When Google Ads does not pay off for an online shop

Percentage margin is the wrong test

Whether Google Ads can work for a shop is decided by the order value in euros, not by the margin in per cent. Two cost blocks do not automatically grow with the order value: the click cost per order (CPC divided by conversion rate) and the fixed costs per order, such as shipping, picking and packing. Together they set a lower limit:

Minimum order value = (CPC / conversion rate + fixed costs per order) / gross margin in per cent

Calculate the order value net of VAT, and the gross margin after cost of goods, payment fees and losses from returns. A worked example with invented numbers: suppose a shop pays €0.50 per click, converts 2% of visitors, has €6 of fixed costs per order and a 40% gross margin. That gives (25 + 6) / 0.4 = €77.50.

The same 40%, two different answers. If you are well below the limit, you do not have an optimisation problem, you have a problem with your offer. The usual break-even ROAS, selling price divided by margin per order, only tells you which value you would need to reach. Whether that value is reachable at all is something you have to settle first. Calculating break-even ROAS with all costs included is only the second step.

Which conversion rate would you need?

Rearrange the formula and you get the test figure for your actual order value: required conversion rate = CPC / (order value × gross margin - fixed costs). For the €40 order value in the example, that is 0.50 / (16 - 6) = 5%.

For context: the Smarter Ecommerce analysis of European e-commerce gives Shopping conversion rates of around 3.5% in the third quarter of 2025 and a peak of 4.5% in the fourth. If you permanently need 5%, you would have to beat even the peak of this market average. Some shops manage it, but I would not base a launch on it. At this CPC the numbers rule the channel out until order value, margin or repeat purchases change.

The CPC moves the result considerably. The same analysis gives Shopping CPCs of roughly €0.29 to €0.30 in the third quarter of 2025 and around €0.34 in November. At €0.30, the example would only need 3%, which looks achievable. I still treat market figures as the favourable case: they come from established advertisers whose numbers probably include brand searches and returning buyers, and a new shop tends to start worse. On top of that, the vendor Channable reports a 15% rise in CPCs from June 2025 to June 2026. So plan with a buffer rather than with the average.

At what margin does it get difficult?

There is no flat percentage limit; it follows from the order value. The required gross margin is (click cost per order + fixed costs) / order value. With €20 of click cost per order and €6 of fixed costs, a €50 order needs 52%, a €100 order 26% and a €200 order 13%.

The table shows the minimum order value in euros, again assuming €6 of fixed costs:

The columns are based on vendor data and are not a forecast for your shop: depending on the sector, Channable reports a CPA of €8.95 to €21.74 for Shopping and €9.90 to €26.72 for PMax. The €30 column is my assumption for a weaker start. Read the table as a warning: with a 25% gross margin, even in the favourable case you can hardly get below a €64 order value.

Repeat purchases lower the threshold

Repeat purchases can carry a CPA above the gross margin. In the €40 example you lose €15 on the first order, and every follow-up order without click costs brings in €10. On average, the new customer therefore has to come back one and a half times. Take this figure from your shop system. How far it stretches depends on how much a new customer may cost.

Also calculate each part of the range with its own CPC, its own conversion rate and the real order value, not the unit price. Several items per order, bundles or a free shipping threshold can lift the order value above the limit. That is why I would work on the offer first, then on the campaign. And if you have a lot of brand searches, your account will show better figures than you can expect from new demand.

Which ranges struggle?

The table is my assessment based on the logic of the formula. The sector alone says little; what matters is which quantity breaks in your range. There are differences all the same: according to Channable, the Shopping CPA for home and garden is more than twice that for baby and pet products.

Check demand and price before you spend money

Google Ads captures existing purchase intent. If there is no demand for the specific product, you first need a channel that creates it. You can check this with a reverse calculation: 30 orders a month at a 2% conversion rate need 1,500 clicks. If the searches for your purchase terms clearly cannot deliver that, you do not need to test. In Keyword Planner, note the difference: forecasts take bid, budget and seasonality into account, historical metrics do not. Use the historical metrics for the question of demand and the forecast for the budget question.

Merchant Center shows your price position: the Pricing tab reports how many products are more expensive or cheaper than the benchmark and where the largest gaps are. For this, Google requires a valid GTIN, so own brands without a GTIN get no comparison and have to compare their key products by hand. If your bestsellers sit clearly above the benchmark, I would expect a weaker conversion rate than the market average, and with it a higher minimum order value. What this means for reach is covered under price position in Shopping.

Closed, not ready yet, or open?

Keep these three states apart, because a test is only worth it in the third. In the first, it burns money for an answer you already have.

How to test with little risk

Treat the test as buying information: decide in advance what loss the answer is worth to you and how many orders the budget can produce at all. Suppose the €100 shop from the example releases €1,000. That is 2,000 clicks and, at a 2% conversion rate, 40 orders. Each order leaves €34 before ad spend.

That fixes the rule before you start: below 30 orders, the answer on the current basis is no. Test only the part of the range above the minimum order value and evaluate brand searches separately, otherwise they flatter the result.

Allow around six weeks, roughly €25 a day in the example, and answer the question without target ROAS. According to Google, the performance of new campaigns can fluctuate until conversion data has settled, after which you evaluate over two conversion cycles, but at least two weeks. For target ROAS, Google names 15 conversions in 30 days as the minimum.

In the example, €300 buys 600 clicks and 12 orders; a single order more or less shifts the €108 result by around a third. That would be chance rather than insight. In that case I would advise holding the money back and working on the offer first.