Ever been in a meeting where someone announces "ROAS is up 15%!" and everyone nods approvingly? I get it. ROAS feels good. It's clean, measurable, and defensible. One number that says "we're doing our job,” but here's the thing: you can have perfect metrics and still be making terrible decisions.

When good numbers lead you astray

Marketing teams don't usually fail because they're lazy or incompetent. They fail because they're optimizing for the wrong thing and getting rewarded for it.

Every KPI makes an implicit promise: "If this goes up, the business wins." But that only works if:

• The metric actually reflects what you care about

• You can't game it in ways that hurt the business

• Everyone agrees on what success looks like right now

Miss any of those, and you end up with impressive dashboards while revenue flatlines. Then everyone acts shocked.

ROAS works great for the wrong goal

Don't get me wrong, ROAS has its place. When your goal is efficiency, better margins, less waste, and smarter spending, it's genuinely useful.

The problem starts when your goal shifts to growth, but ROAS stays your North Star. Growth and efficiency often pull in opposite directions. Pretending they're the same thing is how you optimize yourself into irrelevance.

How to accidentally kill growth while your metrics look great

Here's the pattern I see all the time: ROAS starts dropping, so teams make the "smart" moves. Cut experimental spend. Narrow targeting to people already interested. Double down on retargeting and branded search. Pull back from prospecting because it "doesn't perform."

ROAS goes up. Everyone celebrates. And six months later, new customer acquisition has fallen off a cliff.

What happened? You stopped creating demand and just started harvesting it. You became really efficient at capturing people who were already going to buy, while the pipeline of future customers dried up.

One metric can't do three jobs

Most businesses need marketing to juggle three things at once:

• Close deals efficiently (convert people ready to buy)

• Build the future (create awareness and preference with people who aren't ready yet)

• Improve economics (boost lifetime value, reduce churn)

ROAS tells you how well you're doing the first one. It's basically blind to the second. Which is why ROAS-obsessed companies gradually become harvest machines.

The test: what does this metric make you do?

Forget the theory for a second. Ask yourself: When ROAS drops, what do we actually do? Cut spend? Kill prospecting? Narrow targeting? Reward the team for playing it safe?

If yes, congrats, your KPI is now writing your strategy, and probably not in a good way

The best measurement systems work like this:

Top tier (what actually matters):

• New customer revenue

• Pipeline growth

• Total revenue

Second tier (guardrails):

• ROAS for different channels

• CAC (split by new vs. existing)

• Payback period

• LTV/CAC

Then follow one rule: never celebrate efficiency metrics alone. If ROAS is up but new customer volume is down, you didn't win. You just got more conservative.

How we actually help clients with this

At Direct Agents, a huge part of what we do is helping teams stop managing dashboards and start managing the actual business.

We get everyone aligned on the real goal first. Are we growing? Optimizing for profit? Buying market share? You can't pick the right KPIs until you answer that.

We build measurement hierarchies so teams know the difference between "the outcome" and "a diagnostic metric."

We separate how we measure demand capture (bottom-funnel, high-intent) from demand creation (top-funnel, future buyers), so prospecting doesn't get killed for looking messy in attribution reports.

We reality-check performance with incrementality thinking, so "amazing ROAS" doesn't hide the fact that you're just stealing from your own organic traffic.

ROAS is a tool, not a strategy. The real question isn't whether your metrics look good it's whether they're connected to what actually matters for your business.

If your KPIs let you "win" while growth stalls, you've got a measurement problem, not a marketing one.

Jackson Richards, VP of Strategy