Facebook, Instagram and Google Ads for Kent small businesses: one campaign, real bookings as the scoreboard, a watcher every four hours, and decisions on dated review points.
Clicks are cheap to count and easy to report. They are also not what you sell. A dance class sells a child in the room. A restaurant sells a table booked. A barber sells a chair filled.
Most small-business ad reports stop at the click, because that is what the platform shows first. I do not. Before a pound is spent, the outcome that actually matters is defined, measured and test-fired.
The honest limitation comes first: ads cannot fix a booking page that does not work. I check the destination on a phone before the first pound goes out, and I have found dead booking buttons on live sites more than once.
Before any spend, the outcome event, a booking, an order or a call, is created on your pixel and test-fired, so the platform learns towards the thing you sell. Then there are three instruments, not one: the platform's count, the pixel's count, and your own record, the calendar, the till or the hall register.
When the three agree, the numbers are real. When they disagree, that is the first thing I investigate, before anyone celebrates.
The campaign itself is simple on purpose. One campaign, one ad set, two or three creatives. The copy is held identical, so the creative is the only variable and the test measures one thing.
A read-only system polls the live campaign every four hours and logs every reading. It applies fixed rules: delivery, pacing, cost per booking against your own target, frequency, click-through trend and placement waste. Then it posts a plain-English digest.
It changes nothing on its own. It cannot. Every change to a live campaign is a human hand, mine, with the rollback named before the change is made.
That means you are never reading a report written from memory. You are reading what the campaign did, on the record, every four hours.
On launch day the decision points go in the calendar: the five-day read, the event day, and the day after. The minimum evidence for any change is written down before launch, so nobody moves the goalposts once the numbers arrive.
No creative is switched off on 24 hours of data. Budgets are never moved mid-learning. A campaign that is left alone long enough to be measured is a campaign you can learn from.
I will not launch to a page I have not checked on a phone. I will not switch creative because a day looked quiet. I will not move a budget while the platform is still learning. I will not report clicks as if they were customers.
If the numbers say the campaign is not working, I will say so on the dated review, and I will say what I would change and why.
This suits a small business with something bookable: a class, a table, a chair, a slot. It suits an owner who keeps their own record of bookings and is willing to share it, because that record is the third instrument.
It suits a budget that is real but modest, where every pound needs to be watched and no pound can be wasted on guesswork.
A booking, an order or a call, measured on your pixel and checked against your own calendar or till, never a click.
The watcher posts a plain-English digest every four hours, and the decisions happen on the dates set on launch day.
I check it on a phone before any spend; if it fails, the ads do not run until it is fixed.
No; the system only reports, every change is made by hand with the rollback named, and you see the digest.
It does, inside the campaign; what it cannot do is decide whether your business is winning, and that decision stays with a person on a date.
Email hello@weareconversionforge.uk, or request the free website audit at conversionforge.co.uk/free-audit so the destination is checked first.
A free audit, no pitch: whether people find you when they search, what is quietly costing you customers, and what to fix first.