Most organisations measure their data success by how much they build.
Pipelines delivered. Tables published. Dashboards created.

And yet, trust still erodes, duplication spreads, and decisions remain slow.

In this episode, Roland Brown challenges one of the most entrenched habits in modern data teams: measuring activity instead of value. Building on the foundations laid in Episodes 64, 65, and 66, he explains why traditional data metrics often create the illusion of progress and why real data product success shows up in behaviour, not dashboards.

Roland makes a clear distinction between usage and reliance. A data product can be queried frequently and still not be trusted. It can be reused widely and still be reinterpreted every time. When teams measure volume instead of confidence, failure hides behind busy charts.

The episode introduces three signals that consistently reveal whether a data product is actually succeeding:

• Adoption are the right consumers using the product to make real decisions?
• Reuse is the product reducing duplication and rework, or just feeding more downstream variations?
• Trust do consumers rely on the product without validation, disclaimers, or reconciliation?

Roland explains why adoption is often a lagging indicator and why trust-related behaviours like increased validation, shadow calculations, and side-channel confirmations are some of the earliest signs that a product is in trouble.

Drawing on the product anatomy discussed in Episode 62, he shows how success metrics change when data products are treated as long-lived capabilities instead of one-off deliveries:

• Adoption aligns to decision cadence, not query counts
• Reuse is measured by work eliminated, not consumers added
• Trust reveals itself when products are used without hesitation

A practical example demonstrates how two products with similar usage statistics can have radically different outcomes one stabilising decisions and the other quietly creating more work depending on whether trust is present.

The episode also addresses a common leadership mistake: assuming that low adoption means more training is needed. Roland explains why adoption problems are almost always design or trust problems, not education problems — and why better metrics often reveal uncomfortable truths about ownership, contracts, and intent.

The episode closes with a reframing that ties the entire data product arc together:
data products do not succeed because they are visible they succeed because they are relied on. When adoption, reuse, and trust are measured honestly, teams stop optimising for output and start optimising for confidence.

Discover insights on:

• Why activity metrics hide data product failure
• The difference between usage and reliance
• How to measure reuse without incentivising duplication
• Why trust is the most honest and quietest success signal
• How behaviour reveals value long before dashboards do
• What leaders should measure if they actually care about outcomes

“You don’t measure data products by how often they’re touched.
You measure them by how rarely they’re questioned.”

🎧 Listen to The Data Journey wherever you get your podcasts, or visit thedatajourney.com