Our $100 ChatGPT ad test in Turkey: first campaign notes

Our first ChatGPT ad campaign in Turkey had a $100 spend limit. In the reporting period from 14 to 19 October 2026, Ads Manager recorded $99.81 in spend, 106 clicks and one attributed form event. The useful part of this small test is the record behind those numbers: what we advertised, how we measured the response and what changed while the campaign was running.

This is a campaign for our own advertising service. We have not established a paid-customer result from these figures.

The offer and the reporting period

We advertised our ChatGPT campaign setup and management service under our Turkish brand, AI Reklam Uzmanı. The campaign targeted Turkey and used a clicks objective. Both ads sent visitors to our SEYAR service landing page, where a business could submit an inquiry.

That buying process matters. A service inquiry still needs qualification, a conversation and an agreed scope before it can become a customer. A store selling a physical product has a different path. Our click cost cannot supply the expected purchase cost for either business.

We pulled the figures on 20 October 2026. The requested period ran from 14 October at 00:00 to 20 October at 00:00 in Europe/Istanbul time; 20 October is excluded. The table is a dated reporting snapshot, and attribution or spend can update later. It is not a claim that the campaign had finished.

What Ads Manager reported

All monetary amounts are in USD.

Metric	Reported value	What it describes
Campaign spend limit	$100	The total media budget ceiling we set.
Spend	$99.81	Reported media spend in the selected period; preparation and management time are additional business costs.
Impressions	5,107	Ad displays, rather than a count of unique people.
Clicks	106	Recorded ad clicks; this does not mean 106 different visitors or website sessions.
CTR	2.08%	The campaign-level click-through rate in the report.
Average CPC	$0.94	The reported average for this campaign and period.
Attributed inquiry-form event	1	One click-through conversion and zero view-through conversions.

The conversion report used a 30-day click window and a one-day view window, with conversions assigned to the ad-interaction date. The media and conversion reports cover the same campaign and requested period.

OpenAI’s reporting guide explains the available reporting windows and the different update times for metrics. Saving those settings beside a report makes a later comparison more useful than keeping the totals alone.

The setup changed during the test

On 17 October, we corrected how the browser Pixel retained the click reference on a return visit and added server-side event forwarding through the Conversions API. We also made the phone field mandatory on the landing-page form.

The full period therefore combines different measurement and form conditions. A mandatory field can affect the experience of applying, while an instrumentation change can affect what gets attributed. These aggregate figures cannot isolate either effect.

We also used campaign tags on the links and kept website form-success records separate from platform attribution. Consent, identity and reporting definitions can make clicks, website visits and attributed events differ. Test submissions were kept apart from genuine inquiries.

How we read the single conversion

The campaign objective was clicks; we separately tracked the inquiry-form action. The conversion count says the platform attributed that tracked event to an eligible ad interaction. It does not establish whether the inquiry suited our service, a meeting took place or a payment arrived.

We have not established a person-level match between the attributed event and a particular inquiry, so we do not credit a named prospect to the ad. No applicant identity or contact detail is included here.

OpenAI’s conversion measurement documentation distinguishes receiving an event, matching it to the campaign’s event definition and attributing it to an eligible interaction. The commercial outcome needs its own follow-up record.

What we will use the record for

Our next review will follow the inquiry beyond the form: was it suitable, did it progress to a conversation, was a proposal agreed and was payment received? We will keep changes to the offer, form and measurement setup beside those outcomes.

The two ads were not a controlled A/B experiment, and this single campaign does not establish a reliable acquisition cost or an industry benchmark. We also have no comparable Google or Meta experiment here from which to claim a better channel.

If you are planning a small first test, write down the commercial outcome you need and the conditions you will hold constant. Our reporting guide shows how to keep the settings with the numbers; the lead value calculator helps separate an inquiry from the value of a paying customer.