Data science & analytics
You almost certainly don't need another dashboard. You need an answer to a question, and then possibly to stop looking.
Dashboards are where questions go to be forgotten
The pattern is familiar. Someone asks a question. Rather than answering it, a dashboard gets built so anyone can answer it whenever they like.
The dashboard is used enthusiastically for two weeks. Then it's used for one number. Then the underlying data changes and nobody updates it. Two years later it's still on the intranet, quietly wrong, and someone has built a spreadsheet next to it.
The original question never got a clear answer, and the organisation now maintains a permanent artefact instead.
Sometimes a dashboard is genuinely right — a number a team steers by daily. Often the honest deliverable is a piece of analysis, an answer, and a recommendation, after which the work is done.
What we do
- Focused analysis
- A specific question, investigated properly, answered in writing with the reasoning and the caveats visible. Churn drivers, pricing effects, cohort behaviour, funnel drop-off, segment profitability.
- Experiment design and analysis
- A/B tests set up so the result will actually be interpretable, and analysed so you don't ship a change on noise.
- Metric definition
- Agreeing what your core numbers mean and how they're calculated, before three teams report three different figures to the same meeting.
- Dashboards, where they're warranted
- Built for a specific decision, a specific audience, and a specific cadence. With a note about when it should be retired.
- Data exploration
- Working out what's in your data and what it could support, when you're not yet sure what to ask.
How we approach it
- We start with the decision
- What will you do differently depending on the answer? If nothing, the analysis is interesting rather than useful, and we'd rather find that out before you pay for it.
- We show the working
- Assumptions, method, and limitations stated plainly. Analysis you can't interrogate is analysis you shouldn't act on.
- We report uncertainty honestly
- A number without a confidence interval invites more precision than it deserves. If the data can't answer the question at the resolution you need, that's the finding, and it's a real one.
- We look for the confound
- The obvious relationship in the data is frequently explained by something else. Checking that is most of the difference between analysis and coincidence.
Before you ask.
For sustained work, an analyst is better and we'll say so. This suits specific investigations, one-off questions, or building the foundations before you make that hire.
Data Engineering
Nobody has ever asked us for data engineering. They ask for the dashboard, the model, or the AI feature — and then we find out why it isn't working.
Data & CloudPredictive Analytics
A forecast nobody acts on is a hobby. The prediction is the easy half; the hard half is what happens next.
AI & MLMachine Learning Models
A model that scores 94% in a notebook and a model that earns its keep are different achievements. The distance between them is where this work lives.
Tell us what you're trying to build.
A 30-minute call, no charge and no pitch deck. Describe the problem and we'll tell you how we'd approach it, roughly what it costs, and whether we're the right team for it. If we're not, we'll say so.
30 minutes · No charge · No deck