The Most Dangerous Number in Performance Marketing

More than a decade ago, we took on a sizable ecommerce company.

During discovery, the founder walked us through the business. We discussed margins, growth goals, advertising budgets and the ROAS target he wanted us to maintain.

He’d built a successful company.

We assumed he knew his number.

So we did what we’d been hired to do.

We rebuilt the account around those economics, increased spend and materially accelerated revenue.

And it worked.

Revenue climbed. We hit the ROAS target. Then we hit the revenue targets.

High fives all around.

About three months later, the company hired a new CFO.

She was sharp.

She started digging into the backend financials, looking at the actual economics of the revenue we were generating.

Then she delivered the bad news.

The advertising was hitting its ROAS target.

The business was barely breaking even.

She understandably wanted to know why we hadn’t caught it.

The answer was fairly straightforward. We didn’t have access to the financial information she was using. We’d repeatedly been given the ROAS target by the founder, both verbally and in writing, and instructed to scale aggressively as long as we maintained it.

That’s exactly what we’d done.

Technically, we hadn’t missed the target.

The target was wrong.

That experience changed the way I’ve looked at ROAS ever since.

Today, when a founder or CMO tells me their target is 4x, the first thing I want to know is:

How the hell did you get there?

Who decided 4x was the right number?

It’s a question we’ve asked hundreds of times since.

And one of my favorites:

When was the last time somebody recalculated it?

You’d be surprised how often the answer eventually boils down to some version of:

That’s just the ROAS we’ve always targeted.

That’s not an economic model. That’s institutional memory.

Yet entire customer acquisition programs are frequently built around numbers that nobody has challenged in years.

And sometimes that 4x isn’t protecting the business.

Sometimes it’s preventing the business from growing.

We once had a client targeting a 23x ROAS

You read that correctly.

23x.

When we began working together in 2014, they were spending roughly $10,000 a month with us, primarily in highly efficient Shopping campaigns with significant branded demand.

And they were crushing it.

At least according to ROAS.

The problem was that we were largely harvesting demand that already existed.

If we wanted to create meaningful growth, we needed to move beyond the bottom of the funnel.

And that meant allowing ROAS to come down.

It took time.

A lot of time.

We worked through several CMOs over the years, continually testing farther up the funnel and demonstrating that a lower ROAS didn’t necessarily mean worse performance.

Quite the opposite.

As we incorporated new-customer acquisition and lifetime value into the equation, something became increasingly obvious:

A lower ROAS could put substantially more gross-margin dollars in the bank.

Eventually the conversation changed.

Instead of protecting an extraordinarily high ROAS, the client became comfortable giving us essentially unlimited budget as long as we operated within the economics we’d established together.

That unlocked explosive new-customer acquisition and years of revenue growth.

The 23x wasn’t wrong.

It was answering the wrong question.

Then there’s the problem of whose ROAS you’re looking at

This gets even more complicated today because every platform has its own version of reality.

Meanwhile, the customer only gave the business their credit card once.

We recently analyzed a 30-day period for an ecommerce client where Meta reported a 6.19x ROAS.

So which number was lying?

Potentially none of them.

They were simply assigning credit differently.

Change the attribution window and the number changes again.

That’s why I don’t believe the lesson is to distrust ROAS.

The lesson is to understand exactly what your ROAS is measuring before you make a business decision with it.

Because a 6x you understand can be useful. A 6x you blindly believe can be dangerous.

Here’s where it gets really uncomfortable

With that same client, we materially increased Meta spend.

Meta’s reported ROAS remained healthy.

If we’d looked only at the advertising dashboard, there was plenty to celebrate.

But total Shopify revenue wasn’t increasing with the additional spend.

In fact, it declined.

That shouldn’t happen.

So we kept digging.

What we ultimately found wasn’t simply an advertising problem.

Inventory was constraining the business. Hero products were selling out. A substantial portion of the client’s own revenue plan depended on products that weren’t available to sell.

At the same time, site conversion had deteriorated significantly.

The platform could continue finding interactions it was willing to take credit for.

But the business wasn’t getting bigger.

That’s the distinction that matters.

When platform ROAS holds while blended business performance deteriorates during a scale-up, there’s a good chance your incremental advertising dollars aren’t buying incremental customers.

They may simply be buying more attribution credit.

ROAS is not a business model

I want to be clear about something.

ROAS matters.

We use it every day.

It is an incredibly useful diagnostic for understanding advertising performance.

But it was never designed to tell you everything you need to know about the health of your customer acquisition system.

For that, we need to ask bigger questions.

And perhaps most importantly:

If we spend another dollar, does the business actually get bigger?

That’s the question the advertising dashboard can’t answer by itself.

Fiduciary responsibility cuts both ways

We’ve talked before about the responsibility we believe an agency has when managing someone else’s money.

Sometimes that responsibility means telling a client:

Stop spending.

The platform may say everything is working beautifully, but the underlying economics tell us those incremental dollars aren’t creating profitable growth.

That’s not always an easy conversation.

But there’s another conversation that’s equally important.

Sometimes we have to tell the client:

Spend more.

If a company is protecting a 4x ROAS when its actual economics demonstrate it could sustainably operate at 3x, maintaining that 4x may mean leaving profitable revenue, and new customers, on the table.

In that situation, our responsibility isn’t to protect the prettiest number on the dashboard.

It’s to challenge it.

That doesn’t mean recklessly changing the target overnight.

And determine whether the business continues to make more money as advertising efficiency comes down.

If it does, something important happens.

The lower ROAS stops being something management has to believe in.

The money in the bank validates it.

And while you’re protecting that beautiful 4x, remember that a competitor may be perfectly happy acquiring customers profitably at 3x.

You’re protecting a ratio.

Extreme efficiency has an opportunity cost.

Ask a better question

After more than 20 years doing this, I’ve become far less interested when somebody tells me:

“Our ROAS is 4x.”

My response is:

How the hell did you get there?

Then we can decide whether 4x is protecting the company, or holding it back.

Maybe the better question for the Monday morning meeting isn’t:

“What’s our ROAS?”

Maybe it’s:

“How much more can we responsibly spend?”