Our method

A data project rarely succeeds on technology alone.

It succeeds when the business rules get written down, when access arrives on time, and when acceptance testing was planned for. Here is how we run a project, and what we will ask of you, so that you can plan for it.

Phases, from scoping to handover: 7
Peaks of load for your teams: 3
Of scoping to get started: 2 h
Expert consultants: 11

7
Phases, none of them skipped: 3
Moments that genuinely mobilise your teams: ½ h
Weekly check-in between those moments: 2 h
Free scoping workshop before any quote

Our experience

The three things that make a project work

We have run data projects long enough to see the same causes come round. None of them is technical.

The first is access to the data. The project is approved, the budget is voted, and it takes six weeks to obtain a service account on an ERP nobody remembers administering. We now ask for the access inventory before writing a line.

The second is the gap between written rules and applied rules. Margin is defined in a 2019 procedure; management accounting has computed it differently since 2022, in a workbook that never left one desk. Both figures are defensible. Someone has to decide, and it is not for us to do it, but it is for us to make the gap visible.

The third is acceptance testing treated as a formality. A report nobody has checked against their own figures is not a delivered report: it is a report waiting to be challenged. Acceptance is the moment a project wins or loses its credibility, and it has to be prepared.

The sequence

Seven phases, and what each one costs you

Your share of the work is not spread evenly

It concentrates on three moments. Knowing them in advance means booking the right people in the right weeks, and it is the leading cause of slippage when it is not done.

2-M shareYour share

Scoping
2 hours
50 % pour vous · 50 % pour nous
Your stakes, your sources, your volumes. Nothing to prepare, we listen.
Access and inventory
3 to 5 days
65 % pour vous · 35 % pour nous
The most demanding moment for you: opening access, identifying the right contact for each system, finding out who holds which password.
Data foundation
1 to 4 weeks
5 % pour vous · 95 % pour nous
Extraction, history, raw layer. We work, you get on with your day.
Business rules
1 to 3 weeks
40 % pour vous · 60 % pour nous
The second peak. The rules are written nowhere: they live in the heads of your management accountants, and they have to be drawn out.
Model and reports
2 to 6 weeks
15 % pour vous · 85 % pour nous
Semantic model, measures, reports. Short weekly check-ins.
Acceptance
1 to 2 weeks
70 % pour vous · 30 % pour nous
The third peak, and the most important: your teams compare the figures with their own. A gap found here costs a hundred times less than a gap found in production.
Handover
2 to 5 days
45 % pour vous · 55 % pour nous
Training, documentation, keys handed over. You must be able to change a report without us.

The proportions are orders of magnitude observed on our projects; they vary with the number of sources and the availability of your teams. The durations match a mid-sized project: a focused Power BI project compresses the seven phases into a few days, without skipping any.

What we will ask of you

The list, given up front

These six items decide when the project can start. Gathering them usually takes longer than expected, which is reason enough to begin now rather than after signature.

Read access to each source system

A service account, not the personal login of an employee who will leave one day. For a hosted ERP this often goes through your integrator: the request should be started early.

One contact per system

Someone who knows what the table contains, not only what it is called. An hour of their time is worth three days of reverse engineering.

Your current reports, imperfect as they are

The spreadsheets your teams maintain by hand are the best documentation of your business rules that exists. We do not judge them, we read them.

Someone to arbitrate definitions

A person with the authority to settle between two definitions of revenue. Without them, the project stops at the first divergence, and there is always one.

Two or three business testers

People who know their figures by heart and will spot a missing branch immediately. Plan their availability from the initial schedule, not the week acceptance begins.

Your volumes, in orders of magnitude

Sales lines per year, number of items, depth of history wanted. Those three figures are enough to size the platform and to rule out needlessly heavy architectures from the start.

The relationship

Fixed price or time and materials: what each really implies

The question comes up every time, and it is often settled for the wrong reasons. A fixed price reassures because it fixes an amount; it also fixes a scope, and anything outside it becomes an amendment. Time and materials worries people because it shows no total; it also leaves the freedom to change your mind in week four, which almost always happens.

Our position fits in one sentence: agile while the scope is moving, fixed price once it is settled. Many of our projects start on time and materials for the scoping and the inventory, then switch to a fixed price once the sources are known. That is the only moment a fixed price can be right. Our billing model in detail.

We never quote a fixed price on the basis of a description alone. A quote drawn up without opening the data is a gamble, either at your expense or at ours, and in both cases the relationship suffers before it has even begun.

Going further

The pages that detail each step

Frequently asked

What people ask before starting

How much time will our own teams have to spend on this?

Less than you fear, but not evenly spread. The load concentrates on three moments: opening up access, writing down the business rules, and acceptance testing. On a mid-sized project, count one to two cumulative days per key person on each of those. Between them, a thirty-minute weekly check-in is enough.

Do you work on a fixed price or time and materials?

Both, and the choice is yours. Time and materials lets us start immediately and lets you adjust scope along the way, which suits a need that is still taking shape. A fixed price is quoted after the source inventory, which assumes the scope is settled. We never quote a fixed price before seeing the data: a quote built on a description is a wrong quote.

Can you take over a project someone else started?

Yes, and it happens often. We start with an audit of what exists: what can be reused, what has to be rebuilt, and why. We say it plainly, including when the answer is that almost everything is reusable and the engagement turns out shorter.

What if our business rules are written down nowhere?

That is the normal case, not the exception. The rules live in personal spreadsheets and in two or three people's heads. Part of our job is precisely to surface them, write them down and have you validate them. That documentation is part of the deliverable: it belongs to you and it outlives us.

What if we want to take over afterwards?

That is the aim. Handover is part of the project rather than an option: training for your teams, documentation of the model, legible naming conventions. You should be able to add a measure or change a report without calling us back. We stay available for substantial changes, not for colour tweaks.

The first step

Two hours, minute by minute

Twenty minutes on your systems, forty on how your figures are calculated, thirty on the architecture we would put in place, thirty on the sequence and the load on your teams. Nothing to prepare on your side.

Two hours, and here is how they go

Nothing to prepare on your side. We leave with enough to tell you what we would do.

The workshop is free and carries no commitment. We sometimes conclude that there is no project to run, which is faster for everyone, and it is what makes the next yes worth something.

Let us start by looking at your sources.

Eleven expert consultantsSaint-Priest, near Lyon, France

Two hours with a consultant, free of charge and with no commitment. We look at your systems, your business rules and your volumes, and we tell you how we would run the project: the sequence, the load on your teams, the friction we anticipate.

Request a scoping workshop

Scoping workshop · 2 hours · free · near Lyon or remote