Notebook, pen and coffee mug on a kitchen table in morning light, Pittsburgh rooftops beyond the window

ChatGPT Ads Are Coming: Wait, Watch or Test Before You Spend

In short:

  • OpenAI is testing a visual ad format in ChatGPT with a first group of US advertisers, so for most small businesses it’s something to watch, not something to buy yet.
  • The test is narrow: it shows up during image generation for Free and Go users, and the measurement tools are still being added.
  • Before you spend anything, know what a lead costs you today and have your own way to track calls, forms and bookings, because that’s the only fair way to judge a new ad channel.

Over the past month, a few business owners have asked me some version of the same question: “Should I be advertising in ChatGPT?” It’s a fair question. A lot of people now go to AI tools to ask for recommendations, and ads usually follow wherever people’s attention goes.

For most Pittsburgh small businesses, my short answer right now is: not yet, but get ready. The longer answer depends on what’s actually available, what kind of business you run and whether you can measure the results. So here’s how I’m thinking about it, and the simple “wait, watch or test” decision I’d walk through with you before you spend a dollar.

What OpenAI actually announced

In early October, OpenAI published a post called Building advertising for the way people use AI. Here’s what it says:

  • ChatGPT reaches 1.2 billion people each week.
  • OpenAI is introducing a visual ad format. It’s being tested first during image generation, for Free and Go users in the US, with an initial group of advertisers.
  • It’s expanding measurement integrations. Those are the tools that connect ad activity to what happens afterward.
  • OpenAI says ads do not influence ChatGPT’s answers.
  • Businesses can sign up to advertise through a page linked from that post.

That’s what I’m building on here. When a platform describes its own product in its own words, I can plan around it. Everything else I’m hearing is talk until the company says it.

What’s live and what’s still a test

This is the part that gets lost when headlines move fast. Here’s how I’d sort the announcement:

  • Visual ads: a test. It’s limited to the US, to Free and Go users, to the moment someone is generating an image and to a first group of advertisers. That’s a narrow slice. It isn’t a full ad platform you can log into and run a campaign on next week.
  • Measurement tools: still being built out. It’s good that OpenAI is adding them. It also tells me the reporting side isn’t finished yet.
  • The answers themselves: not for sale. According to OpenAI, buying an ad doesn’t change what ChatGPT says in its answers. So an ad here is an ad, not a way to become the recommendation.
  • Sign-up: open. I’d treat signing up as raising your hand, not as a campaign plan.

A word on the big number. 1.2 billion people a week sounds huge, and it is. But a Pittsburgh plumber doesn’t need 1.2 billion people. They need the few hundred people in their service area who need a plumber this month. Until I know whether a local business can reach those people specifically, the reach number doesn’t tell me much about your budget.

When I’m deciding where a small business should put its money, I want channels that are live and have reporting I understand. Right now, ChatGPT ads are still a test.

Wait: who I’d tell to hold off

I’d tell you to wait if any of these sound like you:

  • You don’t know what a lead costs you today. If you can’t roughly say what you spend to get one phone call or booking from Google, Facebook or word of mouth, you won’t be able to tell whether a new channel is better or worse.
  • Your website isn’t ready to catch people. An ad sends someone to a page. If that page is slow, confusing or missing a clear way to call or book, your ad money leaks out the bottom.
  • Your ad budget is small and already working. If a modest Google or social budget brings in steady work, you don’t need to move part of it to an unproven channel this quarter.
  • Nobody has time to check results every week. New channels need attention. If no one is going to look at the numbers, wait.

Waiting doesn’t mean you’re falling behind. You’re letting bigger advertisers pay for the early learning while you get your own house in order.

Watch: what I’d keep an eye on

“Watch” means you’re not spending yet, but you’re paying attention on purpose. Here’s what I’m watching:

  • When ads open beyond the first group. The questions I’d want answered: Can a local business target by area? Is there a minimum spend? Do ads show up anywhere other than image generation?
  • What the measurement tools actually report. I want to see whether they can show calls, form fills and bookings, not just views and clicks.
  • Your own customers. Start asking new customers whether they used ChatGPT or another AI tool while they were looking for someone like you. If nobody says yes after a couple of months, that’s useful to know. If several do, that’s useful too.
  • Your landing pages. Use the waiting time to make sure each main service has a page with one clear next step. Any ad you run later, on any platform, will need one.

Here’s an easy way to stay on top of it: put a 15-minute reminder on your calendar once a month to check for updates. That’s plenty.

Test: who might be a fit, and how I’d set it up

Some businesses will reasonably want to be early. The first format shows up while people are generating images, so my own read is that businesses with visual products or services, like home decor, design, food or events, may fit more naturally than, say, a bookkeeping firm. That’s my opinion, not something OpenAI has said.

If you decide to test ChatGPT ads once you can get in, or any new AI ad placement, here’s how I’d set it up:

  1. Set a fixed budget and an end date. Pick an amount you’re comfortable treating as the cost of learning, and a date when you’ll stop and look.
  2. Write down your baseline first. What does a lead cost you from your current channels? The test has to beat that number, or at least match it.
  3. Use one offer and one landing page. Don’t send people to your homepage. Send them to a page built for that ad, with one clear next step.
  4. Tag your links. Add tracking tags (UTM parameters) to the ad link so those visits show up on their own in your website analytics.
  5. Ask people how they found you. Ask on your contact form and when you answer the phone. It’s low-tech, and it catches what software misses.
  6. Count leads, not impressions. Views and clicks are easy to report. Calls, booked appointments and real inquiries are what matter.

What to measure first

If you only track three things, make them these:

  • Cost per lead. Total spend divided by real inquiries. Compare it to your baseline.
  • Lead quality. Were these people actually looking for what you offer, in the area you serve? Ten wrong-fit inquiries are worth less than two good ones.
  • What happened next. Did any of them book, buy or come back? Even a rough tally in a spreadsheet tells you more than a dashboard full of clicks.

A platform’s report shows how the platform thinks its ads did. I always check it against the business’s own records: the phone log, the inbox, the calendar. When the two disagree, I trust the business’s records.

My quick decision guide

Here’s the short version to save or share with your team:

  • Wait if you don’t know your current cost per lead, your site isn’t ready, or nobody has time to review results.
  • Watch if your marketing is working and you want to be ready when these ads open up more widely. Check monthly, ask customers how they found you and tighten up your landing pages.
  • Test if you have a clear baseline, a page built to convert, a budget you can afford to learn with and your own way to track real leads.

In my experience, getting in first matters a lot less than knowing your numbers before you start. That part you can work on this week, no matter what OpenAI announces next.

If you’d like a second set of eyes on whether your business should wait, watch or test, or help getting your tracking and landing pages ready first, I’m happy to talk it through. Let’s connect.

Keep reading: Connecting Brand, Technology and Franchise Growth

Tags :

Facebook
Twitter
LinkedIn