How we work
The best results come when your team and ours work in tandem. Here is what that looks like, from the first call to the handover, including why some projects run longer than others.
From first call to handover
A first conversation
Thirty minutes on your questions, your data and what is getting in the way. No slides, and no obligation to continue.
A written proposal
A plain-language plan with scope, timeline and price, so you know what you're buying before it starts. If the scope is still moving, we give a range and say why.
The work
We build in short stretches and review progress with you every week. Your knowledge of the business shapes each step, so surprises come early and cheap.
Handover
Documentation, a walkthrough for your team and a support window, so you can run it without us.
Working in tandem
We don’t disappear for six weeks and come back with a report. Your people and ours work through it together, and each side has a part to play.
What your team brings
- Context on the business: what matters, what has been tried, what is off the table
- Access to the data and tools involved, usually read-only
- One person who can answer questions and make decisions
- A short weekly check-in, plus an hour or two with the people who will use the result
- Honest feedback when something is not working
What we bring
- Methods, models and tooling, and judgement about which to use
- A weekly update with what is done, what is next and what is blocked
- Plain explanations of results, including where we are unsure
- Documentation and training so your team can run it without us
- A heads-up as soon as scope, effort or timing looks likely to change
Why it matters. The biggest cause of a failed analytics project is a good answer to the wrong question. Regular time together is how we catch that early.
What guides the work
- Work in tandem
- You know your customers, your systems and your constraints. We know the methods and the pitfalls. Neither side is enough alone, so we plan time with your team from the first week to the last.
- Start from the decision
- Every piece of work begins with the question someone has to answer or the decision they have to make. If we can't name it, we aren't ready to build.
- Show our working
- You get the queries, the definitions and the assumptions, not just the answer. Anyone on your team should be able to check what we did.
- Say how sure we are
- Data is rarely conclusive. We tell you what the evidence supports, what it doesn't, and what we'd need to find out more.
- Keep it small, finish it
- A narrow piece of work that ships beats a grand plan that stalls. We'd rather solve one problem properly and move to the next.
- Keep a person in the loop
- Models and AI tools make mistakes, sometimes confident ones. We design so that a person on your team can see why a result came out the way it did and can overrule it.
- Collect only what you need
- More data means more risk and more cost. We help you decide what's worth collecting, and treat your customers' data with care.
Some projects run longer, and we will say why
Timelines depend on scope. Two projects with the same title can differ by months once the data, the people and the systems involved are known.
These are the usual reasons a project takes longer than first thought:
- Access to data or systems takes longer than planned
- The data turns out to need more cleaning than it looked at the start
- More teams or approvers get involved than first expected
- Models need extra rounds of testing and tuning before we would trust them
- The result has to connect to other systems, such as your app or CRM
- The question itself changes once people see the first results
How we handle it. We split larger work into stages with a checkpoint at the end of each. At every checkpoint we review what we learned, re-estimate what is left and agree the next step with you. You can stop, change direction or carry on. Any change in scope or cost is agreed in writing before it happens.
Ways to work together
- Fixed-scope project
- A defined piece of work with a clear outcome, timeline and price, such as an event audit, a warehouse model or a dashboard set. Best when you know what you need.
- Staged project
- Larger or less certain work, split into stages with a checkpoint after each. Best for predictive models, AI tools and platform builds, where the scope becomes clearer as you learn.
- Monthly retainer
- A set number of hours each month for ongoing analysis, reporting support and advice. Best when questions keep coming and you don't have a full-time analyst.
- Advisory sessions
- Regular calls to review your plans, your experiments or your team's work. Best when you have people but would like an experienced second opinion.
Practical questions
How long does a project take?
It depends on scope. A focused analysis or audit takes one to three weeks, pipelines or dashboards four to eight, and a first predictive model or AI prototype usually four to ten. Larger or less certain projects can run for several months, and we give a range rather than a false promise. We tell you early if anything changes.
What happens if a project runs longer than planned?
We talk about it as soon as we see it, explain the cause and offer options: cut scope, extend the timeline or stop at a checkpoint. Nothing is added to the cost without your agreement in writing.
How much of my team's time do you need?
Usually a short weekly call and quick answers to questions between calls. Projects move fastest when one person on your side can make decisions, and when the people who will use the result are involved early.
Do you sign an NDA?
Yes, if you need one. We're happy to sign yours or provide a simple one.
Who owns what you build?
You do. Code, documentation and dashboards are handed over to you when the work is paid for.
What if we only need a few days of help?
That's fine. Small pieces of work, such as reviewing an experiment or fixing a pipeline, are a normal part of what we do.