How to test an advertising channel in six weeks

Key takeaways

To test an advertising channel, change one thing, track every response and decide the pass mark before you start. Six to eight weeks is long enough for most channels. A well-run test ends with a clear decision even when the result is disappointing.

• 1 channel, 1 area, 1 offer at a time.
• 6 to 8 weeks, or about 20 tracked enquiries.
• Set the pass mark before spending anything.
• Record the result and what you would change.

What is a channel test?

A channel test is a small, time-limited campaign run to answer one question: does this channel bring customers at a cost we can accept? It differs from simply trying something, because it has a budget, a tracking method, an end date and a rule for what happens next.

• Social media marketing vs traditional advertising: what works better for SMEs?
• The advertising channel scorecard: pick where to spend

What do you need before you start?

You need a baseline, a tracking method and a number. The baseline is your normal weekly enquiries. The tracking method ties responses to the channel. The number is the most you can pay for a customer, worked out from what a customer is worth.

• How to track offline advertising: six methods that work
• Cost per lead by channel: comparing social media and offline

What are the steps?

Follow the same seven steps for any channel, online or offline.

• 1. Write the question the test must answer.
• 2. Choose one area or audience.
• 3. Choose one offer and one message.
• 4. Set the budget and the end date.
• 5. Add tracking and start the log.
• 6. Run it without changes.
• 7. Compare the result with the pass mark.

How much should the test cost?

The test should cost enough to produce a readable result and no more than you can afford to lose. For paid social that is £300–£500 a month for two months. For leaflets it is two drops to about 5,000 homes, roughly £600 to £1,500. Radio and outdoor need more.

• The minimum budget to test an advertising channel properly
• How much does advertising cost a UK small business?

Why should you change only one thing?

Change one thing so you know what caused the result. If you start leaflets, a new website and a price cut in the same month, a rise in enquiries cannot be credited to any of them. Keep everything else steady, including your other advertising.

How do you compare two channels?

To compare two channels, run them at the same time, in separate but similar areas, with the same offer. Or run them one after the other in the same area, allowing for season. Then compare cost per enquiry and cost per customer, not the number of calls.

• Leaflets vs social media ads: which wins local customers?
• How to compare advertising ROI across online and offline

How do you read the result?

Read the result against the pass mark you set. Under the mark: scale up carefully. Close to it: improve the advert and run once more. Well over it: stop. Check the follow-up as well, since a channel can be blamed for enquiries nobody answered.

• Losing leads to slow follow-up? When to outsource it
• Advertising payback period: how long until it pays for itself?

What makes a test misleading?

Four things make a test misleading.

• It was too small to rule out chance.
• It ran in an unusual week, such as a holiday.
• The advert was weak, so the channel took the blame.
• Tracking missed part of the response.

• Advertising mistakes small businesses make, and the fixes
• Seasonal advertising: timing spend across channels

What does this look like in practice?

A pattern we see with firms that say they have tried everything: each attempt was a single advert with no tracking and no end date, so nothing was learnt from it. The second time round, run as a proper test, at least one of the supposed failures usually turns out to work.

• See client work
• Social Media Marketing for UK businesses

Next step

Write one sentence: 'We will spend £X on Y in Z area for six weeks, and continue if a customer costs under £N.' That sentence is your test plan.