Short answer: Yes, the Dutch market is ready for ChatGPT Ads. The users are there, many questions are about purchases, and since the end of August every Dutch business owner can advertise themselves. Most online stores, however, are not ready yet. If you can't see what an order earns per channel after all costs, you can't judge this new channel fairly either.
Since 24 August 2026, Dutch users have been seeing ads in ChatGPT. A week later, the self-serve Ads Manager opened to European businesses. The question store owners are now asking is whether they should get in. In this article we set out what is available, what still causes friction and how to set up a test that actually tells you something.
Where does ChatGPT Ads stand in the Netherlands?
In nine months, ChatGPT Ads has gone from a test in the US to a channel in dozens of countries. The Netherlands was part of the European rollout across 31 countries, according to NL Times and Digiday.
| Date | What happened |
|---|---|
| 9 February 2026 | First test in the US, only via large advertisers |
| 5 May 2026 | Self-serve Ads Manager in beta in the US |
| 24 August 2026 | First ads visible in the Netherlands |
| 31 August 2026 | Self-serve Ads Manager opens in Europe |
A few things to know before you start:
- Who sees the ads: only users on the free plan and on Go, the cheapest paid plan.
- What they look like: ads are labelled. According to OpenAI, they don't influence the answers and conversations are not shared with advertisers.
- How you pay: most spend now runs on pay-per-click, according to Digiday.
- What you can target: in Europe, at country level. Region, city or postcode targeting is only available in the US (Indexlab).
- Who can advertise: businesses registered in the EU, Norway, Iceland, Liechtenstein or Switzerland.
Why the market is ready
The Netherlands is among the fastest adopters of ChatGPT. According to an overview by Searchlab, around 4.8 million Dutch people use ChatGPT at least once a month. That is more than a quarter of the population.
More important than the number of users is what they ask. In June, OpenAI said that around 20 per cent of queries have a direct commercial intent. Someone asking which sofa fits a small living room, or which running shoe suits overpronation, isn't idly scrolling. They're in the middle of a buying decision.
That is the moment that once made Google Ads big. The difference is that there are still few advertisers in the auction. If you learn how the channel works now, you'll probably learn it more cheaply than in a year's time.
Why there's still friction
There are four reasons to be cautious. None of them is a reason to stay away.
- There are no benchmarks. There's no public data on click-through rate, conversion or ROAS. Media agencies quoted in Digiday still call the platform unproven. So you can only compare your results against your own figures.
- Measurement is difficult. People read an answer, compare options and come back later via Google or directly. If you only look at the last click, you'll see little and draw the wrong conclusion.
- Privacy rules are stricter than in the US. Under the GDPR (AVG in Dutch), personalised targeting is only allowed with explicit consent. In Europe, you therefore advertise mainly on context and intent, not on a profile.
- Trust is at stake. People use ChatGPT precisely because it doesn't sell anything. A pushy ad stands out more here than in a feed full of other ads.
Worked example: a six-week test
A test only makes sense if you know in advance when it counts as a success. So start by calculating your break-even CPA: the contribution margin before advertising per order. You can read how that works in CAC and payback period: how much can a new customer cost?.
Suppose you spend €20 a day for six weeks. Your average order is €80 excluding VAT (21% in the Netherlands), and after product costs, shipping, packaging, payment fees and returns you keep 40 per cent of it. The cost per click and conversion rate below are assumptions: there are no benchmarks yet.
| Item | Calculation | Result |
|---|---|---|
| Ad budget | 42 days × €20 | €840 |
| Clicks | €840 ÷ €0.60 per click | 1,400 |
| Orders | 1,400 × 1.5 per cent | 21 |
| Revenue excl. VAT | 21 × €80 | €1,680 |
| Contribution margin before advertising | 21 × €32 | €672 |
| Result after advertising | €672 − €840 | −€168 |
| Actual CPA | €840 ÷ 21 | €40 |
| Break-even CPA | 40 per cent of €80 | €32 |
At first glance, this is a loss-making test. Even so, you don't have to stop it straight away. First look at three things:
- How many orders came from new customers? A new customer who comes back can cost more than their first order earns.
- What did your total revenue do? If your direct traffic rose or more people searched for your brand name, you won't see that in the ChatGPT clicks. So also look at your MER, as explained in Calculating ROAS, POAS and MER.
- Which ads did work? An average often loses out to the best half of your ads.
Break-even CPA = average order value excl. VAT × contribution margin before advertising as a percentage
If after six weeks you're below your break-even CPA, or just above it with lots of new customers, the channel is worth scaling. If you're well above it and your total revenue stays flat, you know that too. Either way, you've got an answer for €840.
Our view: the market is ready, most stores aren't
Whether the Netherlands is ready for ChatGPT Ads is really the wrong question. Consumers are already asking a chatbot their buying questions, whether stores keep up or not.
The real question is whether your store is ready. Most stores look at the Google Ads dashboard, the Meta dashboard and their revenue in Shopify. Three sources that each tell their own story and together often claim more revenue than actually comes in. Add a fourth channel without benchmarks and after three months you still won't know anything.
In our own homeware store, we're testing the channel with a small daily budget. What we're learning isn't primarily whether ChatGPT Ads works. It's that you can only judge it if your own figures are in order: revenue per channel, margin per product and what a new customer really costs you.
A year from now, there will probably be two kinds of online stores. Those that ran small, well-measured tests now and know where ChatGPT Ads fits in their mix. And those that get in once it has become crowded and expensive.
What you can do now
- Get your basics in order. Know what each channel earns in contribution margin, not just in clicks. See Calculating contribution margin.
- Set up conversion tracking before you spend your first euro. Without tracking, you learn nothing from a test.
- Agree in advance when the test is a success. Record your break-even CPA, your budget and the duration.
- Look beyond the last click. Also watch direct traffic, searches for your brand name and new customers over the same period.
- Write ads that help. People are asking a question. An ad that gives an answer fits better than a shouty discount.
- Test one country at a time. Start in the Netherlands and only then compare with Belgium or Germany.
Frequently asked questions
Can I advertise in ChatGPT myself as a Dutch online store?
Yes. Since 31 August 2026, the self-serve Ads Manager has been open to businesses registered in the EU. Your ads do have to pass OpenAI's policy review.
Do all ChatGPT users see ads?
No. Only users on the free plan and on Go see ads. Users on a more expensive plan don't see them.
What is a good ROAS for ChatGPT Ads?
There are no reliable benchmarks for that yet. Work with your own break-even point: the ROAS or CPA at which you break even after all variable costs.
How much budget do I need to test?
There's no fixed minimum that's right for every store. Choose an amount you can afford to lose and that generates enough clicks to see something, for example €15 to €25 a day for six weeks.
Want to see what an order really earns you per channel, including when you start testing ChatGPT Ads? Ecommetrics puts your figures from Shopify and your ad channels alongside your product and shipping costs, excluding VAT. Join the waiting list.