How Much Should You Spend on a New Ad Network Test?

One of the questions I get fairly often is:

How much should I spend testing a new advertising network?

The answer is frustratingly similar to my answer for, "How much should I spend on paid overall?"

It depends.

More specifically, it depends on what you're trying to learn.

I don't think budget is actually the first question you should ask.

Starting with budget is one of the fastest ways to limit your opportunity. You either underfund a test that never had a realistic chance of producing meaningful results, or you let an arbitrary budget dictate the size of the opportunity you're willing to pursue.

The reality is that if you want your business to continue growing, you have to test. Every business eventually reaches a point where simply optimizing existing channels produces diminishing returns. Sustainable growth requires finding new audiences, testing new ideas, and occasionally exploring entirely new channels.

Of course, testing comes with risk.

So does not testing.

The businesses that never experiment often find themselves falling behind competitors that continue to test and learn.

That's why, before deciding on a budget, I think there are three more important questions to answer.

1. What are you actually trying to learn?

Before you decide how much you're willing to spend on a test, you should know what you're hoping to get out of it. After all, your budget should be driven by the value of the answer to your question.

The dream outcome is obvious. You discover a channel full of your ideal customers, they're converting below your CAC target, and you've found your next growth engine.

Ten or fifteen years ago, that happened more often.

Today, it rarely does.

The paid media ecosystem is much more mature, which means most tests aren't about striking gold. They're about reducing uncertainty and making better decisions.

Maybe you're measuring incrementality. Maybe you're trying to understand how a new audience behaves. Maybe you're evaluating where another channel fits within your broader marketing mix.

Those may not sound as exciting as discovering your next billion-dollar channel, but they're incredibly valuable. Every good test gives you information you can use to make better decisions going forward.

Every once in a while, though, something genuinely new comes along.

That's how I think about ChatGPT Ads.

The objective isn't necessarily to prove it can become a major acquisition channel today. It's to understand the platform while it's still evolving.

How does it work? What targeting capabilities exist today? How do visitors behave after they reach your website? What would need to improve before making a larger investment?

For the first time in probably a decade, I've caught myself thinking:

There might actually be oil in those hills.

Maybe there isn't.

But there are enough encouraging early signals that I think it's worth finding out.

2. How much can you afford to lose?

Once you know what you're trying to learn, the next question becomes:

How much uncertainty are you comfortable taking on to answer it?

I don't mean: "How much money are you willing to light on fire?"

I mean: What's the largest bet you can comfortably make if this test doesn't produce the outcome you hoped for?

Testing is inherently uncertain.

If you only invest in things you already know will work, you'll never discover your next growth opportunity. At the same time, betting the farm on an unproven channel is an unnecessary risk.

The bigger risk is never experimenting at all. Businesses that stop testing eventually stop discovering new opportunities. Competitors find new audiences first, growth plateaus, and every dollar continues flowing into the same increasingly mature channels.

It's also worth recognizing that unsuccessful tests aren't necessarily failures. Every test teaches you something.

Maybe the audience wasn't right. Maybe the messaging missed the mark. Maybe one creative approach dramatically outperformed another. Or maybe you simply didn't invest enough to generate a meaningful answer (more on that later).

Those lessons shape future strategy long after the test is over.

The goal isn't to eliminate uncertainty. It's to decide how much uncertainty your business can comfortably absorb while continuing to move forward.

3. Finally, are you testing... or just trying to stay busy?

Sometimes the motivation is less strategic.

Maybe there's leftover budget that needs to be spent.

Maybe leadership wants to see innovation.

Maybe everyone feels like they should be trying something new because everyone else is.

None of those are inherently bad reasons.

But before investing in a new initiative, it's worth pausing and asking the two questions we've already discussed.

• What are we hoping to learn?

• What are we comfortable risking?

Without those answers, it's easy to run a test that never had a clear definition of success or failure. And when that happens, even a potentially promising opportunity can look unsuccessful simply because nobody agreed on what the test was supposed to accomplish in the first place.

Good testing isn't about staying busy.

It's about reducing uncertainty and making smarter decisions with every experiment.

So how much should you spend?

If you're running a serious incrementality test...

Spend enough to generate a meaningful answer.

Channels further from conversion require more time, more touchpoints, and more patience before their impact becomes measurable.

A test that's too small...

...often tells you nothing at all.

Underfunded tests are a bigger waste of money than inefficient tests that produce valuable learning.

If you aren't giving the test enough runway to answer your original question, you're unlikely to walk away with useful insights.

On the other hand...

If your goal is simply to understand an emerging platform, your budget can be much smaller.

That's how I'm thinking about ChatGPT Ads.

I'm not trying to prove it can immediately become a major acquisition channel. I'm trying to understand the environment.

• Is my audience there?

• How do they engage?

• Does this feel like a platform that's evolving in a direction worth paying attention to?

Those questions don't require a massive investment. They require enough investment to learn something meaningful.

Just don't expect a small exploratory test to uncover your next billion-dollar growth engine overnight.

The Bottom Line: Deciding on Spend for an Ad Network Test

The right testing budget isn't determined by the network. It's determined by the value of the answer you're trying to buy.

Know what you're trying to learn. Know how much uncertainty your business can comfortably absorb. Then invest enough to answer the question you're asking.

Because successful testing isn't really about proving yourself right.

It's about making better decisions over time.