ChatGPT is now a CPC ad platform, test budget before CPCs normalise

ChatGPT is now a CPC ad platform. Conversion tracking included. Recommended opening bids of $3 to $5 per click.

Most practitioners are filing this under "interesting, not ready." That is probably a mistake.

The Google/Meta duopoly has been described as impenetrable for a decade. What makes ChatGPT different is not the format or the targeting. It is the position in the decision journey.

Someone searching Google is looking for options. Someone asking ChatGPT has usually narrowed down the question. They are past awareness, already evaluating. That is a different quality of intent, and it shows up in what they type. A query like "best PPC agency in London under £5k retainer" is not a research query, it is a shortlist query.

The unknown that decides everything

The critical question is whether the ad placement disrupts the intent or complements it. If the model surfaces an ad that answers what the user was going to ask anyway, conversion rates will surprise people. If it feels like an interruption between a question and its answer, it will perform like an interruption. We will not know which until accounts have a few months of data.

Early bids of $3 to $5 suggest OpenAI is pricing conservatively to build volume. That window will not stay open. The first cohort of advertisers in any new auction get the best CPCs and the most data to learn from. The second cohort pays more for less signal.

What to actually do

The right move is not to redirect significant budget. It is to allocate a test envelope before the channel matures.

• Pick one product or service line with a defined ICP and a strong landing page. Do not test ChatGPT ads against your whole catalogue, the first lessons get drowned in noise.
• Allocate a small test budget, $2,000 to $5,000, with no expectation of profitable ROAS in month one. Treat it as a learning fee.
• Set up conversion tracking properly from the start. ChatGPT's reporting will be thinner than Google's. Your own GA4 / Conversion API setup needs to be the source of truth.
• Document the creative patterns. Headlines that read like answers to a question outperform headlines that read like brand statements. The same instinct that works on Reddit works here.
• Watch for query data. Even early reporting will tell you which questions are converting and which are not. That intelligence informs your Google Ads and SEO work even if ChatGPT itself never scales for you.

The bigger picture

The practitioners who learn the creative format and audience behaviour early will have a head start that money cannot buy later. That has been the pattern on every new ad platform since Google Search in 2002. The window between "interesting" and "competitive" is usually six to twelve months. It is open now.

Curiosity is not a media strategy. But neither is waiting until the platform is crowded to start learning. The cost of being early is a small test budget. The cost of being late is paying premium CPCs without the institutional knowledge of how the format actually works.

If you want a view on whether a test envelope on ChatGPT makes sense for your specific account, book a free audit and we will model the opportunity against your current paid-search spend.