TL;DR: I spent just over NZ$100 testing ChatGPT advertising during the launch of Get Your Business Sorted. Over four days the campaign generated 3,363 impressions, 27 clicks and two completed Business Snapshots, while teaching me some valuable lessons about landing pages, website performance and tracking. Although I didn’t gain a new client immediately, I would absolutely run another campaign, and this time I expect much stronger results.

When a new advertising platform appears, the temptation is to wait until everyone else has worked out how to use it.

I’ve never been very good at waiting.

When OpenAI launched advertising through ChatGPT, I decided to give it a go. Get Your Business Sorted had just launched, so it seemed like the perfect opportunity to test whether ChatGPT users would respond differently from people on Google or Facebook.

My budget was deliberately small. I wanted to learn before spending serious money.

After four days and just over NZ$100 spent, I came away with some useful lessons.

Some were exactly what I expected.  Some surprised me.

Why I decided to try ChatGPT advertising

For years, most businesses have thought about online advertising in terms of Google or Facebook.

That is changing.

More and more people are asking ChatGPT for advice before they search Google. If someone is looking for help with growing their business, solving a problem or choosing a service provider, there is a good chance AI will become part of that process.

As someone who works with websites, automation and AI, I felt I should understand the platform first-hand rather than simply read about it.

The campaign

The campaign ran from 1 July to 4 July, alongside the launch of Get Your Business Sorted.

I limited the campaign to approximately NZ$100, although it eventually spent around NZ$110, which was close enough for my purposes.

I also set a maximum bid of $5 per click, expecting that would produce around five clicks per day.

This was never intended to generate hundreds of leads. It was a learning exercise.

The advertisement itself was deliberately simple.

• Clear headline

• Clean branding

• One obvious call to action

• A free five-minute Business Assessment

The first lesson arrived almost immediately

Within the first couple of days, it became obvious that I had two problems.

The first was technical.  I had done a lot of publicity for Get Your Business Sorted, and the hosting wasn’t coping.

The extra visitors exposed the fact that my website hosting couldn’t keep up.  I had been intending to switch to better hosting, and this was the time. Fortunately, moving the site to faster hosting solved that issue quickly.

The second problem was more interesting.

Almost nobody was completing my original 60-question Business Assessment.

Looking back, that shouldn’t have been a surprise. Asking someone who has only just discovered your business to answer 60 questions is asking for a fairly big commitment.

So I built a new Business Snapshot with just 18 questions.

That single change immediately improved completion rates.

If I’d made that change before launching the campaign, I suspect the results would have been stronger.

The numbers

After four days, the campaign had produced:

• 3,363 impressions

• 27 clicks

• Approximately NZ$110 spent

• 2 completed Business Snapshot assessments

At first glance, two completed forms might not sound exciting.

However, advertising should always be judged against its purpose.

My purpose wasn’t simply collecting forms.  It was learning how this new platform behaves.

What happened with the leads?

Both assessment submissions came from New Zealand businesses, which was my target audience.

One person requested a follow-up discussion, so I contacted them and sent my Calendly booking link. Although they haven’t booked yet, I have enough information to include them in my follow-up process.

The second person chose not to request contact. I couldn’t identify enough information about their business to justify pursuing them directly, although they will continue receiving my regular business emails.

Neither lead has become a client yet.

That happens.

Marketing is rarely about immediate sales. Most businesses need several interactions before they decide to engage.

The hidden result

One outcome doesn’t appear in the advertising statistics.

Website traffic increased significantly during the campaign.  That matters because every visitor becomes another opportunity for future marketing.

Some people will come back weeks later.  Some will see future articles.  Some will encounter my business again through search engines or social media.

Marketing rarely works in isolation. Each activity builds on the previous one.

What I changed afterwards

Once the campaign finished, I turned my attention to measurement.

Google and Meta have offered tracking pixels for years. OpenAI now provides its own measurement pixel as well, and I installed it on my Gecko Gully website.

The installation was surprisingly straightforward.

That means future campaigns can provide much better information about what visitors actually do after clicking an advertisement.

Without measurement, you’re making educated guesses.  With measurement, you can improve every campaign.

Would I do it again?

Absolutely.

Not because the first campaign produced spectacular results.  Because it taught me how the platform works.  Every new advertising channel has a learning curve.

The businesses that usually benefit most are the ones willing to learn while everyone else is still watching from the sidelines.

I also believe something much bigger is happening.

People who are actively looking for advice, guidance and solutions are increasingly turning to AI before they visit Google.  That makes ChatGPT an interesting place to advertise, particularly for businesses offering professional services.

We’re still in the early days.  That’s often where the best opportunities are.

What’s next?

My immediate plans are fairly straightforward.

The OpenAI tracking pixel is now collecting data, which means future campaigns should become increasingly effective.

I’m also considering running Facebook advertising again, using my much larger custom audience of previous website visitors. Because the OpenAI pixel is already installed, it will also record visitors who arrive through those Facebook campaigns, giving me a much richer picture of how people move between platforms.

Rather than treating Google, Facebook and ChatGPT as separate marketing channels, I think businesses will increasingly need to see them as parts of one connected marketing system.

That is where I expect the biggest gains to come from over the next few years.

Final thoughts

Would I recommend ChatGPT advertising today?

Yes, with realistic expectations.  Treat your first campaign as research rather than expecting instant sales.  Make sure your website is fast.  Keep your landing page simple.  Ask for only the information you genuinely need.

Most importantly, measure everything you can.

Those lessons apply whether you’re spending $100 or $10,000.