for the same reason it was informative, because it describes a kind of household rather than tracking a particular one.
5. Independent Buying Platforms
This is the open route, and it is where the audiences the closed platforms refuse become reachable.
The category has three layers, and organizers routinely mistake one for another.
The demand side platforms are the buying engines. The Trade Desk, Google’s DV360, StackAdapt, Simpli.fi, and Basis are among the names an organizer will hear. They hold the connections to the inventory, they run the auctions, and they accept audiences from many sources including the geoframing vendors and the identity providers above. Some sell directly to advertisers with a self-serve interface. Others sell only through partners, which is why a fair office usually meets one of these platforms wearing somebody else’s badge. Google appearing here is not a contradiction of Section 3. Display & Video 360 and Google Ads are separate buying environments with different audience rules: Display & Video 360 supports third-party audience segments where Google Ads’ customer list tool does not.
The geoframing vendors build the location audience and buy media with it. They are the ones who will draw a boundary around your grounds, or around a comparable event, and produce a set of devices seen inside it. White Paper One named this category and made the central point about it: most of them resell the same underlying suppliers and buying platforms, so the technology is broadly common and what differs is service, reporting, and candor. El Toro built its approach on coordinate-level capture and address matching. Choozle is self-serve and publishes which of its products are unavailable in which states. GroundTruth built its business on foot-traffic verified targeting. Feathr is purpose-built for events, associations, and nonprofits, and publishes its own lookback windows, audience minimums, and state restrictions. Propellant Media, Thumbvista, and similar agencies resell these capabilities as a managed service. As in Section 4, these are illustrations of a category rather than recommendations.
The managed service agencies sit on top of both, take a brief, and run the campaign. For a fair office with no dedicated digital staff this is frequently the right answer, and the mark-up should be weighed against the cost of the mistakes it prevents rather than against the media alone.
What you get here that you cannot get anywhere else
Three things, and it is worth being precise about them because the rest of the category’s pitch is available elsewhere cheaper.
Historical presence audiences. A set of devices observed inside a boundary during a past window, which is the capability the closed platforms do not have at any price. This is the engine of the device versions of the plays in Sections 9 and 10, and it is subject to the audience floor from White Paper One: the vendor must be able to find enough devices inside the boundary to accept the order.
Address-level household targeting. A list of physical addresses matched to households and reachable across devices in the home. This is what lets direct mail and digital run against the same households. How exposed it is depends entirely on where the addresses came from. Addresses from your own ticketing records, or selected because they fall in a segment, start from a file rather than from observed movement. Addresses produced by linking devices seen at a place back to homes are household linkage in White Paper One’s sense: built from device identifiers, bound by the same floor, and restricted by the same statutes. Ask which one you are being quoted.
Conversion zone measurement. The ability to draw a second boundary, usually your gate, and report how many exposed devices later appeared inside it. Section 12 explains carefully what this does and does not prove, because it is the single most oversold number in this business.
How to buy from this category without being taken
White Paper One’s rule holds and gets sharper here: published limitations are the strongest available signal of quality. A vendor who states in writing which products cannot run in which states, what the minimum audience size is, and how far back the lookback window reaches is expecting to be held to it.
Four additional questions belong in every conversation in this category.
- Whose data is this, and what happens if that supplier changes policy? Most vendors in this layer are reselling. A vendor who cannot name the upstream supplier is asking you to accept a supply chain you cannot inspect.
- What is the state execution map today, and who maintains it? Not what is legal in general. What your specific campaign can run, in the specific states your audience lives in, this month.
- What share of the budget is media and what share is fee? A managed service is a legitimate purchase. A managed service billed as media is not.
- What do I keep when the campaign ends? Audience files, reports, and the underlying counts. The answer is often less than the organizer assumed.
The cost structure nobody explains
Device audiences are priced against a scarcity that gets worse every year. The share of phones sharing location with a participating app keeps shrinking, the lookback windows keep shortening, and the states keep narrowing where the product runs at all. That means the same audience costs more to assemble each season and covers less of your crowd than the year before.
This is the arithmetic behind White Paper One’s ordering, and it is why this section sits after Section 4 rather than before it. The open route buys you something real that nothing else can buy. It is also the only route in this paper whose supply is contracting.
6. Streaming Television, Streaming Audio, and Direct Mail
These three sit together because they share a mechanism. None of them is a location product. All three are destinations that a location-derived or profile-derived audience can be delivered into, and all three are undersold to event organizers who assume they are priced for national brands.
Streaming television
Connected television inventory is bought programmatically through the same demand side platforms as display, which means an audience assembled by any of the three routes in Section 2 can be pointed at a living room screen. For an event this matters more than the equivalent display impression for reasons that have nothing to do with targeting. The advertisement runs full screen, it is often bought in non-skippable formats, and it arrives in the room where household plans get made, which is the actual decision unit for a family buying four fair tickets.
Two practical notes. Creative is the constraint, not inventory: a fifteen second spot cut from social vertical video looks like exactly that on a television, and the difference in response is visible. And frequency management across a small trade area needs an explicit cap, because a tight geography with a modest budget will otherwise show the same household the same spot enough times to become an irritant your box office hears about.
Streaming audio
Audio deserves more attention from fairs than it gets, for one structural reason: it is particularly well suited to reaching people while they are driving, which for a regional event is the same activity as deciding how far they are willing to travel. Audience delivery works the same way as streaming television, through the platforms, and inventory is generally cheaper.
The creative rule is different. Audio has no visual, so the event name, the date, and one reason to come have to survive being heard once while someone is merging onto a highway. Most event audio spots fail because they were written to be read.
Direct mail
Mail is in this section rather than in a chapter of its own because the modern version of it is an audience product. An address-level household audience, of the kind Section 5 describes or the kind a segmentation provider in Section 4 produces, can be sent to a mailbox and to the screens inside the same household. The mail file generally comes from the address, identity, or segmentation provider that built the audience, not from the demand side platform, which buys digital media only; the same onboarded audience feeds both. The pairing is the point: the same household sees the same offer on paper and on a screen in the same week. Whether that pairing lifts response for your event is a testable question, and the holdout in Section 12 is how to test it.
For fairs specifically, mail has an advantage the digital channels do not: it can stay on a refrigerator door until the event. A dated event with a fixed window is close to the ideal use case for a medium whose weakness is that it arrives whenever it arrives.
The cost discipline is straightforward. Mail is expensive per household and cheap per household that actually attends, and which of those two figures you are looking at depends entirely on how well the audience was built. Mailing a rented list is the version of this that gives mail its bad reputation. Mailing the households your own profile identified is a different exercise with different economics.
7. First-Party Email and Text Messaging
The previous sections dealt primarily with paid-media platforms. This section turns to first-party communication channels you control directly, and it is the most valuable and most legally dangerous section in the paper.
Email is typically among the lowest-cost forms of repeat reach an event organizer has. It is also the asset most likely to be sitting unused inside a ticketing platform because nobody asked for the export.
Three practices separate an email program that works from one that gets ignored.
- Segment by behavior, not by demographic. A household that bought on the day of the announcement and a household that bought at the gate are different buyers, and the announcement of next year’s headliner is worth sending to them at different times with different framing.
- Send from the event, to a schedule the audience can predict. An annual event that mails twice in March and vanishes trains its list to ignore it.
- Treat the list as a permanent asset with an annual maintenance cost. Deliverability decays. Addresses go stale. A list that has not been cleaned in three seasons will quietly stop reaching a meaningful share of the people on it, and nothing in the send report will say so.
Text messaging, and why it needs a lawyer's attention rather than an intern's
Text can be among an event’s highest-response channels, and it carries the most legal exposure of anything in this paper. The rules tightened twice recently and Texas tightened them further than the federal floor.
The federal position, in plain terms. For marketing texts covered by the Telephone Consumer Protection Act, prior express written consent is generally required, and state law and carrier rules can add to it. Statutory damages run from five hundred to fifteen hundred dollars per message, with no cap in the aggregate, which is why a single non-compliant send to a ticket list is an existential rather than an administrative problem. Since April 2025 a consumer may revoke consent by any reasonable method, not only by replying STOP, and the revocation has to be processed within ten business days. The broader requirement, that a revocation in one channel applies to all future marketing contact from the same business, has been delayed to January 31, 2027, which is inside the planning horizon of anyone reading this for a 2027 season.
Separately from the law, the carriers enforce their own layer. Since September 1, 2023, unregistered application-to-person traffic sent to U.S. numbers over ordinary ten digit numbers has been blocked rather than throttled, so a program has to be registered as a brand and a campaign before a single message will arrive. Registration is not consent and consent is not registration. You need both, and a program can be perfectly lawful and still deliver nothing.
Two content rules that catch fairs specifically. Quiet hours apply. And alcohol-related messaging is age-restricted under carrier rules and may require specific verification and age gating, so a text promoting the beer garden, the wine walk, or a distillery sponsor should be reviewed before launch, even where nothing in the law prohibits it. Plan the sponsor obligations around that before you sell them.
Texas, in detail, because it is where a great many of these events are
Texas Senate Bill 140 took effect on September 1, 2025 and expanded the state’s telemarketing statute to cover text messages explicitly. It created a private right of action through the Deceptive Trade Practices Act, with statutory damages that stack, treble exposure for willful conduct, and no limit on the number of successive claims a single individual may bring. Permitted hours are narrower than the federal standard, and Sundays start later.
The registration question caused a season of confusion and has since been clarified. A constitutional challenge filed in September 2025 was resolved that November, and the state’s position in those filings is that businesses running consent-based text marketing programs are outside the registration requirement. The exposure that remains, and it is substantial, is private litigation over consent itself.
The practical instruction for a Texas fair is short, and it is recommended practice rather than statutory deadline. Keep the consent record, including the exact language shown at the point of opt-in and the date. Keep it for at least four years past the last contact, which covers the period in which federal claims can typically be brought. Make an opt-out propagate to every system as quickly as you can, and never later than the federal processing deadline above. And do not import a list from a sponsor, a partner venue, or a previous promoter, because the consent that list carries was not given to you.
The one sentence to take away
These are first-party relationships you control rather than audiences you rent every season, although delivery still depends on carriers, mailbox providers, platform policies, and the law. They are also the channels where the organizer, rather than a vendor, carries the legal exposure directly. Build them first and govern them properly.
Four campaigns do most of the work an event can ask of location data. The first three are presented in order of certainty, which is also the order in which they should be attempted. The first is available to any event with usable first-party records. The device versions of the second and third are available only to events and portfolios that clear the audience floor; the profile versions are available to far more events. The fourth turns from demand you can already see to demand you have not yet developed. Its analysis is available to any event with ticket addresses, and its activation depends on the platform and the state.
8. Reaching Your Own Past Crowd
The first play uses no purchased data at all.
What it is. Take the audience you already hold, meaning registered cashless accounts, ticket buyers, gate scans, wifi registrations, and email subscribers, and reach it directly across the closed platforms, streaming, and mail. Then build a modeled expansion of it to prospect against, seeded with your best attenders rather than all of them.
The strongest source, if you have one: a closed-loop cashless system. Where an event requires its own wristband, chip, or app for every purchase on the grounds, and registration is required before funds can be loaded, every registered spending account becomes a record the event holds itself. Set up correctly, that registration record carries a home ZIP code, giving the event a first-party view of its trade area rather than relying entirely on a panel estimate. It also ties geographic origin to behavior on the grounds: spending by ZIP code, by day, by hour, and by vendor, which no purchased visitation dataset supplies directly. For a free festival with no ticket-buyer file, a mandatory registered cashless system can provide the most comprehensive first-party trade-area dataset available to the event while also tying geographic origin to onsite spending behavior. Visitation data still covers what it cannot: attendees who never spend, the years before the system was installed, and events you do not own. Run both in the same season and compare them.
Whether it delivers that depends on setup, and four points decide it.
- Registration has to carry the ZIP. Require registration with name, email, and ZIP before a wristband, chip, or app account can be activated or loaded, whether registration occurs before the event or onsite. Because funds cannot be loaded without registration, every spending account begins with a geographic record the event controls. Cash or card payments used to load the account do not supply the ZIP; the registration does.
- Coverage is spender-level, not head count. One parent may load funds for an entire family, and children or guests may never have separate spending accounts. Treat the result as household- or spending-account-level coverage rather than an attendance count.
- The contract decides ownership and access. Confirm in writing that the event owns or has full permitted use of the registration and transaction data, what the cashless provider may do with it, and that the data can be exported in usable form after the season.
- The terms decide the permitted uses. State the intended purposes at registration, including payment, trade-area analysis, and the event’s own marketing, and make sure those uses are reflected in the event’s privacy notice and consent process. Vendors should receive only the transaction information they need to operate, and sponsors should receive aggregated reporting unless the event has a separate lawful basis for sharing identified records.
Card data collected independently by a concessionaire on its own payment terminals is not automatically part of the event’s first-party audience file. Unless the event’s agreements give it lawful access to that customer information and permit its use for the event’s own marketing, treat it as data held by the concessionaire and its payment providers rather than as an event audience record.
How a closed-loop system works on the grounds, and how to choose one, is the subject of the third white paper. For this play, the point is narrower: an event that has one already holds one of the best-qualified first-party audiences in this paper and a strong first-party view of its trade area.
Ask for the ZIP code at the cash window. Cash walk-ups are the crowd ticket records miss, and the simplest fix happens at the point of sale. Have box office staff ask each cash buyer for a home ZIP code and record it with the number of tickets in that sale: in the ticketing or point-of-sale system if it has a field for it, on a tally sheet if it does not. It takes seconds, identifies no one, and covers every cash transaction rather than a sample, so the walk-up share of your trade area is counted rather than estimated. Keep it a question, never a condition of the sale. Where the same window also takes cards, ask every buyer the same way rather than as part of the card payment; in California, requesting a ZIP code as part of a card transaction has been held to violate state law.
Gate surveys complete the map, not the list. White Paper One’s intercept survey asks for a home ZIP code and nothing that identifies the respondent, which is what keeps it quick and unintrusive. Those ZIP codes fill in what your ticket, registration, and cash-window records still miss: walk-ups the window did not capture and, at a free festival without a cashless system, most of the audience. They tell you which areas to point the profile route in Section 4 and ZIP-level geotargeting toward. They do not become a list you can reach, and the survey works because it asks for so little.
Why it comes first. It is the only play with no third-party audience acquisition cost, no precise-location product availability question, no 50,000-device geoframing floor, and no vendor lookback window. The audience still carries state privacy and consent obligations, which is why the opt-out step below comes first. It is also the highest-responding audience an event has, for the unglamorous reason that people who came last year are the people most likely to come again.
How to run it.
Export the list, all of it, from every system that holds one: the cashless platform first where there is one, then ticketing, email, and everything else. Reconcile the duplicates. Suppress the opt-outs before anything is uploaded rather than after.
Split it by recency. Last season, two to three seasons ago, and lapsed beyond that. These three want different messages and justify different budgets, and treating them as one list is the most common way this play underperforms.
Match it into the closed platforms, and build the exclusion at the same time. The exclusion is what stops your prospecting budget landing on people who already bought.
Seed a modeled audience from the best of it. Not the whole file. The households that bought early, bought more, or came multiple years.
Send the same audience to a profile platform and read what comes back. This is the step organizers skip, and it is the one that makes every later play cheaper: the segments that come back are the input to Sections 9 and 10.
What it costs. Media, plus the staff time or service cost of cleaning the list, integrating it, and onboarding it. The media is cheap because the audience is small, warm, and does not need to be bought twice. On Google, targeting the list requires an account that clears the history and spend threshold in Section 3; below it, use the list on Google for suppression and run the targeting on Meta.
What can go wrong. The list has never been cleaned, matches poorly, and the organizer concludes the channel does not work. It is the list, not the channel. Budget one season of maintenance before judging the result.
9. Targeting a Comparable Event
The second play buys something you cannot build.
What it is. Identify an event that draws the audience you want, meaning a comparable fair, a festival in an adjacent market, a rodeo, a state fair two hundred miles away, and reach the households that attended it.
Two ways to run it, and they are not equivalent.
The device route. A geoframing vendor draws a boundary around the comparable event’s grounds during its dates and assembles the devices observed inside. This is the more precise version and the more restricted one. It is subject to the state availability map, the lookback window, and the audience floor, and in a state that prohibits the sale of precise geolocation it is not available at all. Because the floor is counted against the comparable event’s devices rather than yours, a large comparable can clear it where your own event would not, which is the one circumstance where a standalone organizer can buy a device audience at all.
The profile route. Your visitation platform reads the comparable event’s trade area and segment composition, and a segment audience is built from that description. Less precise, broader, and available in places the device route is not. It also does not decay on the device clock.
An organizer who can run only one should run the profile version. It is generally available more broadly than device activation, avoids the precise-geolocation sale bans discussed here, and compounds, subject to state privacy law, vendor policy, platform eligibility, and minimum audience size. An organizer running both should use the device audience for the near-term ticket push and the profile audience for the season-long prospecting, which is a distinction of purpose rather than of quality.
Choosing the comparable well. This is where the play is won or lost. The right comparable is not the biggest event nearby. It is the event whose attenders resemble the attenders you want more of, which is a question your own profile from Section 8 has already answered. Organizers routinely target the largest regional event and reach a crowd with no reason to drive to theirs.
The professional courtesy question. Nothing prohibits this and everyone in the business does it. It is still worth deciding in advance how you would answer if a fellow organizer asked, and worth remembering that the arrangement is symmetrical.
10. The Shared Headliner
The third play is the one only a booked event can run, and it is the one with the clearest line to the buying decision.
What it is. An act on your bill is playing eleven other dates this year. Every one of those dates assembles an audience of people who demonstrably will pay to see that act. Reach those households, in the markets close enough to yours to matter, with an advertisement built around the act rather than around your event.
Why it works. When the comparable date isolates the act, and a single-headliner show is the cleanest case, this is the strongest behavioral signal in the paper: the audience has demonstrated a willingness to attend a live event featuring that artist. Where the source date is a festival or a multi-act bill, you know the household attended, not why, and the audience should be treated as a comparable-event audience rather than an artist audience.
How to run it.
Get the routing early. This is a booking conversation, not a marketing one, and it is the reason this play belongs to organizers who have a talent buyer at the table rather than a vendor.
Decide the radius by drive time, not by distance, and take it from your own trade area rather than a rule of thumb. The trade area analysis in White Paper One shows how far your attendance actually travels. The honest test is whether the household would make the trip, not whether the pin lands inside your circle.
Build the audience by whichever route is available in that state, per Section 9.
Write the creative around the act. This is the play where generic event advertising fails hardest. The household is responding to an artist, and the advertisement has to lead with the artist.
The limit, stated plainly. This play can be built and it can be measured only within the constraint White Paper One established and Section 12 repeats: location data measures days and places, not performers. You can reach the audience of a shared date. You can compare the response of the market where you ran it against the market where you did not. What you cannot do is claim that the resulting ticket sales prove the artist’s value on your bill, because your bill has other acts on it and the day has weather.
11. Developing the Markets You Are Missing
The fourth play turns the fence around.
The first three plays reach people who have already shown what they want: your own buyers, a comparable crowd, a headliner’s fans. This one reaches the households that should be buying and are not. Most geofencing asks who was at the event. This play asks where the next customer should come from, and it moves location tools from capturing an audience to developing a market.
What it is. Use the addresses on your own ticket orders to measure how deeply you sell into each Census block group in your trade area. Find the block groups that resemble your strongest ones and sit within your drive time but buy far less. Then point advertising at them, by whichever method fits the objective and the state.
Why block groups rather than ZIP codes. A ZIP code is a mail delivery area, and a single one can hold a suburb that buys heavily and a neighborhood that has never heard of you. A block group is a Census Bureau area that generally contains 600 to 3,000 people, sits inside a census tract, and never crosses a county line. It is the smallest area for which the American Community Survey publishes household and income estimates, and it is the unit the leading visitation platforms use to report trade areas, which is why White Paper One described aggregated visitation in census block groups. Working at this level lets your own sales, public Census estimates, and a visitation platform’s profile describe exactly the same piece of ground.
Three layers, kept separate. This play depends on not blending three different things. Buyer geography comes from your ticket addresses and tells you where purchasers live. Market intelligence comes from Census estimates, segmentation, or a visitation platform and tells you what those places are like and which ones resemble your best. Activation comes from the advertising platform and decides whether a chosen area can actually be reached, and at what precision. A vendor who presents one product as all three is describing a chain, and each link should be checked separately. Figure 5 shows the sequence.
How to run it.
Export the orders. Order number, street address, city, state, and ZIP code, from every season you hold. Billing addresses are enough. Nothing is uploaded to an advertising platform at this stage, and nothing leaves the office.
Assign each order to a block group. The Census Bureau’s free Census Geocoder accepts batch files of up to 10,000 addresses and returns, for each match, the state, county, tract, and block codes as separate columns. Join them into the 15-digit block code; its first 12 digits are the block group code. Set that column to text before a spreadsheet touches it, or leading zeros disappear and long codes turn into scientific notation. Choose the 2020 Census vintage and use it for every step that follows. Boundaries are adjusted during the decade, and the first-12-digits shortcut is reliable only against the census geography the blocks were numbered in. Figure 6 shows how the codes nest.
Count and compare. Tally orders and tickets by block group, add the number of households in each from the American Community Survey, and divide. Tickets per thousand households is the penetration rate. It is the same calculation White Paper One ran by ZIP code, now on ground a fraction of the size.
Pick the targets. A target is not simply a low number. It is a block group inside your drive-time trade area whose households resemble the ones that already buy, by segment or by the profile your visitation platform returns, and whose penetration sits well below theirs. A block group that is low because its households look nothing like your audience is correctly low, and advertising will not change that.
Pull the boundaries. The Census Bureau publishes every block group’s outline in its free TIGER/Line boundary files, and its TIGERweb service can return selected block groups as GeoJSON, the format most platforms that accept custom shapes will take. Use the same vintage as the geocoding.
Two ways to reach an area, and they are not equivalent.
The area route. Upload the target block groups as geofences. At least one demand side platform, Simpli.fi, publishes support for custom shapes of any complexity and for bulk GeoJSON uploads, though the capability, the minimums, and the state availability should be confirmed for your account. What a polygon geofence reaches is devices observed inside it, which is not the same thing as the households that live there. In a residential block group those devices belong mostly to residents, along with commuters, contractors, delivery drivers, and anyone passing through. The route runs on precise device location in real time, so it sits on the state execution map. In the states that prohibit the sale of precise geolocation, and in others where vendors apply their own limits, minimum radii of a mile or more are common and neighborhood-level tactics are switched off. Most block groups are smaller than that. There the area route falls back to the ZIP code or radius that contains the target, and the block-group analysis still decides which one.
The household route. Select the residential addresses inside the target block groups, from a mailing list, a segmentation provider, or an address file the provider builds, and reach those households through the address-level targeting described in Section 5 and the mail described in Section 6. This is the route that actually reaches residents. Its exposure follows Section 5’s rule: addresses selected because they fall in a geography or a segment start from a file rather than from observed movement. Minimums apply here too; one platform’s published guidelines require at least 100 residential addresses per campaign.
Choose by objective. Use the underpenetrated block groups to identify priority markets, then choose the activation method that fits: an area geofence for people present in those areas, household targeting when the intent is to reach the people who live there. An organizer who cannot run either at block-group precision loses less than it seems by broadening delivery to ZIP codes, because the intelligence was the valuable part, and it still decides where the money goes.
Scale is the real limit, not the device floor. White Paper One’s fifty-thousand-device floor applies to historical audiences assembled from past presence at a place. An area geofence is a live campaign, with no list to assemble and no floor on the purchase. Its limit is size. A block group holds roughly 250 to 1,200 households, only some of their phones share location with the apps that carry advertising, and only some of those appear during the campaign. A single block group may yield a few hundred reachable devices or fewer, too few to spend a budget, to clear a platform’s delivery minimums, or to read against a holdout. Run the target block groups together as one geography, often twenty or more, rather than as one campaign each: the analysis stays precise, and only the delivery is pooled. For small, precise targets the household route scales better, because a matched household can be reached on its television and other screens whether or not a phone is sharing location, and mail reaches every address.
Adding audience layers. Once the geography is set, most independent platforms let you narrow it with demographic, interest, or behavioral segments: age, household income, music preference, past visits to comparable venues. The block group tells you where to look; the segments tell you whom to prioritize inside it. Two cautions apply. First, block groups are small, and every layer joined by “and” cuts the pool again. Four layers on a handful of block groups can leave too few reachable households for a platform to serve, or too few to measure. Target the priority block groups together as one geography and use one or two layers, joined by “or” where the layers are optional. Second, these segments are modeled estimates rather than verified facts, and segments built from past visits to places are device-derived location data, unavailable where the sale bans apply.
What it costs. The geocoder and the boundary files are free. The work is staff time in a spreadsheet and, if the resemblance test is to be done properly, the segmentation described in Section 4, which many organizers will already hold by this point. Media is bought as usual, on whichever route the state permits.
How it proves itself. This is the play best suited to the holdout in Section 12. Choose two comparable sets of underpenetrated block groups, advertise to one, hold out the other, and compare their penetration rates next season from the same ticket file. The measure is your own sales, in public geography, against a control chosen in advance.
What can go wrong.
- Buyers are not the whole audience. Ticket addresses identify purchasers, not every attendee. They show where buyers live, not where the whole crowd lives, and cash walk-ups do not appear at all.
- Small numbers mislead. A block group with three orders is not a stable measure of anything. Combine adjacent block groups, pool several seasons, or set a minimum number of purchasers before calling an area underpenetrated.
- Mixed vintages break the join silently. Geocode, count, and pull boundaries in the same census geography. In Connecticut, the Census Bureau replaced counties with nine planning regions beginning with 2022 products, so tract and block group codes changed even where the boundaries did not.
- A geofence sold as reaching residents is reaching whoever was there. Ask which of the two routes you are buying, in those words.
12. What You Can Honestly Measure
This section is deliberately conservative, because the measurement claims in this industry are where credibility is lost.
The three things that can be proven
- Delivery. Impressions served, to how many unique households, at what frequency, inside which geography. This is directly reportable, can be checked against platform logs, and should be demanded in a report you can read without the vendor present.
- Response. Clicks, site sessions, and above all ticket purchases attributable through a tracking parameter to a specific campaign. Where your ticketing platform can carry a parameter through to the transaction, this is the strongest number you will get.
- Difference between markets. Run a campaign in one market and not in a comparable one, then compare ticket sales by origin. This is the closest thing to evidence available to an event, and it costs nothing beyond the discipline of holding one market out.
The two things that are commonly claimed and rarely proven
Foot traffic attribution. A vendor draws a boundary around your gate and reports how many exposed devices appeared inside it. The number is real in the narrow sense that those devices were observed. It is not evidence that the advertisement caused the visit, because the households most likely to see an advertisement for your fair are the households most likely to attend it regardless. Without a holdout, this figure measures correlation and presents it as causation. Ask every vendor quoting a visit lift what the unexposed comparison group was. The answer tells you whether the number means anything.
Attribution to a performer. Repeating White Paper One because it matters more here: location data measures days and places. A day’s draw belongs to the whole bill, the weather, the ticket price, and whatever else was happening that weekend. Attribution becomes possible where an act is isolated, meaning a single headliner night, the same act across two years, or a day otherwise unchanged from the prior edition. Those comparisons carry real weight. The rest is directional and should be labeled as such in anything shown to a board.
How to build measurement that survives scrutiny
Hold one market out, every season, without exception. It costs a small amount of reach and it is the only thing that converts a report into evidence.
Instrument the ticket path before the campaign starts, not after. A parameter added in week three cannot be applied to week one.
Ask the gate. White Paper One’s intercept survey asks how people heard about the event. Tally the answers by channel every season, and compare the answers from ZIP codes inside the campaign markets with those from the holdout market. It is self-reported, and people tend to credit whatever they saw last, so it is a check on the holdout rather than a substitute for it: the holdout shows whether the campaign worked, and the survey suggests which channels people noticed.
Record what you did, in enough detail to repeat or refute it. Which audiences, which routes, which markets, which dates, which creative. Two seasons of that record is worth more than any single season’s dashboard.
Measure campaigns with holdouts and compare seasons for strategy. A holdout is the stronger evidence of whether a campaign worked, because weather, lineup, price, competition, and the economy all change between seasons. Year-over-year comparison answers a different question: whether the trade area widened, whether new segments appeared, and whether the origin of attendance shifted, which only resolve at the scale of the season.
13. Budget and Sequencing
An organizer with a fixed budget and no prior program should spend it in this order. Each step is useful alone, and each makes the next decision better.
First, and before any media spend: activate what you own. Export the lists, clean them, match them, build the exclusions. The cost is staff time. Nothing purchased later performs until this exists, because the exclusion list and the seed audience both come from here.
Second: profile activation. Read your own audience, identify the segments, and buy reach against a description of them. It is generally available more broadly than device activation, reaches more channels, avoids the precise-geolocation sale bans, and compounds across seasons, subject to state privacy law, vendor policy, platform eligibility, and minimum audience size. For an organizer who will only ever fund one purchased audience, this is the one. The block-group analysis in Section 11 tells you where to point it first.
Third: streaming, with the audience you already built. Television and audio inventory pointed at the audiences from steps one and two. No new audience purchase, a new surface.
Fourth, and only where the law and the arithmetic both permit: device-based advertising. The comparable event play and the shared headliner play, run through the open route, in the states where they are available, with the state execution map in writing, and only where the audience clears the vendor’s floor. This is the most restricted, most perishable, and most expensive category, and it is fourth for those reasons rather than because it does not work.
Fifth: direct mail against the same audience, if the per-household economics survive contact with your ticket price.
What to hold back
Reserve a share of the budget for the last three weeks. Event purchasing is compressed and late, the audience is warmest when the date is nearest, and a plan that spends evenly across the season will have exhausted itself before the moment it was buying for.
Reserve one market as a holdout. Every season. It is the cheapest measurement instrument in this paper.
What not to do
Do not start with device-based advertising. It is the most restricted and most perishable of the routes, and it answers no questions about your audience.
Do not buy a new audience when the constraint is creative. A better audience will not rescue a spot that fails to say what the event is, when it is, and why to come.
Do not commit to a multi-year subscription before a single season has been measured.
Reading the data is one thing. Buying against it is another.
TSE Entertainment has represented event buyers since 1975. We sit on the buyer’s side of the table exclusively, and our fee is built into the deal structure rather than added on top of it. We do not sell the analysis in this paper and we do not run it for you. What we do is take the findings you develop and weigh them in the offer we recommend, and in the routing conversation that makes the third play in this paper possible at all.
If you are building a lineup for 2027 and want your own audience findings weighed in the buying decision, we would welcome the conversation. tseentertainment.com
14. Where to Start
f this paper produced one action rather than a program, it should be this one.
This month. Export every list you own and find out how many unique households you actually hold. Most organizers do not know the number, and it determines everything else.
This quarter. Match that list into the platforms you already advertise on, including TikTok if your audience skews young, build the suppression, and send the same list to a profile platform to learn which segments you draw. You now know who your audience is and you can reach the part of it you already have.
This season. Add one purchased audience, by the profile route, aimed at the segments you learned matter, in the markets adjacent to the ones you already hold, chosen from the underpenetrated block groups Section 11 identifies. Hold one market out. Instrument the ticket path.
Next season. Compare. That comparison, and not any single campaign report, is the thing that turns this from a marketing expense into an input to the buying decision.
Everything else in this paper is refinement. An organizer who does only the four steps above, in that order, will be ahead of most of the field, and will have spent almost nothing on data.
Glossary
Terms defined as they are used in this paper, adding to the glossary in White Paper One rather than repeating it. Vendors sometimes use them differently, which is itself worth watching for.
Application-to-person messaging. Text messages sent by a business system to consumers rather than from one person to another. In the United States, this traffic over ordinary ten digit numbers must be registered before carriers will deliver it.
Brand and campaign registration. The carrier process for approving an application-to-person text program. The sending business is registered as a brand and each messaging use as a campaign. Registration is separate from consent, and a program needs both.
Census block group. A Census Bureau area generally containing 600 to 3,000 people, nested inside a census tract and identified by a 12-digit code. The smallest area for which the American Community Survey publishes household estimates, and the unit visitation platforms use to report trade areas.
Closed-loop cashless system. A payment system in which attendees load funds onto an event-issued wristband, chip, or app account that can be spent only at the event. With required registration, it is a source of first-party origin and spending data.
Connected television. Television programming streamed over the internet to a television set, with advertising bought programmatically and delivered by household.
Conversion zone. A boundary, usually drawn around your gate, that a vendor uses to count how many devices exposed to an advertisement later appeared there. Without a holdout it measures correlation, not cause.
Customer Match. Google’s tool for advertising to a customer list you collected yourself. Available to policy-compliant accounts for exclusion and observation; targeting requires meeting account history and spend thresholds.
Demand side platform. Software that buys digital advertising across many sites, apps, and streaming services through automated auctions, and accepts audiences from outside sources.
Display & Video 360. Google’s demand side platform. A separate buying environment from Google Ads, with different audience rules, including support for third-party segments.
Frequency cap. A limit on how many times the same household or device sees an advertisement in a given period.
GeoJSON. A standard text format for map shapes. The usual way to hand a block group boundary to an advertising platform that accepts custom polygons.
Holdout market. A comparable market deliberately left without a campaign so that results in the markets that received it can be compared against it.
Identity resolution. Linking records that refer to the same household or person across different systems, so that an audience can be recognized on platforms other than the one where it was built.
Incrementality. The share of a result that would not have happened without the campaign. Estimated by comparing an exposed group or market against a held-out one.
Lookalike audience. An audience a platform builds from people who resemble a seed list you provide. Also called a modeled or similar audience.
Onboarding. Converting offline or customer records, such as names, email addresses, and postal addresses, into identifiers that advertising platforms can recognize.
Prior express written consent. The form of consent generally required under the Telephone Consumer Protection Act for the marketing texts and calls it covers: a written or electronic agreement that identifies the number and authorizes marketing messages to it.
Programmatic. Automated buying and selling of advertising through platforms and auctions rather than through orders placed directly with each publisher.
Suppression list. A list of people deliberately excluded from a campaign, such as past buyers during prospecting or anyone who has opted out.
TIGER/Line files. The Census Bureau’s free boundary files for every census geography, including block groups, keyed to the same codes as its published data.
Vendor Question Checklist
Print this and take it into the meeting. It is written to be asked out loud.
About the audience
- Which of the three routes in Figure 1 is this, in plain terms?
- Where does the underlying data come from, and what happens if that supplier changes policy?
- How far back does the lookback window reach, and how quickly does the audience decay?
- What is the minimum audience size, and what happens if my trade area falls below it?
About availability
- Which of your products cannot run in which states today? In writing.
- Which of those restrictions are statutory and which are your own policy?
- Who maintains that map, and how often is it reviewed?
- What happens to an audience already delivered when a household opts out?
About activation
- Which destinations are live for my account at my spend level today, and which are roadmap?
- Is activation included in this quote or a separate purchase?
- If you list Google as a destination, which Google product, audience type, and integration do you actually use?
- Is audience expansion switched off by default, and can you show me the setting?
About area targeting
- Will you accept custom polygons, such as Census block group boundaries uploaded as GeoJSON?
- Does this geofence reach devices present in the area, or the households that live there?
- What is the smallest area or radius you will run in each state my audience lives in?
- After the audience layers are applied, how many households or devices remain, and will the campaign still serve?
About money
- What share of this budget is media and what share is fee?
- What is the minimum commitment, and can I buy one season before subscribing?
About what I keep
- What reports, counts, and audience files remain mine when the campaign ends?
About measurement
- If you quote a visit lift, what was the unexposed comparison group?
- Will you support a holdout market, and will your reporting show it separately?
Sources and Further Reading
Platform policies, vendor terms, and state statutes referenced in this paper change frequently. Every item below should be confirmed as current before it is relied on. Sources were reviewed September 23, 2026, and the census geography and area-targeting sources on September 30, 2026.
Regulatory: location data
- Virginia SB 338, restricting the sale of precise geolocation data, effective July 1, 2026.
- Oregon HB 2008, enrolled text, prohibiting the sale of personal data that locates a consumer or a linked device within a radius of 1,750 feet, effective January 1, 2026, with the Legislature’s measure summary.
- Maryland Online Data Privacy Act, effective October 1, 2025, prohibiting the sale of sensitive data.
- Connecticut Public Act 26-64, enacted from Senate Bill 4, amending the Connecticut Data Privacy Act to prohibit the sale, sharing, or transfer of precise geolocation data, effective October 1, 2026.
- Texas Data Privacy and Security Act, defining precise geolocation as sensitive data and requiring consent to process it.
- FCC v. AT&T, Inc. (2026), decided June 4, 2026, on FCC forfeiture orders arising from carrier sales of customer location data. Procedural history: the No. 25-406 docket, recording the June judgment and remand to the Fifth Circuit, and the No. 25-567 docket, recording the denial of Verizon’s petition for rehearing on August 17, 2026. The carriers’ later refund requests are reported by Broadband Breakfast.
Regulatory: text messaging
- 47 C.F.R. § 64.1200, the Federal Communications Commission’s rules under the Telephone Consumer Protection Act, including prior express written consent and revocation of consent.
- FCC Report and Order FCC 24-24, adopted February 2024, permitting revocation of consent by any reasonable method.
- FCC Order DA 26-12, released January 6, 2026, extending to January 31, 2027 the effective date of the requirement that a revocation apply to all future robocalls and robotexts on unrelated matters.
- Texas Senate Bill 140 (89th Legislature), effective September 1, 2025, bringing text messages within the state’s telephone solicitation statute.
- Texas Secretary of State, Telephone Solicitation FAQ, confirming that a business sending text messages with prior consumer consent is not required to file the Chapter 302 registration statement.
Carrier and messaging industry requirements
- Twilio, full blocking of unregistered 10DLC traffic, effective September 1, 2023.
- CTIA, Messaging Principles and Best Practices, the industry’s consent, opt-out, and age-gating practices for non-consumer messaging.
- CTIA, Short Code Monitoring Handbook, version 1.9, identifying alcohol as SHAFT content that must be age-gated appropriately and may be subject to additional carrier review.
Platform policies
- Google Ads, Customer Match policy, including the first-party collection requirement, the account history and spend thresholds for targeting, and the 540-day membership limit.
- Google Ads Help, Lookalike segments in Demand Gen, including the 2026 transition to suggestion mode, and Lookalike segments in Video campaigns.
- TikTok Ads Help, About location targeting, including ZIP code targeting, the 3,000-location limit per ad group, and the absence of location exclusions.
- TikTok Ads Help, Supported IDs and formats for a customer file, listing email, phone, and mobile advertising IDs.
- TikTok Ads Help, How to create a Lookalike Audience.
- Variety, TikTok U.S. Joint Venture Deal Set to Close in January, on the ownership structure of TikTok USDS Joint Venture LLC.
- Display & Video 360 Help, audience list targeting, including third-party audience lists.
- Display & Video 360 announcements, including the April 2026 expansion of lookalike segments to YouTube line items.
- Meta, Custom Audiences terms, including advertiser representations of rights, permissions, and lawful basis, and Custom Audience Terms of Service for the Marketing API.
- Meta Ads Manager, location and audience-expansion settings described in Section 3. These options change and are not documented on a stable public page; they are stated as reviewed in September 2026 and should be confirmed in the live campaign interface before each campaign.
Segmentation and location intelligence
- Spatial.ai, PersonaLive segmentation and activation destinations.
- Experian, Mosaic, including the 2026 rebuild and the February 2027 retirement of Mosaic V7 syndicated audiences.
- Azira, Azira One launch announcement, September 1, 2026, and the Azira One beta program page.
- Placer.ai, Unacast, Foursquare, Advan, Environics Analytics, and AGS product documentation, as listed in White Paper One.
Census geography and area targeting
- U.S. Census Bureau, Census Geocoder, free address geocoding in batches of up to 10,000 records, returning state, county, tract, and block codes.
- U.S. Census Bureau, TIGER/Line Shapefiles, block group boundaries, with block groups defined as generally containing 600 to 3,000 people.
- Federal Register, Change to County-Equivalents in the State of Connecticut, June 6, 2022, adopting Connecticut’s nine planning regions as county equivalents.
- Simpli.fi, How Modern Marketers Reach Local Markets With Geo-Location, March 2026, describing custom-shape geofences and bulk GeoJSON uploads.
- Simpli.fi, Geo-Fencing Overview, describing geofence audiences as built from areas visited by devices.
- GroundTruth, What is GroundTruth doing to ensure user privacy?, listing the states where it requires a one-mile minimum radius and disables neighborhood and location-based audience tactics.
- GroundTruth, Content and Targeting Guidelines, including a minimum of 100 residential addresses per campaign.
TSE Entertainment, LLC has represented event buyers since 1975, booking talent for fairs, festivals, casinos, theme parks, rodeos, and corporate productions. TSE is based in Austin, Texas, and operates exclusively as a buyer-side representative.
This paper is provided for general information. It is not legal advice. Organizers should review any activation program, and any text messaging program in particular, with counsel familiar with privacy and telemarketing law in the states where their audience resides.