Why the Readiness Question Comes Before the Media Plan
Most of what decides a ChatGPT Ads pilot happens before anyone opens a campaign builder. Whether your buyers sit in a live market, whether your category is allowed, whether the people you sell to are on a tier that sees ads at all, whether a click can be traced to a lead in your CRM: none of that is a campaign setting. It is either true on the day you launch or it is a gap you will discover with money on the table.
So every question below is about your company. The channel itself, the formats, the buying lanes and how we run it, is on our ChatGPT Ads management page. Here you get the instrument we use internally before we recommend a pilot to anyone, published in full, with the weights printed so you can argue with them.
One set of dates sits under every check below. As of September 2026, ads can appear for users on the Free and Go plans, including logged-out ChatGPT, in the US (since February 9, 2026), Canada, Australia and New Zealand (since April 2026), the UK, Mexico, Brazil, Japan and South Korea (all live by August 11, 2026), across 31 European countries, the EEA plus Switzerland, from August 24, 2026, and India, where OpenAI announced ads on August 27, 2026. Score against your own markets, and re-score when that list moves, because it has moved several times this year.
The 28 Checks, and What Each One Is Worth
Tick a check only when it is verifiably true today. If you are not sure, leave it unticked: an unsure answer is an unready one, and half credit here would flatter you into a pilot. The weights sum to 100 and are printed on every check. Two checks in the first section are marked as stop checks, and they behave differently: if either one is untrue, the verdict is wait no matter how high the rest of the score climbs.
Market and Eligibility 0 / 24 pts
Two facts decide whether you can advertise at all: whether your buyers are in a market where the ads run, and whether OpenAI's ad policies permit what you sell. Both are stop checks. The other four ask whether you have planned for the way access actually works right now.
If either stop check is untrue, stop here. The remaining 76 points describe a campaign you are not allowed to run yet, and the useful question becomes when the market or the policy changes.
Audience Reachability 0 / 20 pts
Ads can appear for users on the Free and Go plans, including logged-out ChatGPT, and never for accounts OpenAI identifies as belonging to users under 18. Plus, Pro, Business, Enterprise and Education are ad-free, so part of a buying committee may never see a thing you run. This section asks how much of your buying committee sits on the side where the ads appear.
A low score in this section is the one gap money cannot close. Budget buys impressions where your buyers are; it cannot move them onto a tier that shows ads. When these five come out weak, the honest read is that the spend belongs in the channels where your committee already is.
Targeting Fit 0 / 18 pts
Targeting is mostly contextual. You write natural-language context hints at ad-group level and the system matches them against the conversation. As of September 2026, OpenAI documents location, platform and custom audience targeting plus context hints; job-title, demographic, firmographic, lookalike and website retargeting audiences are not among the documented options. Custom audiences built from your own customer or prospect lists exist, but not for campaigns in the EEA or Switzerland. Six questions about whether your offer survives that.
Missing points here rarely stop a campaign. They stop you learning from one: the ads run, the clicks arrive, and nothing tells you which conversation produced them. Write the hints and sharpen the landing page first, since both are cheap and both are yours to keep.
Measurement Readiness 0 / 26 pts
This section carries the most points because it decides whether the pilot returns knowledge or anecdotes. OpenAI reports aggregated campaign metrics by design, and conversion reporting depends on the OpenAI Pixel or Conversions API you install, so everything you learn rests on instrumentation you own. Every item is work that belongs before the first impression.
Everything unticked in this section is week-one work, and it is the reason we wire measurement before a cent moves. The mechanics are in our guide to ChatGPT ads analytics: tags, pixel, server-side events, and what the aggregate reporting can and cannot show.
Budget Posture 0 / 12 pts
A pilot buys information. These five checks ask whether the money is set up to buy any, and they carry the fewest points on purpose: every one of them can be fixed in an afternoon by the people already in the room.
This is the easiest section to repair and the most often skipped. A pilot with no end date and no agreed metric does not fail, it just never finishes, and twelve months later nobody can say what it proved.
Your readiness score updates as you tick. Ask for the verdict whenever you are done.
- Market and eligibility0 / 24
- Audience reachability0 / 20
- Targeting fit0 / 18
- Measurement readiness0 / 26
- Budget posture0 / 12
What Your Score Tells You to Do Next
Three verdicts, and the stop checks sit above all of them. The bands come from the same threshold we apply internally before recommending a pilot to a client:
| Score | Verdict | What to do about it |
|---|---|---|
| 75 to 100 | Run now | Your buyers are reachable, your targeting works without identity data, and a click can be traced to a lead. Build a small pilot with a fixed window and one agreed metric. Our ChatGPT Ads management page describes the six steps we would run it through. |
| 50 to 74 | Prepare now, run later | The channel fits, the plumbing does not yet. When the missing points cluster in measurement, the repair earns its keep across every channel you run, whatever you decide about this one. Close those gaps, then re-score. |
| Below 50 | Wait | Too many pieces are missing at once for a pilot to teach you anything worth the spend. Put the money where measurement already works and revisit this in a quarter. |
| Any score | Wait | A stop check came out untrue. Market availability and category eligibility are not gaps you can work around, so the score is beside the point until one of them changes. |
Two honest limits on that verdict. A high score says you are ready to learn something from a pilot; it does not forecast what the pilot returns, and nobody can, in a channel that opened in February 2026. And readiness decays: availability, formats and policy have all moved during 2026, so a score is a photograph, not a certificate. Re-run it before the next budget conversation.
How the Weights Were Set
Measurement carries 26 points, the largest share, because it is the only section whose failures are invisible while they happen: the campaign runs, the dashboard fills, and the numbers cannot answer the one question you funded them to answer. Market and eligibility carries 24 and owns both stop checks, because those two are the only items on the list that no budget and no agency can route around. Budget posture carries 12, the smallest share, because its checks are the ones a single meeting can settle, which makes them cheap to fix and easy to forget.
What the scorecard cannot see is worth naming too. It does not know whether your creative is any good, what your competitors are willing to pay for the same conversation, or whether your sales team follows up on a lead within the hour. Those decide outcomes; this decides whether you are in a position to read them. It sits alongside the other diagnostics in our resource library, where the tracking and budget instruments live if this scorecard sends you back to them first.
What Buyers Ask About This Scorecard
Do I need an ad account or a budget to use this?
No. Everything scores in your browser, nothing you tick is transmitted or stored, and no email is asked for. What you do need is about 20 minutes and honest answers from whoever owns tracking, since the measurement section is the one nobody can answer from memory.
Why do two checks override the whole score?
Because they are the only two that cannot be worked around. If your market has not opened, or your category is not permitted under OpenAI's ad policies, then perfect tracking and a beautiful offer buy you nothing. Every other check on the list describes a gap you can close with work.
What should I do if the verdict is wait?
Wait. Nothing on this page needs an agency to fix, and a wait verdict is a real outcome here rather than a softer way of asking you to book a call. Note the checks you missed, watch the market and policy dates, and score again when they move. Our guide to advertising on ChatGPT covers what those changes look like in practice.
Is a high score a prediction of results?
No, and treat any scorecard that claims otherwise with suspicion. This one measures readiness: whether a pilot would produce a clean answer. Results depend on your offer, your market and your follow-up, and the channel is young enough that nobody holds years of evidence about any of it.
Can I use this to brief an agency or my own team?
Yes, that is the most useful thing to do with it. The unticked checks are your work list and the printed weights put it in order, while the section bars show where the gaps concentrate. Start with measurement: that part has to be finished before any spend, and postponing it never makes it cheaper.
Want a Second Read on Your Score?
Send us your five section scores plus two lines on who you sell to and where. A senior paid media specialist at COSEOM® replies with our own read of the same five sections, what we would fix first, and whether we would run a pilot at all. If the honest answer is wait, that is the answer you get.
Get a Second Read