When should data teams worry about platform costs? 💸

The way data grows and the way we price its storage and computation follow entirely different trajectories relative to business growth. Businesses don’t pay a linear fee for more insight; they pay an exponential tax. If revenue is growing by x, your data is probably doubling at 2x, and the cost to handle it might be ballooning by 3x or 4x. That’s not a pessimistic figure; it’s just the arithmetic of the modern data economy.

Every query, every event produced, every real-time dashboard, every machine learning model used to power recommendations for customers will chip away at your profitability. You’ll tell yourself that you created a “machine good at printing money”, but don’t forget that it is also good at “printing a giant electricity bill.”

And while everyone acknowledges at some level that growth in data costs outpaces business growth — throwing terms like “big data” and “data scaling challenges” into meetings — few actually pause to examine the mechanics. What causes this growth? Why do costs snowball? And, most importantly, what should we be doing about it?

Less is more

Many teams assume, incorrectly, that you can “optimize” your way out of a data cost spiral — that the right tool or vendor can flip a switch and make the problem disappear. My experience tells a different story. Unless you’re deliberate about how you manage access policies, retention policies, tiered storage, data discovery, incremental processing, and aggressive pruning of unused artifacts, no shiny tech can save you. Without a thoughtful strategy, your systems will grow bloated, inefficient, and unmanageable, regardless of how much money or tools you throw at the problem.

The counterintuitive truth is that escaping this spiral doesn’t require massive investments in new tools, vendors, or complex infrastructure. It requires something much harder to come by: discipline. Discipline to recognize that not all data is equally important, to make tough decisions about what to delete, and to ignore the temptation to collect and store everything. Less data doesn’t mean less insight — in fact, it often means the opposite. A leaner data ecosystem delivers sharper, faster, and more actionable answers by cutting through the noise.