Performance Max for a small shop: when the budget says it is too early

Below 60 conversions a month, PMax with a target ROAS is usually too early

I would only use Performance Max with a target ROAS once a campaign reaches around 60 conversions a month over several months. That is my assessment and a working threshold, not a Google requirement. Others draw the line somewhere between 30 and 50.

The number only helps you once you translate it into your own figures. PMax is too early if the budget cannot buy the conversions, or if you cannot evaluate the test result. Both can be calculated before launch, and the second problem barely features in many recommendations: even a campaign that runs smoothly is of little use if, at the end, you decide on the basis of chance.

Budget check: 60 times order value divided by target ROAS

Your target ROAS determines what a conversion may cost: order value divided by target ROAS. Multiplied by 60, that gives the monthly budget the working threshold requires.

With €2,000 a month, Shop A would be well funded, while for Shop B it would be too early. That is why a blanket budget threshold is of little use: it ignores order value and ROAS, the two figures that determine the result. If your planned budget is below your figure, the 60 conversions at the desired ROAS do not add up mathematically, and more patience will not change that.

For cash flow, divide the monthly budget by 30.4. Shop A comes out at around €40 a day. According to Google, a single day can cost up to twice that, but over the month you will not be charged more than the daily budget times 30.4. The check only tells you about money, though. It does not answer whether there is enough demand for your range.

Readability check: at 20 conversions, ROAS 4 is not a result

Google advises letting a PMax campaign run for at least 6 weeks before evaluation and excluding the 1 to 2 weeks of ramp-up from the analysis. That leaves around four weeks on which to base your judgement. How much the result of those four weeks is worth depends on the number of conversions. As a rough approximation, the purely random uncertainty of a conversion count, expressed as a 95% range, is plus or minus 2 divided by the square root of the number of conversions. That is my own calculation.

Suppose your break-even ROAS is 3. At 20 conversions, the range extends down to 2.2: the test could be showing a profit or a clear loss, and you cannot tell the difference. A clear signal there only comes from a measured ROAS below about 2.1 or above about 5.5; anything in between means carry on or stop at the loss limit. At 60 conversions, the lower edge sits roughly on your break-even, which you can work with. That is why I stay with 60 rather than 30. Order values of different sizes make ROAS even noisier than the table shows, so the ranges are, if anything, on the generous side.

If your target is well above break-even, you can manage with fewer. If you work close to break-even, even 60 are not enough.

What happens in the campaign with few conversions

According to Google, automated bidding needs typically 1 to 2 weeks, but sometimes up to 6 weeks, and performance and spend can fluctuate during that time. Google names conversion volume as one of the factors for the pace, alongside the conversion cycle and the campaign settings. Tight targets and low daily budgets can slow the ramp-up further. In my assessment, a small shop often hits several of these brakes at once.

Two properties of PMax add to this. The campaign offers only automated bid strategies, so you cannot fall back on manual CPCs. And it can serve on YouTube, Display, Search, Demand Gen, Gmail and Maps. My assessment: with a large budget, that is reach; with a small one, it mainly means less insight into and less influence over what each euro is spent on. For your margin, this means you pay for the fluctuation out of a budget that is poorly placed to absorb it.

What to do instead below the threshold

Below the threshold, use the setup whose result you can control better at your volume. The thresholds are my decision rule; only the 15 comes from Google. As of November 2026, it applies to target ROAS for Search and Shopping and is counted at the conversion tracking level, while my 60 are per campaign.

The test route in the middle row matches Google's advice for preparing for the season to start with Maximise conversion value when the budget is limited. Store Growers likewise advises consolidating below 30 conversions per campaign. How to weigh up the two campaign types per product group is covered in PMax or Standard Shopping.

Decide in advance what the test may cost

Before launch, calculate what a bad run would cost: test budget times (1 minus achieved ROAS divided by break-even ROAS). Suppose a shop spends €1,200 a month for six weeks, which comes to around €1,650. The break-even ROAS is 3, but only 2 is achieved. That leaves a loss of around €550 after cost of goods and advertising. If you cannot carry that amount, it is too early, even if all the other figures fit. If you can, write it down before launch as the stop limit, so that you do not decide on gut feeling in week three.

When switching pays off

I would switch once these conditions are all met at the same time:

• One campaign reaches around 60 real purchases a month, over several months and not just in season.
• Your budget passes the check for the ROAS you actually need.
• You know your break-even ROAS and can carry the cost of the test.

The groundwork in Standard Shopping is not lost: according to Google, conversion data from previous campaigns can help the campaign settle faster. You still wait the six weeks before the first evaluation.

Older accounts can test earlier

With faulty tracking, the calculation is worthless, so the reconciliation with the shop system comes first. It also only counts real purchases. 60 add-to-cart conversions do not meet the working threshold.

An account with a long history does not start from zero and can test earlier. A long conversion cycle or strong seasonality pushes the threshold up, because the data reaches you later or only in part of the year.

A columnist at Search Engine Journal considers PMax worth testing precisely with a limited budget and low conversion volume, because fragmentation does damage. If you would otherwise spread your budget across many small campaigns, a single PMax campaign can indeed do better. I think the argument is right, but it is an argument against fragmentation. You achieve the same with a single Standard Shopping campaign, and there you keep the manual options. How PMax fits in beyond that is described in the overview of Performance Max for online shops.