Why the Simplest Questions Are the Hardest to Answer

Are we making money on this job? It sounds like a simple question. In most growing businesses, it isn’t.

The answer usually exists. It’s spread across several systems that were never designed to talk to each other. The accounts show one part. The job tracker shows another. Nobody can see all of it at once.

That gap ran through our webinar, The Data Odyssey, presented by Ed Hardie, Senior Software Engineer at The Curve. Most businesses don’t have a data problem. They have a joining-up problem.

Ed has seen it from both sides. He’s spent much of his career around senior leadership teams, asking for dashboards. He’s also been the one building them.

Every business generates more useful data than it realises. Even Ed’s weekend mustard stall does. He pulls reports from his till to see what sold where, then decides how much to make for the next market.

A growing business has far more going on. Each new system does its own job and collects its own data. Buy another company and you inherit a whole new set. Ed has even seen different people using different systems to do the same job.

Each system holds part of the answer. None of them holds all of it. That’s the gap a data warehouse closes. It organises your data, combines it across systems, and keeps working as the volume grows.

One margin question opened up the whole business

The clearest example came from our own work.

The client sends engineers to jobs all over Europe. With the cost of everything going up, including travel and accommodation, there was a feeling that margins were being squeezed. They wanted to know which customers were actually profitable.

The accounting system showed money in and money out. But the real cost of a job sat elsewhere: how many people went, how long it took, what equipment they needed. That lived in the job tracking system, the purchase ledger and the ticketing system.

We agreed a formula with the client for the cost of serving a customer. Working it out in a spreadsheet for one client was an ordeal. Doing it for every client, every month, wasn’t realistic.

So we brought those systems into a data warehouse and let it do the joined-up thinking. The result was a dashboard. Pick a client and a period, and see the margin on any job.

Then the questions kept coming. We added sales data, so the sales team could compare a prospect with similar clients. If a new customer looked expensive to serve, they could price for it before agreeing the deal. Sales commission, once calculated by hand, was automated in the same tool.

It started with checking the margin on work already done. It ended with a way to judge work the business hadn’t yet agreed to. In Ed’s experience, that’s how these projects usually go. Once people see their data joined up, they start finding ways it could make their own jobs easier.

“Many businesses don’t realise the potential of their data.”

You no longer need a vast budget to start

Ed showed how far things have come with a story from his school days. His GCSE IT teacher told the class about the national Cattle Tracing System, which records the birth, sale and death of every cow. It’s a textbook warehousing problem: huge volumes, many sources and a lot to organise. As Ed remembers it, running analytical reports meant switching the system off.

That was the technology of the time. Today, the tools are better and the cost has fallen dramatically. The Cattle Tracing System itself is being replaced in England, with farmers moving to the new service in mid-2027.

Managing your data properly used to need a large team and a vast budget. Now you can start small and add to it as you go.

Bronze, silver, gold: how it works

Ed kept the technical part short. We favour an approach called a medallion architecture. It has three layers, and he explained them using a real warehouse.

Bronze is the loading bay. Data arrives from your systems and is dumped on the floor. Getting it in matters more than tidying it.

Silver is the shelving. Software checks the data, removes duplicates and catalogues it. You know what you’ve got and where it is.

Gold is the order ready to ship. Software runs the calculations a report needs. For the client above, that meant combining job details, time and expenses to work out the margin.

From there, the gold data feeds whatever people use to read it: Power BI, a custom dashboard or an AI chat interface.

Three ways to build one

Like a website, a data warehouse can be bought as a package, assembled from services or built bespoke.

A managed platform, such as Databricks or Snowflake, offers an all-in-one package with little hassle. Costs rise with use, and things get complicated outside the common use cases.

A connected ecosystem, such as Microsoft’s Azure services, OneLake and Power BI, fits together neatly. Pricing is based on usage, and setting it up properly takes effort.

A custom build uses open-source tools you host and control, without big-tech pricing or oversight. It needs real expertise to build and run.

There’s no universal winner. The right choice depends on your situation.

Join your data up and AI can use it

There’s one more reason to get your data in order. Once it’s catalogued and documented, an AI model can find its way around it.

We’re building this for another client now. A manager will be able to type a question like “Which was our most profitable month in 2026, and which jobs contributed?” and get a report on the spot.

It still needs guardrails, and someone to check the output. But it means fewer questions waiting for someone to build a new report.

Start with one question that matters

You don’t need to join everything up at once. Pick one question that matters to the business. Prove you can answer it properly. Then build from there.

If you’d like to see what’s possible first, take Ed’s tip. Download Power BI, which is free from the Microsoft Store, and connect it to a spreadsheet. The built-in guides will show you how quickly plain data becomes a useful report.

If you know your data could be telling you more, our Data Discovery is the right place to start. We invest a lot of time up front listening and understanding, because that’s what leads to the right advice. When you’re ready to build, our Data Platforms & Engineering and Power BI & Fabric teams can take it from there.