Most companies we meet are not short on data. They are short on results from it. 

A story you will probably recognize

Take a regional retail chain we can all picture. Sales are flat, margins are under pressure, and the board wants answers. So the company invests. It stands up a cloud data warehouse, connects every till and every online order, and hires three analysts. Within a few months there are more than forty dashboards, and every store manager receives a daily email full of charts.

And still nothing moves. The reason is simple. Nobody ever asked which decision the dashboards were meant to improve. The store managers already knew sales were soft. What they needed was to know, which products should they stock less of. Which promotions actually paid for themselves. Which three stores to fix first. None of that lived on the forty dashboards, because the project began with technology instead of a question.

Now compare that with a medium sized manufacturer that did it the other way round. Before anyone mentioned tools, the leadership team named the decision that hurt them most. They were quoting delivery dates to customers and missing them, which cost them repeat orders. So the question was narrow and clear. How do we know, at the moment we quote, whether we can actually deliver on time?

That one decision shaped everything that followed. The data team pulled only the numbers that mattered: machine capacity, supplier lead times, and the order book. They built one model and one simple view, not forty dashboards. Within a quarter the sales team was quoting dates it could keep, on time delivery improved, and the finance chief could point to fewer penalty charges and more repeat business. Same kind of technology as the retailer. A completely different result.

We see the mirror image in software companies too. One client had a beautiful churn dashboard that refreshed every hour. Impressive to look at, useless in practice, because no one owned the decision it was supposed to feed. Once the customer success leader agreed that the number would trigger a specific outreach routine every week, the very same data started saving accounts. Nothing changed in the technology. What changed was that a human decision was finally attached to it.

The problem is almost never the technology

Notice what all of these stories have in common. In every case the technology worked. The warehouse was solid, the dashboards were accurate, the model was fine. The failure, when it came, was one of sequence. The companies that struggle start with the platform and hope the value will appear. The companies that succeed start with the decision and let the platform serve it.

Start with the business, every time

So what does starting with the business actually look like? In our experience it comes down to being honest about three things before a single tool is chosen:

- The first is the goal. What are we truly trying to make better, faster, or cheaper this year? Not in the abstract, but in terms the board already cares about.

- The second is the decision. Which choices do we make again and again that better information would genuinely improve? Pricing, stock levels, credit terms, where to send the sales team next week.

- The third is the people. Are the managers who will actually use the insight in the room while it is being designed? If they are not, you are building a report that will be admired once and then ignored.

Only when those three are clear does it make sense to talk about architecture, metrics, and the technology stack. Put plainly, your data strategy is not a separate document that lives in a folder no one opens. It is part of your business strategy, or it is nothing.

What changes when the business leads and data follows

When the order is right, three things change. Priorities get clearer, because the team knows what matters and stops pouring effort into what does not. The work starts to bite, because every insight is tied to an action rather than to a reporting habit. And agreement across the business becomes far easier, because the data now supports goals the leadership already owns rather than competing with them.

No dashboards gathering dust. No analyses that nobody reads. Just a steady supply of decisions made with more confidence and less guesswork.

A simple test before you spend anything

If you take one thing from this, make it this test. Before you approve any data initiative, ask a single question. Which decision will this improve? If you and your team cannot name that decision in one sentence, you are not ready to build yet.

And that is good news, because it means the most valuable work costs almost nothing. It is not a giant strategy programme. It is a handful of honest conversations between the business and the data people, held early enough to actually steer the direction.

Get started

This is the work we do at Avito. We help leadership teams do things in the right order, so that Data Management and Data Governance stop being technical projects off to the side and start behaving like what they really are: a reliable way to make better decisions, faster, with numbers you can trust.

If your dashboards are busy but your decisions are not getting any easier, that is usually the place to start. We would be glad to have that first conversation.

 Get in touch for an informal conversation