MarketReadHow MarketRead works, from your website to ads that run themselves
Nothing exists in your ad account until MarketRead builds it, nothing spends until you press Go live, and everything can be paused in one tap. Here is the whole journey, and the rules it runs under.
- about 60 seconds
See your ads, free and with no sign-up
Paste your website. MarketRead reads it, looks at your photos, works out who buys from you and where, and shows you up to four finished ads in about a minute. If you don't like them, you've lost nothing.
- about 2 minutes
Connect your Meta ad account
You approve MarketRead through Meta itself, so no password is shared, and you can withdraw access from your own Meta settings whenever you like. You need a Facebook Page and an ad account with a card on it; if you've ever boosted a post, you have both.
- about 1 minute
Confirm the brief
Everything is filled in from your preview: what a result means to you (an enquiry, a booking, a sale), where your customers are, and your monthly budget. Change anything that's wrong, and add photos from your phone if you want.
- under a minute
It builds the campaign, paused
Campaign, audience and ads are created in your own ad account, every one of them paused. You see exactly what was built, and the daily budget, before anything exists that could spend a penny. Building is free.
- one tap
You press Go live
Your 14-day free trial starts. Each part is switched on and then read back from Meta to confirm it really is on. If any step fails, everything goes back to paused and you're told: a half-live campaign is never shown to you as live.
- daily
It runs the account, every day
Losing ads paused on evidence, winners given more budget in safe steps, tired ads replaced, the month kept on pace. Every decision written down in plain English, and a summary in your inbox each Monday.
The rules, in full
No black box. These are the thresholds MarketRead runs on, and each decision in your log states the numbers it used, so you can always check its working.
| Rule | What it does | When it's allowed to |
|---|---|---|
| Learning week | Leaves new ads alone | For Meta's first week, nothing is changed unless an ad is clearly broken, because changes during learning restart it. |
| Pause a losing ad | Pauses it and moves its money to the others | Only when it's at least 90% certain the ad's true cost per result is 50% over your target, judged on what it has actually spent. |
| Pause a dead ad | Pauses it | When it has spent two and a half times your target cost per result with nothing to show. An ad that has been working gets more rope: it is paused only once its dry spell since its last result reaches four times your target. |
| Back a winner | Raises its budget by 20% | When it's at least 80% certain the ad beats your target, with about ten results in a week. One step every 72 hours, never past your monthly ceiling. |
| Replace a tired ad | Makes a fresh ad and puts it live | When an ad is paused, or people have seen it too often and its cost is rising. At most twice a week. |
| Keep the month on pace | Trims the daily budget | Whenever spend runs ahead of the month. Your monthly budget is a ceiling and is never exceeded. |
We'd rather be late and right than early and wrong with your money: small budgets are noisy, so the bars are deliberately conservative.
What we measured before building it
We replayed the rules over £437,790 of real UK ad spend: 21 lead-generation campaigns, July 2023 to August 2026, walking forward one day at a time and using only the numbers that existed on that day.
We report one figure from that replay: money that went to an ad already provably past its target on the evidence available at the time. We don't publish a "results you would have gained" number. That needs an assumption about how a winner would have scaled, which nobody can measure, and most tools quote it anyway.