Rupon Anandanadarajah: SaaS Metrics Are Driving Decisions, but Not Always the Right Ones

SaaS metrics were supposed to make everything clearer. And in a lot of ways, they did. But Rupon Anandanadarajah has spent enough time inside high-growth product teams to know that the data telling you things are going well and the reality of things going well are not always the same thing.

His argument isn’t that SaaS metrics are useless. It’s that they’re being used wrong — and that distinction matters more than most companies want to admit.

The Proxy Problem

Here’s the thing: most SaaS metrics don’t measure outcomes. They measure behaviour that’s supposed to indicate outcomes.

Activation rates suggest early value. Retention signals ongoing usefulness. Conversion implies product-market fit. But Rupon Anandanadarajah points out something teams rarely stop to question — are these proxies still accurate? Does a rising activation rate actually mean users are getting value, or have they just clicked through a shorter onboarding flow?

Sometimes it’s the latter. Strip onboarding down too aggressively and activation climbs while long-term engagement quietly falls apart. Users completed the process; they just didn’t understand the product.

The metric improved. The business didn’t.

Why Teams Keep Doing It Anyway

This isn’t stupidity. It’s structure.

SaaS companies run on quarterly pressure. Investors want updates. Executives want traction signals. Product teams get evaluated on measurable output. So naturally, teams gravitate toward initiatives that will visibly move numbers — fast.

Anandanadarajah sees this as a feedback loop that builds slowly and then compounds. Over time, decision-making tilts toward optimisation over understanding. And optimisation without understanding is just a very efficient way of running in the wrong direction.

The catch? It can look like progress for a long time before it doesn’t.

When Teams Optimise in Silos

Fragmentation makes this worse. Marketing chases lead volume. Product targets activation. Customer success owns retention. Each team hits its numbers. The business as a whole — not so much.

Anandanadarajah describes this as local optimisation at the expense of global outcomes. It’s not that anyone’s doing a bad job; it’s that no one has agreed on what success actually looks like at the system level. Improvements in one area can quietly introduce friction somewhere else. That friction rarely shows up in anyone’s dashboard.

What the Numbers Can’t Tell You

Behavioural data shows what happened. It won’t tell you why.

That gap is where decisions go wrong. Without qualitative context — customer interviews, support conversations, watching someone actually use the product — teams are reading signals without the story behind them. A drop in feature usage might mean the feature is bad, or it might mean users don’t know it exists, or it might mean a UI change buried it three releases ago.

Same metric. Completely different problem.

The Fix Isn’t Fewer Metrics

Anandanadarajah isn’t calling for less measurement. He’s calling for better interpretation.

SaaS metrics should start conversations, not end them. When a number moves, the productive question isn’t “did it go the right way?” — it’s “what behaviour caused that, and is that behaviour actually what we want?” That’s a slower process. But it’s the one that produces insight you can build on.

He also pushes companies to define success around customer outcomes rather than product activity. Did users complete the workflows that matter? Are they measurably better off than before? Those answers are harder to get — and much harder to game.

Where This Is Heading

As SaaS markets tighten and growth gets harder, shallow optimisation loses its cover. Companies can’t paper over weak retention with acquisition volume forever. At some point, the product has to actually work for people.

Anandanadarajah thinks the shift toward more disciplined use of SaaS metrics is already in motion — not because companies suddenly became more thoughtful, but because the old approach stopped working.

That’s usually how change happens.

The companies that get ahead of it — the ones that pair data with judgment, align teams around shared definitions of value, and treat metrics as questions rather than answers — those are the ones who’ll have a real edge. Not just a good-looking dashboard.