Why every company needs a data strategy

Technology · 2026-04-21 · 8 min read

By Arjun Mehta, Engineering & product

Collecting data is easy. Agreeing what a customer is turns out to be the hard, expensive, political part.

Most companies do not have a data problem, they have a definitions problem. Sales counts an active customer one way, finance another, and product a third, so three dashboards disagree and every meeting relitigates the numbers instead of using them.

A strategy starts by naming a dozen metrics that matter, writing one definition for each, and giving each a single owner who can change it. Everything downstream inherits from that document. It is dull work and it removes more friction than any tool purchase.

Then centralise storage before you centralise dashboards. Raw events in one warehouse, transformations in version control, dashboards built on those transformations. The alternative — every team wiring its own extract — produces numbers that drift apart quietly.

Access should be broad and logged rather than narrow and negotiated. Data locked behind a request queue gets replaced by private spreadsheets, which is the worst outcome for both governance and accuracy.

Finally, judge the whole thing by decisions changed, not reports produced. If no dashboard has altered a plan in six months, the pipeline is a cost centre with good visual design.

Tags: data, analytics, business

Arjun Mehta — Arjun writes about web performance, developer tooling and applied AI for ESPYCRUX. He has spent the last decade building and maintaining production web apps, and most of his articles start from something that broke in one of them.