By Anthony Gaita
TL;DR
- Store-visit attribution connects an ad someone saw to a real visit they made, which lets you prove media drove foot traffic and not just impressions.
- The number in your dashboard is an estimate. Measurement companies watch a sample of location-sharing phones, then scale that up to represent your full audience. Understanding this makes the number more useful.
- A strong setup rests on three things: good coverage in your markets, a solid match rate, and a control group to measure lift against.
- For screen-off channels like programmatic audio and CTV, store-visit lift is a primary way to prove impact, because those channels can’t be measured by clicks.
Every multi-location advertiser wants to know whether the money they spent on ads got people through the door. Store-visit attribution is the tool built to tell them. Someone sees your ad, walks into a store a few days later, and attribution draws the line between those two events. That line is what turns ad spend from a leap of faith into something you can measure. Here’s how it works under the hood, and what the number in your dashboard is really telling you.
What store-visit attribution is measuring
The goal is to connect an ad someone saw to a physical visit they made. Someone sees your display, video, audio, or DOOH ad, and later shows up at your store. Attribution ties those two moments together. If you can tie them together, you can show the ad did something, and that’s the reason advertisers pay for it.
The mechanism is location data from a mobile phone. When a phone that was served your ad is later spotted inside the boundary of a store location, that phone gets counted as a visit. The ad exposure and the location signal get matched using a device ID or a probabilistic match. That match is the engine of the whole product.
How the number is built
Start with the problem the measurement company has to solve. They want to know how many people who saw your ad later walked into your store. But they can’t see inside every phone in the country. They can only see the phones that have agreed to share their location, usually through apps like weather or navigation tools that ask for location access and pass that data along.
That group of location-sharing phones is the sample. Think of it like a political poll. A pollster doesn’t call all 250 million adults in the country. They call a few thousand, then scale the results up to represent everyone. Store-visit measurement works the same way. The company watches its sample of location-sharing phones, counts how many of them saw your ad and later showed up at your store, then multiplies that up to estimate the true total across your whole audience.
So when your dashboard says 1,240 store visits, the measurement company didn’t watch 1,240 phones walk into your stores. It watched a smaller number in its sample, maybe a few dozen or a few hundred, and scaled that up to 1,240 based on how big the sample is relative to your full audience.
This is why the size of the sample matters so much. If a company has strong coverage in your city, its sample includes plenty of local phones, and the estimate is solid. If it barely has any phones in your area, it’s scaling a tiny number up to a big one, and small errors get magnified. Same reason a poll of 50 people is shakier than a poll of 5,000. Before you trust a store-visit number, it’s worth knowing how many phones the company actually sees in the markets where your stores are.
Match rate: how much of your audience is actually visible
The sample is one limit. Match rate is the other, and it’s the piece most advertisers never ask about.
When your ad runs, it reaches a lot of phones. Match rate is the share of those phones the measurement company can actually connect to its location data. If your ad reached 100,000 phones and the company can tie 20,000 of them to location signals, your match rate is 20 percent. The other 80,000 are invisible to the measurement. Any visits from that group are estimated, not seen.
A higher match rate means more of your audience is being watched directly instead of guessed at. When you know the match rate, you know how much of your result is real observation and how much is the model filling in blanks. A store-visit number built on a 40% match rate is standing on firmer ground than one built on 10%, even if both land on the same total.
Why the method keeps getting sharper
The way location and identity data flow to advertisers changed, and measurement changed with it.
Apple’s App Tracking Transparency reset the baseline. As of Q2 2025, the industry-wide opt-in rate sat at 35 percent, meaning roughly two-thirds of prompted iPhone users decline cross-app tracking. That 35 percent is up only slightly from 34.5 percent in 2024 and 34 percent in 2023, so the smaller pool of shareable data is the steady state, not a temporary dip. On top of that, iPhone users can now hand an app their approximate location instead of exact coordinates. Approximate isn’t precise enough to tell whether someone walked into your store or the strip mall next door.
Measurement companies answered by getting better at estimating. They lean harder on statistical modeling and work to build bigger, more representative samples so the scaling stays accurate even as raw signal gets scarcer. Cookie and identifier deprecation pushed the same direction, moving the field away from tracking one specific phone across apps and toward broader modeled approaches. The result is a category that leans more on estimation than it did five years ago and, with a capable provider, reads more reliably for it. The advertisers who get the most out of it are the ones who understand what sits behind the number.
Setting it up to prove real results
A store-visit program gives you data you can actually take seriously when three things are in place.
Good coverage in your markets. A national average won’t help if your stores sit in areas where the measurement company sees very few phones, so confirm coverage where you actually operate.
A match rate you can see. Knowing how much of your audience is observed versus estimated tells you how much weight the result can carry.
A control group. This is the one that turns a store-visit count into proof. Split your audience in two. One group sees your ad, the other doesn’t. Then compare store visits between them. If the group that saw your ad visited more, that difference is lift, the visits your advertising actually caused. A raw visit count only tells you people who saw your ad also went to stores. Lift tells you the ad is why. Without a control group, you can’t separate the people your ad moved from the people who were already going to show up.
Get those three right and store-visit attribution becomes a number you can defend to a CFO.
Why this matters most for screen-off media
Screen-off media is where store-visit attribution earns its keep, and it’s the part Floodlight cares about most.
Programmatic audio and connected TV are mostly click-free. Nobody clicks a Spotify midroll or a streaming ad the way they click a search result. Last-click attribution can’t register these channels at all, which leaves an open question: how do you prove audio and CTV drove business? Store-visit lift answers it. It measures the outcome that matters, a visit to your store, without waiting on a click that was never coming.
For a multi-location advertiser running audio or CTV, store-visit lift is often the strongest near-real-time signal that the spend moved someone into a store. That’s why the setup carries so much weight. When store visits are the main proof for a channel, good market coverage, a strong match rate, and a real control group are what turn that proof into budget you renew without a second thought.
If you’re running audio or CTV to drive foot traffic and you want store-visit measurement built to hold up under a CFO’s questions, that’s what Floodlight sets up.
Frequently asked questions
What is store-visit attribution?
Store-visit attribution connects an ad someone saw to a physical visit they later made. When a phone that was served your ad is spotted inside a store’s boundary, the ad exposure and the visit get matched, so you can show media drove foot traffic and not just impressions.
Is the store-visit number in my dashboard an exact count?
No. It’s an estimate. Measurement companies watch a sample of phones that share location data, count how many saw your ad and later visited, then scale that up to represent your full audience. Strong coverage in your markets makes that estimate more reliable.
What is match rate, and why does it matter?
Match rate is the share of the phones your ad reached that the measurement company can actually tie to its location data. A higher match rate means more of your result is observed directly rather than modeled, so the number stands on firmer ground.
How do you prove audio or CTV drove store visits?
Because those channels are click-free, use store-visit lift. Split your audience into a group that sees the ad and a control group that doesn’t, then compare visits between them. The difference is the lift your advertising actually caused.