Radio has a measurement problem that has nothing to do with whether it works. A spot airs, someone hears it in the car, and an hour later they search for you on their phone. There is no link to click, no pixel to fire and no cookie to follow. So most radio advertisers end up judging stations by gut feel, by a promo code a few listeners remember, or by audience estimates alone.
There is a better way, and it is the same way linear TV is measured: line up exactly when each spot aired with what happened on your website and your phones in the minutes that followed, and measure the lift against what would have happened anyway. This guide explains how that works for radio, what you need, and where radio is harder than TV.
Why promo codes and vanity URLs undercount radio
Most radio advertisers already try to track response with one of these:
- A promo code read on air
- A vanity URL or a dedicated landing page
- A dedicated phone number per station
- A "how did you hear about us?" question at checkout or on the call
Keep them all. But each one depends on the listener remembering and repeating something they heard once, often while driving. Most don't. They search your brand name, type your main web address or call the number on your website. The result is a count that is too low, and low by a different amount on every station, which quietly distorts which stations look best.
What you need to measure radio
- Your station logs. The record of every spot that actually ran: station, market, date, time and creative. They arrive as Excel files, PDFs or CSVs, in every format imaginable, and that is fine.
- Read-only access to Google Analytics 4. Minute-by-minute web traffic, by location, from the analytics you already have. Nothing to install, no pixel, no tag changes.
- Optionally, your call data or order data — from your call-tracking provider or your store — if calls or orders matter more than site visits.
How radio attribution works
The method is identical to the one used for broadcast, cable and satellite TV:
- Build a baseline. From months of your own traffic, work out how many visits you would normally get at that minute, on that hour, on that day of the week — accounting for spots that aired in the past.
- Find the lift. The response credited to a spot is the traffic in its response window minus what the baseline expected anyway.
- Let the window fit the data. The window closes when traffic returns to normal, instead of after a fixed number of minutes.
- Share credit fairly. When two spots air close together, they split the response rather than both claiming it.
Because the baseline comes from your own history, the result is incremental response: the visits, calls and sales that would not have happened without the spot.
Local radio: measure each market against its own traffic
This is where radio differs most from national TV, and where most of the difficulty lies.
A national spot can be measured against your national traffic: one series of minutes per day. A local spot on a station in one city should only be measured against traffic from that city — its market or state — so that a busy evening in one place is not credited to a spot somewhere else.
Do that for a schedule running across dozens of stations in many markets, minute by minute, and the data volume multiplies quickly: minutes in the day, times markets, times stations, times devices. It is a big-data problem, which is one reason radio is measured far less often than TV. But it is the only way to get local answers that hold up, and it is the default for every local spot we measure.
What kinds of radio can be measured
Anything that airs at a known time in a known place:
- Local broadcast radio, market by market
- National and network radio
- Satellite radio
- Live reads and host endorsements — get the time the read ran, and it is treated like any other airing
What cannot be measured this way is audio without an airing time: podcasts and on-demand streaming, where every listener hears the ad at a different moment. Without a timestamp there is no moment to measure the lift against.
Small spots and how much you need
Radio schedules are often long lists of inexpensive spots, and a cheap spot produces a small lift. One small spot on its own is hard to read. Many of them together are not: the same station and daypart, measured over weeks, builds a clear picture of what it delivers for the money.
Two things make results clearest:
- Steady baseline traffic, so a spot's response stands out from normal variation
- Spots of roughly $100 and up, so each airing has a chance to move the needle
Measured this way, the result is often surprising. The inexpensive station with small, steady upticks can deliver more response per dollar than the marquee spot with the obvious spike.
What the report shows
The output is your station log with new columns next to every spot: incremental visits, incremental calls or orders if you include them, and cost per response for that airing. From there it rolls up by station, market, daypart, day of week, hour and creative.
The weekly question stops being "is radio working?" and becomes:
- Which stations and dayparts delivered response at a cost we are happy with?
- Which ones delivered almost nothing for the money?
- Which creative works, and in which markets?
Radio and TV in one report
If you run both, they are measured with the same method and reported side by side, so cost per response on a radio station can be compared directly with cost per response on a TV network. That is often the most useful view of all: where the next dollar should go across your whole linear media plan.
Common questions
Can radio be measured if my listeners don't visit the website right away?
Yes. The response window follows the data rather than a fixed cutoff, so response that builds more slowly is still counted as long as it rises above the baseline.
Do I need a promo code or a separate phone number per station?
No. Attribution is based on time and market, not on what the listener types or says. Codes and dedicated numbers are a useful cross-check, not a requirement.
Can you measure live reads?
Yes. Give us the time each read ran and it is measured like any other spot.
What about podcasts and streaming audio?
Not with this method. They have no single airing time to measure against.
Getting started
Quality Analytics has measured linear TV and radio response for advertisers and agencies since 2014, across hundreds of campaigns and hundreds of millions of dollars of media. Send us your station logs and give read-only access to GA4, and we will have you set up the same day and return your first report the next business day. We run up to a month of your airings before you pay anything, and plans start at $399 a month, month-to-month. Book a demo to see it on your own data.