What is data science?
A plain-English guide: what data science actually does, how it differs from dashboards and from machine learning, and when it earns its keep.
Last updated
The definition.
Data science is the investigation: clean the mess, test what actually moves the number, then choose a model only if it pays. It turns the history you already keep into answers you can act on — what is driving the figure, what is likely next, and which experiment is worth running.
It sits under data and analytics, next to business intelligence. Intelligence is the reporting layer. Science is the question underneath.
What the work actually is.
Four steps, always in this order. Skip the first three and the model is guesswork.
A question worth answering
What is driving churn, what you will sell, which leads are real. A clear question beats a vague model.
A table you can trust
Spreadsheets, CRM, shop exports — cleaned, joined, and defined the same way finance would. This is most of the work, and often the part that pays first.
A test, not a hunch
Statistical checks against your own history. If the signal is not there, that is the finding — and we will say so.
A model only if it pays
Where a forecast, score or flag would change a decision you already repeat, that is machine learning integration — the same job, wired into the workflow.
When it is worth it — and when it is not.
It pays when you have history, a repeated decision, and a question that would change Monday if you knew the answer. Demand, cash, churn, lead quality, the odd number that should never sneak up.
It does not pay as a science project. If a trusted report answers it, that is business intelligence, and we will not add a model for the sake of it. If the job is reading and writing language, that is LLM integration.
How the work is scoped, shipped and kept running is the same as the rest of our tech consulting.
Frequently asked questions
Is data science only for large companies?
No. The useful version is a clear question against the history you already keep — what is driving churn, what you will sell, which leads are real. You do not need a lab. You need a question worth answering.
How is data science different from business intelligence?
Business intelligence is the trusted view of what happened. Data science is the investigation: why it happened, what is likely next, and which experiment is worth running. Most businesses need the view first.
How much history do we need?
Less than people expect. Many useful findings come from a few years of invoices, CRM and shop exports. If there is not enough to learn from, we will be straight about that before anyone trains a model.
Do we get a report, or a model?
Whichever answers the question. Sometimes a cleaner table and a finding is the win. A model is only built when a forecast, score or flag would change a decision you already repeat.
Who uses the result day to day?
Your team. We do the analysis and keep any model honest. There is nothing to learn except the answer — and, where a model is wired in, a button or a flag in the workflow you already use.
Tell us the number you want explained — or predicted.
Send a few details and we will set up a short call. No obligation, no hard sell.