The problem is not that you lack knowledge. It is that each analysis sits on its own, so year two starts over instead of building on year one. A knowledge graph connects the conclusions, so a new analysis automatically receives the context relevant to it.
In year one you analyse the market. Five important conclusions come out. They get written down.
In year two the same market needs analysing again. You start over and hope somebody remembers what it said.
That is the classic problem. Every analysis is an island. You find something important, and then it sits in a folder nobody opens again.
What a knowledge graph is
Imagine analyses, decisions, market data and competitor insights sitting in one connected structure rather than in separate files.
Not hierarchical, not a folder tree. A network, where each piece of knowledge is linked to what it affects.
One node might be that online sales in your industry grew noticeably last year. Another is that a competitor launched an online product in the third quarter. A third is that you decided to stay with physical channels.
Separately, those three are just notes. Connected, they are an explanation, and the next time somebody analyses the market, the whole connection comes with it.
You do not start over. You start from here.
How it works
An analysis has input and output. Input is what it works on, output is what it finds.
In a knowledge graph the output of one analysis becomes input to the next. Each time an analysis runs, the conclusions are stored as nodes with links to what they concern. When somebody later analyses the same area, the relevant context is retrieved automatically.
The practical difference is what the analysis has to go on. Without the context it works on an isolated dataset. With it, it also knows what you found last time, and can address what has changed.
That does two things: the analysis gets deeper, and it gets faster, because the basis does not have to be re-explained every time.
Why it matters
Most smaller companies run the same cycle.
Year one. You spend three months on strategy. A dozen analyses come out. They get documented, the meeting happens, the decisions get made.
Year two. The strategy needs revising. The documents are somewhere, maybe in a shared folder, maybe in an email thread. In practice you start over, and reorder some of the same analyses.
Years three to five. Nobody can recall precisely what was decided in year one, or why. So you analyse again without knowing whether you already had the answer.
It is not that anyone is bad at searching. It is that knowledge sits scattered, and nobody knows what to search for.
With a structure that connects things, you do not need to search. A competitor analysis in year two contains, by itself, the fact that the competitor opened a new channel the year before and what that channel did to your position. You wrote it down then. Now it surfaces exactly where it is needed.
The value compounds
This is where it gets interesting, and also why it is hard to see at the start.
Year one is the first pass. You find a number of things.
Year two builds on year one. You find fewer new things, but sharper ones, because you can now see what has moved rather than only what is.
Year three has context from two full rounds. Patterns become visible that could not be seen before, because they only appear over time: a segment slowly becoming less profitable, an assumption you have repeated three years running without testing, a competitor moving consistently in one direction.
It is not that the system gets better at delivering data. It gets better at connecting it.
The difference between a database and a graph
A database stores documents. A graph stores relationships.
Store a report in a database and you can search for the word margin and find every report mentioning margin. That is useful.
In a graph the question is different: you analysed margin in the context of a particular customer segment, you later analysed that segment again, and here is what the change means for the margin.
That is not the same information retrieved faster. It is a relationship that existed in none of the documents separately.
How to start
You do not need five years of data to begin.
- Pick three areas. Typically market, competitors and your own capabilities.
- Write conclusions as conclusions. Not as minutes. One line that stands alone: what did we find, and what was it based on.
- Note what you did not know. That is what makes the next analysis sharp, because it knows where the gap was.
- Repeat in three months. The first time you notice no difference. The second time you do.
Point three is the one most often skipped, and it is the most valuable. A conclusion without its uncertainty cannot be checked later.
Why the graph is built into the analysis on our side
A knowledge structure that depends on somebody remembering to maintain it does not get maintained.
In 360° Sprint the nodes on the board are connected, and the connections determine the order the analysis runs in. A node receives its context from the ones upstream through the knowledge graph, so the competitor analysis genuinely knows the conclusion from the market analysis. Results collect in Insights with run history, so two analyses of the same area can be held against each other over time.
It is not something anyone has to remember to do. It is a consequence of the analyses being connected rather than sitting beside each other.
Why being able to return to the basis matters at all is covered in strategic intelligence as a working method, and how the phases of the work connect is in the strategy chain.
The real gain is not that you know what you knew. It is that new knowledge builds on the old automatically, instead of replacing it.
The process it belongs to is described in strategic planning for owner-managers.