- The information lives across shared drives, inboxes and spreadsheets.
- Finding the right file means knowing which colleague to ask.
- Settling a disagreement between two reports takes days and several people.
“To the extent that an organization is not completely flexible in its communication structure, that organization will stamp out an image of itself in every design it produces.”Melvin E. Conway, How Do Committees Invent?, Datamation, April 1968
What this phase buys you
Work can start without an archaeology expedition. What you own is known, located and owned by someone.
Every source the work touches, where it actually lives, and one named person against each row. A spreadsheet is a perfectly good first catalogue. This is where data stewardship starts, because a steward without an inventory has nothing to steward.
Not this. Not a catalogue product. Buying one before you know what you have gives you a second thing to keep up to date.
A new starter can find the right dataset or document without asking a colleague which one it is.
Every project pays the same discovery cost again, and one person quietly becomes a single point of failure because only they know where things are.
What you actually do
Four moves. None of them needs a platform, and none of them is glamorous. That is the point: this is the work that gets skipped, which is why so many companies are stuck a phase earlier than they think.
Inventory before you integrate
List what actually exists, where it lives and who owns it. A spreadsheet is a perfectly good first catalogue. The value is in having looked, not in the tooling.
Name an owner per source, not per system
Your estate is scattered along the lines your teams talk, which is Conway's point above and is why finding anything currently means knowing who to ask. So ownership follows the meaning of the data, not the software it sits in. The person who would notice it was wrong is the owner.
Kill the copies you can
Most estates carry several versions of the same extract because someone once needed it differently. Retire what nothing reads.
Make one path findable end to end
Do not boil the ocean. Take the single path the work depends on and make it locatable without a human in the loop.
The usual mistake
Buying a catalogue product and calling it done. A catalogue nobody maintains is a second thing to keep up to date, not an answer.
- The information cannot be found. The information exists but is scattered across drives, inboxes and individual machines with no index. Every task begins by asking a person where something is.