Scaling, and knowing when to stop
Performance at a small budget is not a rate you can multiply. It is the cheapest slice of the audience. Decide the stopping rule before you are attached.
Something is working. This is the point where most accounts get worse, because the two obvious moves, spending more and launching more, both change the conditions that made it work.
Read the numbers that mean something
The platform will show you dozens of metrics. Three of them decide whether you are making money, and the rest explain movements in those three.
- Cost per conversion, against what a conversion is worth to you. The scoreboard.
- Return on what you spend, if you sell something with a price attached. Same scoreboard in a different currency.
- Volume at that cost. A good cost per conversion on four conversions a week is not a business, it is a data point.
- Everything else, click rate, cost per thousand, hook rate, watch time, is diagnostic. Use them to explain why the first three moved, never as the target.
Reconciling the platform with the bank
What the ad account reports and what your bank shows will not agree. The gap has widened as tracking has become less reliable, and arguing with it wastes weeks. Treat it as a known offset and measure it once a month.
- Each month, write down what the platform claims it produced and what actually arrived that you can attribute to advertising.
- Work out the ratio between them. Two thirds, half, whatever it is for you.
- From then on, judge campaigns against each other using the platform's figures, which are internally consistent, and judge whether the channel is worth running using the bank.
- Recheck the ratio when anything changes: the site, the tracking, a browser release, the offer. It moves.
Asking buyers how they found you is worth doing alongside this, and is also unreliable. People forget, and they credit the last thing they saw. Use it as a sanity check on the direction, never as a number.
Scaling without breaking it
- Raise budget gradually rather than multiplying it. Large jumps push delivery back into learning and you lose the stability you were paying for. Increments of around a fifth, a few days apart, are a common rule of thumb and the principle matters more than the figure.
- Widen before you deepen. More places and more people at the same creative usually beats forcing more spend through the same narrow setup.
- Add creative as you add budget. More spend through the same adverts means the same people see them more often, which brings fatigue forward.
- Expect the cost to rise. At higher spend you are reaching people who were harder to convince. The question is not whether it rises but whether it is still under your ceiling.
That last point is where optimistic forecasts die. Performance at a small budget is not a rate you can multiply. It is the cheapest slice of the audience, and it does not extend indefinitely. Doubling the budget and getting sixty per cent more conversions is a normal and acceptable outcome, not a failure, as long as the cost per conversion is still inside what a customer is worth.
Knowing when to stop
Decide the stopping rule before you start, in money and in time, and write it down somewhere you will see it when you are attached to the campaign.
- A spend ceiling with no conversions, derived from what a customer is worth rather than from how long you can bear it.
- A date. Not only an amount, because a slow burn passes the ceiling without ever feeling like a decision.
- A condition for stopping the whole channel, not just one campaign. If three different offers across three angles all failed with tracking working, the honest conclusion may be that this channel is not where your customers are.
Written out, illustrative: a customer is worth 600, so the ceiling is 1200 spent with nothing to show. The date is six weeks from launch. The channel condition is three offers across three angles. Any one of the three being hit means stop and reassess, not stop and quit, and the difference between those two is what the written rule protects.
Stopping is not failure, it is information bought at a known price. The failure is the account that has been running for eight months at a loss because nobody wrote down what would count as enough.
Work to do
- Write your cost per conversion ceiling from what a customer is worth in profit, and put it where you will see it.
- Compare what the ad account reported last month against what actually arrived, and write down the ratio between them.
- Write your stopping rule as three lines: an amount, a date, and the condition for stopping the whole channel.
- If you are scaling, plan the budget increases as steps a few days apart rather than one jump.