Note ·

A maturity ladder for internal sources

Open dataOpen standardsMetricsEnterprise architectureKnowledge management

Read access is the bottleneck, and bottlenecks need a number. Tim Berners-Lee’s five-star ladder for open data was built for governments publishing to the public, and it transfers almost unchanged to a company publishing to its own agents. One star: the data exists somewhere, a PDF or a spreadsheet on a share. Two: it is structured and machine-readable, an export, a report API. Three: it is in an open format rather than a vendor’s. Four: things in it have stable identifiers, so a record can be referred to from elsewhere. Five: it links to other sources by those identifiers. Most internal systems I have met sit at two, and the jump that matters for a context warehouse is from two to four, because without stable identifiers the taxonomy cannot join anything.

The second number is time. From the day someone asks to read a source for the warehouse to the day its first record lands, per source, in days. It captures everything the ladder does not: the owner’s reluctance, the ticket queue, the API that exists in the licence and not in practice, the MCP server that is a demo. It is a number a leadership team can be shown, and it tends to move once it is shown.

The caution from Patents per Capita Is a Vanity Metric applies: both numbers can be gamed, by publishing things nobody needs at five stars or by counting a stub as a first record. The defence is to measure only the sources an actual proposal was waiting on.

References

    Open the full search →