What Data Intelligence Means for Small Businesses

The term gets used to mean too many things

"Data intelligence" can mean a multi-petabyte warehouse with a dedicated analytics team, or it can mean a single dashboard that answers one question a business owner currently has to dig through a spreadsheet to answer. For a small or mid-sized business with no data team, it almost always means the second thing -- and conflating the two leads to either over-building (a warehouse nobody queries) or under-building (more spreadsheets).

The actual starting question

Not "what data do we have" but "what question do you keep re-answering manually." That's usually a much smaller, more specific thing than "we need a dashboard" -- it's "how many of last month's leads converted, broken down by where they came from" or "which of our products are actually profitable after returns." If you can name the specific question, the data intelligence work is building the smallest reliable thing that answers it directly, not a general-purpose analytics platform.

Why this matters for scoping the work

A dashboard built around one real, recurring question is something people actually open, because it replaces a task they were already doing badly. A dashboard built around "give us visibility into our data" in the abstract tends to get built once, demoed once, and opened rarely after that -- because it wasn't anchored to a specific moment of pain.

This is the same scoping discipline described in how we approach software generally: understand the real, specific problem before building anything, rather than starting from the solution shape.

What this looks like technically, without a data team

You don't need a separate analytics warehouse to answer most of these questions for a business at this scale -- the data usually already lives in the same operational database the rest of the business runs on. The work is mostly: getting the query right once, presenting it as something non-technical people can read without translation, and keeping it current automatically instead of as a one-time export.

That's a meaningfully different (and cheaper) project than "build us a BI platform," and naming that difference honestly is part of scoping the engagement correctly. See Services for where this fits among the things we build.