A 2x ROAS is not good or bad. Your margin decides.

01What ROAS is, and what it quietly hides

Return on ad spend is simple. Money in from the ads, divided by money spent on the ads. Spend 1,000, make 2,000 back, that is a 2x. Useful number. Half a number.

What it leaves out is everything that is not the ad bill. Making the thing. Delivering it. Staff. Software. Payment fees. Refunds. Tax. ROAS is a revenue ratio, and revenue is not what keeps a business alive.

Revenue is not the thing that keeps a business alive. What is left over is.

So 2 businesses can show the same revenue on the same ad spend and be in completely different health. The box looks the same from outside. Inside, one has plenty left and the other has nothing.

02Your break even ROAS, in 1 line

Here is the whole calculation. Take 1 and divide it by your gross margin, written as a decimal. 50 percent margin is 0.5, so 1 divided by 0.5 is 2. That is your break even ROAS. Below it you lose money on every sale. Above it you make money.

3 businesses. 1 identical ROAS. 3 completely different situations. That is why a benchmark borrowed from somebody else's business is worse than useless. It is confident and wrong.

Now put your own margin in and redo that table for yourself. The only row that matters is yours. This is the kind of small bit of maths I keep sitting inside the command center I build for an owner, next to the numbers, so nobody has to go hunting for it on a bad day.

03Why a small return on big spend beats a big return on nothing

Here is the trap that keeps small businesses small. A founder is proud of a 5x. Every time they raise the budget the return drops, so they stop raising it. The return stays beautiful. The business stays exactly the same size.

Do the maths with round numbers. A 5x on 2,000 a month of spend makes 10,000 of revenue. At 50 percent margin that is 5,000 of gross profit, and 3,000 of it went on the ads. 2,000 left.

Now a 2.5x on 20,000 of spend. That is 50,000 of revenue, 25,000 of gross profit, 20,000 on the ads. 5,000 left. Half the return. More than double the money you keep.

Scale profits, not the ROAS number.

A high return on a budget too small to matter is a vanity number with a spreadsheet attached. Put your own 2 options through those same 2 lines before you decide which one you would rather own.

04Set a base ROAS and let the rule decide

Once you know your break even, set a base ROAS a bit above it. That is your line. Above the line you raise the budget. Below it you cut. Now the decision is not a judgement call you make while tired.

Write it down. Everybody's instincts get worse as the spend gets bigger. A rule you set on a calm Tuesday is better than a decision you make on a bad Friday.

How far above break even to sit is the only real judgement here. Too close and an ordinary bad week pushes you into losing money before the rule reacts. Too far above and you choke spend that was actually working. Use how jumpy your own account has been over the last 3 months, not somebody else's number.

053 days, not yesterday

This is the discipline that saves the most money, and it is the one people skip. A single day tells you almost nothing. Traffic comes in lumps. Some sales get counted days late. One big order throws the whole picture off.

So take a rolling 3 day average and compare that to your base. Above it, raise the budget by a small step. Below it, come down. Then do it again tomorrow. Small, frequent, rule based moves beat big, rare, emotional ones every time.

One day is a footprint. 3 days is a direction.

This is the part I now hand to the machines. In the command center I build for an owner, the daily numbers land in one place and a Telegram assistant tells me what moved, so the 3 day check is a glance instead of a morning. The rule did not change. The work of running it did.

06Do not read the number off the platform

The ROAS in your ads manager is a modelled figure, built from partial data, and it is produced by the party you are paying. That does not make it a lie. It makes it a claim.

Read revenue where the money actually lands. Your payment account, your bank, your own sheet. Then do the division yourself. The 2 numbers will differ, sometimes by a lot, and the one from your bank is the one you run the business on.

If you cannot answer those, the number on the screen is not telling you what you think it is telling you. Same reason a buyer never believes your 20 page document and does believe 1 clear page, which I wrote about in nobody reads your 20 page proposal. The simple true number wins over the impressive one.

07The thing that really ruins the number

Platforms give the credit to the last click. But a real buying journey often runs across several ads over days. One opened the loop. One built trust. One broke the doubt. One collected the money. Only the last one gets the credit.

So the ads with a poor reported ROAS are often the ones doing the work that made the winning ad possible. Switch them off because the number looks bad, and 2 weeks later the winner gets worse and nobody can work out why.

That one mistake is expensive enough to deserve its own article, and it is the main reason I do not let a single ad be judged on its own reported number.

08When ROAS is the wrong number to look at

If the sale happens on a call, or the client stays with you for a year, or the first purchase is meant to lose money, then ROAS is measuring a slice of the thing and reporting it as the whole thing.

For most service businesses the pair that matters is what a client is worth over their whole life with you, against what it cost to win them. Plus how long you wait to get that money back. A 1.2x on the first sale is fine if the client is worth many times that over 2 years and you can afford to carry the gap.

So: is a 2x good or bad? It is good if your break even sits below it, you are reading revenue from your own bank, you are judging on 3 days, and the spend is big enough for the profit to matter. It is bad if your break even sits above it. And it is bad if you genuinely do not know what your break even is.

09Questions, answered