Should I run ads before anyone has paid?

Ads only tell you faster

If strangers will not use your product for free, ads will not fix that. You would be paying for visitors who leave. So find out first whether strangers want it at all. If people can try or buy your product without talking to anyone, the test is simple: does a stranger use it for free, without being pushed? If there is no free way in, the test is whether a stranger asks for a call, a trial or a price on their own.

Shops and phone games are different. They are sold through paid ads from day one, so the question there is not whether to run ads. It is whether each sale pays back what the ad cost, within a time you can afford to wait.

One week, then decide

If people buy your product on the spot and it is cheap, you do not need to wait two or three weeks to see whether ads work. Run the test for one week. If nobody pays in that week, stop the ad and go back to the landing page and the product. That only works if the week's spend brings enough visitors for a sale to be possible in the first place. The next section works that out.

If buying takes a trial, a call or a pilot, one week is too short. Then the result you look for is the first step a buyer takes, such as a trial started, a demo booked or a reply, and the test runs for as long as that step honestly takes. Either way, decide the spend and the stopping line before the test starts. Once money is going out, it is very easy to give it one more week.

How big the test should be

1. Decide your whole marketing budget first: the total you can afford to spend on getting customers over the next few months.
2. Give one test a small slice of that budget, about 5%. Prep66's founder spent £5 a day on his one test. He set that figure from his own budget, not from the price of the product. Work yours out from your own budget, in your own currency.
3. Check that the slice is big enough to show anything. Find out roughly what a click costs on that channel, and use a fair guess at how many visitors turn into a paying customer. If the money cannot buy enough visitors for about three sales, the test cannot tell you anything. Then pick a cheaper channel, judge an earlier step such as signups instead of sales, or accept that the test is too small to trust.
4. Have the landing page and its headline finished before any money goes on ads.
5. Judge the result at the end of the week and decide. Never let an ad run on because you hope it will pick up.

What the ads did on Prep66, Ancestorii and AquaPaws

• Prep66, a Reddit ad. £5 a day, aimed at two aircraft engineering subreddits in the UK and EU, from 27 August to 1 September 2026: about £25 spent, one signup, nobody paid. The founder switched it off without waiting for it to "warm up". To be fair to the ad, a test that small does not prove an ad is dead on its own. What settled it was that in the same days the free channels brought more people, and a paying customer.
• Ancestorii, the launch. about £210 in its first seven days of ads brought about 15,000 visitors and no paying customer. The site asked strangers to pay before they had tried anything.
• Ancestorii, two months in. By then it had spent more than £500 on Meta ads and £76 on Reddit ads, on top of everything else it tried, and it had four signups in total. Later, after a rebuild, about £4 a day on Meta ads brought around four signups a day. Signups reached 125. Still nobody had paid.
• AquaPaws, a Shopify store. A video was made on Fiverr in the style of the cat videos going viral at the time, and run as an ad at £30 to £50 a day. It worked: around ten strangers bought. But one sale did not cover a day's ad spend, and the money ran out. The ads brought buyers. The cash did not last.

The usual advice, and why this disagrees

Guides written by agencies that run ads (read on 17 September 2026) say the first four to six weeks of a campaign are for collecting data, and that a real answer takes two to three months. For a product people buy on the page in minutes, that means paying for weeks after the answer is already in. Ancestorii's founder could see in the first week that the ads were not working. He kept paying for months anyway, hoping it would change.

What to record for every ad

For every ad, every day, write down what you spent, how many clicks it brought, how many signups, how many payments, and where each signup came from. Put the cost next to the signups and the payments. A test with no record teaches you nothing, whatever it cost.