Checked 9 October 2026.
An online shop in Somerset that delivers across Britain can consider ChatGPT advertising today. A Sherborne business that only serves customers nearby needs to check whether its service area can actually be targeted before spending.
UK self-service access is available. Our recommendation is to treat it as a measured trial once the audience, offer and tracking fit, rather than moving a working advertising budget because the channel is new. OpenAI lists the United Kingdom as available in its Ads Manager country guide.
What you are buying
ChatGPT Ads are sponsored placements that can appear below a response. OpenAI says they are separate from its answers, and payment gives advertisers no control over how the answer ranks or recommends a business. Ads may reach eligible Free and Go users; Plus, Pro, Business, Enterprise and Edu are ad-free. That means the audience is a subset of ChatGPT users. OpenAI’s explanation of ads.
The practical opportunity is to put an offer beside a relevant conversation. Whether that produces a customer depends on the conversation, the offer and what happens after the click.
Can you reach customers in Dorset or Somerset?
For most campaign types, OpenAI documents regional, city and postcode targeting where supported. Availability varies by country, so check the locations offered inside your account. New product-feed campaigns have a different restriction: their geographic targeting and exclusions are country-level only. Current campaign setup guidance.
We have not confirmed that Sherborne, Dorset, Somerset or your particular postcodes are selectable. For a business with a small service area, that is a condition to check before committing any media budget.
Context hints describe the offer and situations where it might be relevant. They are neither exact-match keywords nor targeting rules, and they do not guarantee delivery for a particular audience. Putting “Dorset customers” in a hint is therefore no substitute for checking the location controls. OpenAI’s ad-group guidance.
A hypothetical furniture retailer might describe compact dining tables for people furnishing a small home, then send visitors to a page showing dimensions, materials, delivery and prices. The hint, ad and landing page should describe the same offer. A broad list of places and products gives you a less useful test.
Check the category rules too. UK availability does not mean every business can advertise: legal services outside the US are currently listed as not allowed, and health and financial categories have specific restrictions. OpenAI’s eligibility table.
Is it worth testing for your business?
This is how we would make the decision. The recommendations are our judgement, not forecasts of platform performance.
- You sell eligible products across the UK: Consider a limited trial. Your delivery area fits the campaign, margins are known and purchases can be measured.
- You sell an eligible service that customers research before enquiring: Consider a limited trial. The landing page answers buying questions and you can record which enquiries qualify.
- You serve only a small local area: Check targeting first. The selectable locations must fit the area you can genuinely serve. Do not rely on context hints as a geographic boundary.
- Your website or enquiry tracking is unreliable, or the budget cannot absorb an unsuccessful test: Wait. Fix the page or measurement, and protect channels already bringing you commercially useful customers.
There is no sound basis for assuming ChatGPT will be cheaper than Google or Meta for your business. OpenAI says it does not yet have performance benchmarks across advertisers, industries or campaign types. Advertiser FAQ.
Published pilots show why you need your own evidence. Windmill Strategy reports spending US$1,470 on a month-long B2B agency-services test, generating 259 clicks and no tracked form conversions. That test ran in June and July, with the controls available then. Windmill’s pilot.
InterTeam reports more than 150 qualified leads across three B2B software clients and a lower cost per lead than its own Google campaigns. These are self-reported results from different offers and sales processes; neither example predicts the outcome for a Dorset shop or Somerset service business. InterTeam’s case study.
Compare any trial with your own existing campaigns on accepted enquiries, customers and margin. A lower click price alone does not make a channel a better investment.
What should the trial cost?
The standard minimum daily budget for an account billed in pounds is £15 per campaign. That is a platform minimum, not our recommended test budget or a promise of enough data. Setup, creative, landing-page work and management can add to the media cost. Campaign budgets.
A daily budget is an average. With an unchanged £15 daily setting, spend can reach £30 on an individual day within the applicable £105 seven-day limit. Changes to the budget affect those limits, so read the account’s displayed controls. Daily budget rules.
For a fixed trial, we would favour the campaign-total budget as the stricter control on overall media spend. Agree separately what setup and management will cost, when the trial ends and what would justify continuing. Campaign budget options.
Start with one offer and one audience problem. Spreading a small budget across many products and ad groups can leave you with too little evidence about any of them.
Measure the trial before you launch it
1. Define the outcome. For a shop, that might be a completed purchase and its contribution after costs. For a service business, define an accepted enquiry: the right service, area, budget and timing. Decide what acquisition cost the business can support before looking at clicks.
2. Choose the objective around that outcome. Views, clicks and conversion objectives are supported. Conversion optimisation needs Pixel and/or Conversions API tracking and a supported standard event; it still charges for clicks or impressions, not per conversion. Choose an event you can record reliably. Conversion campaign requirements.
3. Test the page and measurement. Complete the form or checkout yourself on mobile. Check that the confirmation is recorded once, campaign parameters survive the journey and the team can follow up. Make the privacy notice clear about advertising measurement and any customer data shared. Configure measurement to respect consent before launch: the ICO says advertising storage and access technologies require consent, and the OpenAI Pixel defaults to consent enabled unless configured otherwise. ICO guidance, Pixel consent controls.
4. Keep a simple enquiry or order record. Record source, location, qualification, sale and value. Compare the same dates and time zone in Ads Manager and analytics. Export results and inspect click-through and view-through attribution separately where available. OpenAI notes that conversion reporting can take 24–48 hours to appear. Reporting guidance.
5. Make the decision on total cost and customer quality. In a hypothetical test, £450 in media plus £150 in setup producing three accepted enquiries means £200 per accepted enquiry. Whether that is affordable depends on how many become customers and the margin they bring. If there are no accepted enquiries, report the spend and zero outcome plainly.
An attributed conversion shows that the platform gave an ad credit. It does not, on its own, establish that the sale would not have happened otherwise. For a small trial, keep that limit in view and allow for your normal sales cycle before declaring success or failure.
If you want help deciding whether to test, send us your service area, offer, current advertising spend and how you measure enquiries. We can discuss the trial alongside your existing paid advertising, with a scope and quote agreed first. Our marketing budget guide can help you set the wider budget, or contact the studio to talk through the numbers.
