Measuring a Radio Ad With the Hourly Report

What a listener does after the spot

A thirty-second spot runs at 7:40 on the morning show. Somebody in a car hears the name, likes the sound of it, and does nothing until they park.

Then they search for the business on their phone, or type the address if it was easy to remember. A few do it at once, more within the hour, some that evening.

That leaves a short cluster of visits after each airing, mostly from phones, arriving through a search for the name or with no source at all.

The cluster is real but small, and it vanishes in the daily totals most people check.

Where the response shows up

• Direct visits, from people who typed the address they heard.
• Searches for the name, landing on the home page.
• Phone visits, especially after spots in drive time.
• Calls, from people who tapped the number in the results.
• Visits later in the day, once listeners reach a computer.

The second is usually the largest and the hardest to see. Analytics hides the words people searched on Google, so a visit from someone who searched your name is given away only by its timing.

Daily totals hide it

A handful of extra visits after each spot, spread across a day, looks like an ordinary busy Tuesday. Weather or a newsletter can move the daily figure more than the spots do.

The response is concentrated in time, so it has to be measured in time. Hour by hour, a cluster twenty minutes after a spot stands out against the same hour on a day without one.

Switching Analytics to hours

In Universal Analytics, the graph above most reports has buttons for hourly, daily, weekly and monthly views. Set the dates to the days the spots ran and choose hourly.

For easier comparison, build a custom report with the date and the hour as dimensions and sessions as the measure, then export it to a spreadsheet.

Check one setting before trusting any of it: the time zone in the view settings. Analytics reports in that zone, so if it is wrong, every cluster appears in the wrong hour.

Get the actual air times

The schedule agreed with the station says roughly when spots will run. Stations keep a log of what actually aired and can usually provide one.

A spot booked for the morning show may have run at 6:52 or 8:15, or moved to another day. Clusters that follow the real times, not the booked ones, are about as clear as this evidence gets.

A fair comparison

Every site has a daily rhythm. Visits rise at lunch, dip mid-afternoon and climb again in the evening, when people browse on their phones.

So compare each hour with the same hour on the same weekday in the weeks before the flight, not with the hour before the spot. Three or four earlier weeks give a fair sense of normal.

Check where the extra visits landed and on which devices. Phone visits to the home page mean something different from computer visits to an old blog post.

A worked example

A flower farm in the valley east of the city sells bouquets from a farm stand and takes pre-orders on its website. Last week it ran spots on a local station ahead of Mother's Day.

Afterwards the owner opened Analytics, saw a busy week, and could not separate the spots from the usual rush before the holiday.

The hourly view looked stranger still: clusters of visits roughly three hours after each spot. The view had been set up by a designer in Toronto and was reporting in Eastern time.

Shifted back three hours, it made sense. Each morning spot was followed within half an hour by a cluster of phone visits to the home page and the pre-order page.

The mid-afternoon spots left nothing that could be told apart from an ordinary afternoon. The time zone has been corrected, which applies from now on, and next season's budget is going to the mornings.

Make the next flight easier to read

Use an address that is easy to say and spell on air. If the domain has hyphens, numbers or odd spellings, listeners will search the name instead, and some will land on a directory.

A short page address read only on the radio, such as your domain followed by /radio, appears in Analytics as its own landing page. Few will type it, but each one is counted.

If you run AdWords ads on your own business name, their impressions can also be broken down by hour, which is about as close as you can get to counting the searches.

Calls and walk-ins count too

Plenty of listeners never open the website. They search the name and tap the phone number in the results, or simply drive over.

Note the time of every call during a flight, and ask new callers and new faces at the counter how they heard about you.

Reading what you find

A clear cluster after most spots says listeners heard the name and went looking. It does not say how many became customers, and it misses everyone who responded days later.

No cluster is not proof of failure either, since radio often works by repetition. What the hourly view does well is comparison: mornings against afternoons, one station against another, this flight against the last.

The counter-case

A business that runs only a few spots a week, or whose site gets few visits, may see nothing either way. The numbers are too small to separate from chance.

Businesses whose customers walk in or phone will learn more from the counter and the call log, and a campaign meant to build familiarity over months is not designed to produce a burst on the day.

Where the website is part of the response, though, check the time zone, get the air times, and read the hours rather than the days.

Before the next flight

1. Check the view's time zone in Analytics.
2. Record normal hourly visits for a few weeks.
3. Choose an address that is easy to say.
4. Ask the station for actual air times.
5. Compare each spot's hour with earlier weeks.
6. Log calls and walk-ins during the flight.
7. Compare mornings and afternoons before rebooking.

Step four is the easiest to leave out. Without the real air times, the hourly figures are matched against a schedule that may not have happened, and clusters get credited to the wrong spot.

Tracing any sudden jump back to its source works the same way, as covered in a spike in visits that came from one link.

Frequently asked questions

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.