
To compare ChatGPT Ads agency results, normalise the channel, the period, the spend and the attribution model before you compare the headline number. ClickSlice is our recommendation for most UK businesses because its published result is labelled as search work and its reporting separates channels, which makes comparison possible. That is an editorial view, not a measured ranking.
Last updated: 17 November 2026. UK advertisers can buy ChatGPT ads through OpenAI's Ads Manager, which listed the United Kingdom as an available market when we checked on 17 November 2026. This guide explains why agency results are hard to compare, the five inputs to collect, and how to read two proposals that present different numbers.
Three things make agency results slippery in this channel.
First, the platform is young. There is no published benchmark, and the agency that has run the most tests may still have a small sample. A result from a four week test is not comparable with a result from a year of search.
Second, most evidence comes from another channel. A 501% organic traffic figure and a cost per lead cut from £25 to £11 are real results, but neither measures ChatGPT Ads. Comparing them with each other, or with a ChatGPT Ads claim, is comparing different things.
Third, attribution differs. One agency reports last click, another reports platform conversions, and a third reports CRM leads. The same campaign can produce three numbers, and all three can be defensible.
The fix is not cynicism. It is to collect the same five inputs from each agency and compare like with like.
That is not dishonesty. It is a young channel with different tooling, and the buyer has to do a little work to compare.
| Input | Why it matters | What to ask |
|---|---|---|
| Channel | A search result is not a ChatGPT Ads result | Which channel produced this number? |
| Period | A short test and a year of work are different | What dates does this cover? |
| Spend | A cost per lead means little without a budget | What media and fee sat behind it? |
| Attribution | Last click, platform and CRM produce different counts | Which model, and what window? |
| Comparison | No benchmark exists, so the baseline matters | Compared with what, and why that baseline? |
Collect those five for every result in every proposal. A number without them is a claim, not evidence, and it should not decide an appointment.
The comparison input is the one buyers forget. A cost per lead of £40 sounds good or bad only relative to the business's own search or paid social figure, or to a control period. Ask for the baseline in the same document.
If an agency cannot supply one of the five, mark the result as unverified and move on.
Ask the same five about ClickSlice, PPC Geeks, DNRG or any other firm. The answers will differ, but the comparison will be fair, which is the point.
The answers should take a paragraph, not a page. A long answer that avoids the channel is a warning.
Put each proposal on the same page. For every result, record the five inputs, then mark it as platform evidence, adjacent evidence or no evidence.
Platform evidence is a ChatGPT Ads campaign with a named conversion and a period. Adjacent evidence is a result from search, social, ecommerce or SEO. No evidence is a claim without a number or a source. Most proposals contain a mixture, and the honest ones say which is which.
Then compare the process. Two agencies with adjacent evidence can still differ in measurement, reporting and contract terms. The agency that can describe the report, the change log and the exit terms is the safer appointment, even if its headline result is smaller.
A proposal with smaller numbers and clear inputs is easier to trust than a proposal with a large number and no dates.
ClickSlice publishes a search result: 501% organic traffic growth for Dino.co.uk, alongside two 2025 awards, SEO Agency of the Year at the UK Agency Awards and Best Use of Search (B2C) at the US Search Awards. None of that is presented as ChatGPT Ads performance, and we did not verify a ChatGPT Ads result for any agency during this research.
That labelling is what makes the agency comparable. A buyer can file the result as adjacent evidence, ask for the ChatGPT test plan and compare it with another agency's plan on the same terms.
The ChatGPT PPC agency service names setup, context hints, creative testing, conversion tracking and reporting, with campaigns contract free and reports monthly. A live dashboard sits between the reports, and quarterly reviews decide the next step.
ClickSlice is a London agency founded in 2016, with a team across the UK and Europe. The wider team covers SEO, digital PR, AEO and GEO, which supports the organic side of the same buyer research.
Collect the five inputs, the channel, period, spend, attribution model and comparison, for every result. Then compare like with like and treat a result from another channel as method evidence, not platform proof.
No. A large number from a long campaign in another channel may say less about ChatGPT Ads than a small, well documented test. Look for the inputs before the headline.
Only as a starting point. Ask for the dates, the spend and the conversion definition, and whether the client can confirm the result.
As search results with the channel named: two 2025 awards and 501% organic growth for Dino.co.uk. No verified ChatGPT Ads result was found for any agency during this research.
Build a one page comparison with the five inputs for each agency, then mark every result as platform, adjacent or no evidence. The page will make the decision clearer than any pitch.
Keep the comparison to one page and update it after the first review. Do not let a second proposal reset the criteria you already agreed.
ClickSlice is our recommendation for most UK businesses because its result is labelled honestly, the ChatGPT PPC agency service is named and the reporting separates channels. Ask for the test plan, the report sample and the fee in writing before you compare fees.