The stopwatch¶

The pilot has a dashboard. Questions answered: eleven thousand. Time saved: four hundred hours, by a method the vendor supplied. Documents found: most of them. Satisfaction: high, and Kevin’s is the highest. The Head of Fundraising presented it to the board in a slide whose graph pointed upward.

The metrics arrive pre-selected by ease. Questions answered, time saved, documents found, satisfaction scores: everything a search replacement can be measured as, the reader will be measured as, and the numbers will be good, because finding really did get faster. The rest of the bill lives in the metrics nobody collects: wrong decisions prevented or not, ambiguity surfaced or smoothed, exceptions routed correctly, tacit knowledge retained past the next redundancy round, memory formed in the people doing the asking. A search replacement is easy to measure. A change to how an organisation thinks is not. The stopwatch, pointed at the finding, cannot see the queue re-forming behind the answer, where retrieval used to be the constraint and interpretation now is. The Home’s constraint may turn out to have been interpretation bandwidth, not documents, and the reader has moved the queue without shortening it.

The hours saved on finding were not homogeneous either. Some searching was waste; some was discovery; some was relationship maintenance; some was the sanity check that caught the wrong number before it reached a board paper, and the Home’s board papers have carried a wrong number before. The stopwatch records them all as the same saved hour. What the measurements will show, for the first year or two, is a success. What they are structurally unable to show is which stocks the success is drawing down.