Originally written in French. Translated by AI — the meaning has been preserved, not the prose.
Angle
A software vendor's advantage now lies in what it understands of its customer's ground, not in what it knows how to produce — but an understanding that stays in the heads of a few people is not a company advantage, it is an individual talent that leaves with them. What turns one into the other is a piece of writing: one context document per opportunity, tying each section to the risk it investigates and distinguishing what the team knows from what it believes, from what gives it doubts and from what it leaves unsaid. That writing is also what makes the understanding arbitrable: as long as it stays spoken, it cannot be contested, cannot be dated and cannot be handed on, so that none of the dimensions of the context — the customer's decision culture, the real user, the buyer's criteria, technical soundness, traceability requirements, the shifting market — can be picked up by those who were not in the room.
Synthesis
Taken separately, these notes look like a list of things one ought to understand better. Placed end to end, they describe a moat shifting, then the condition that decides whether the shift actually happens.
The shift has a simple cause, and it is economic. An advantage that rested on an entry ticket — servers, databases, infrastructure teams — dies the day the same resource can be rented by the hour, because what founded it was never the building block but the ability to afford it. The same mechanism then reaches production itself: more features, more screens and more code cost less every year, more players can attack a market, and producing fast becomes the baseline. What has not fallen in price is the understanding of what to build; scarcity does not disappear, it changes place.
What that understanding covers is not a single category. A customer's decision culture makes two expectations diverge beneath an identically worded need. The user on the ground is not the one in the framing workshop, and generating the screen faster does not bring the two any closer. The buyer assesses a decision where the user assesses a tool, so that an adopted product can be blocked at signature. The technical question stops being feasibility and becomes integration, maintenance, security and holding up in operation. Compliance requirements arrive disguised as ordinary sentences about a figure to be justified or a comparison between sites. And the market moves while you build, which means reading where it is standardizing and where blind spots remain.
These six dimensions share a trait the enumeration hides: none can be observed from head office, and none survives without being written. They are acquired in interviews, on site visits, in pre-sales, in support — so by people, in situations, at a given moment. An organization that leaves them spoken pays for them again with every project, every arrival and every departure.
That is what gives the context document its place in the reasoning. Its point is not to gather facts — a shared folder would do that too — but to impose two constraints no other deliverable imposes. The first is to qualify: marking what is known, what is believed, what raises doubt, what goes unsaid, which brings out the areas where a decision rests on an unverified belief. The second is to order by risk rather than by discipline, which changes what an empty section means: it no longer flags a contributor running late, it flags a risk nobody has investigated.
The shift is therefore paid for in writing work, and that is where it most often fails. Producing one more feature can be shown in a demo; writing down what you do not yet know about an opportunity cannot. The turn hinges on what a team agrees to count as work before it has produced anything visible.
Tensions / contradictions
One tension bears on the price of writing. A qualified context — dated, owned, revisited — costs time to maintain, and that time is taken out of the time the falling cost of production had just freed up. Nothing here says at what volume of opportunities maintaining the document costs more than the understanding it preserves.
Second tension, between two notes in this batch: a customer's decision culture and unspoken constraints are precisely what a team perceives without being able to formulate, whereas writing demands a formulation. Part of what makes the advantage therefore risks staying outside the document meant to preserve it.
Questions
- What, in an understanding acquired in pre-sales, is inevitably lost at the moment it gets written down?
- How does an organization recognize that a competitor has overtaken it on understanding of the context, when that gap shows up in no feature comparison?
- On what signals do you decide that a context document has aged, when the dimensions it describes evolve at different rates?