9 Jan 2026 · Journal
Reading drop-off without vanity metrics
A 62% drop-off looks like a verdict. It is often a counting error wearing a percentage sign.
Vanity metrics in User Journey Mapping Analytics usually arrive as stage-to-stage percentages with no eligibility rule. “Viewed page → started form: 38%.” Started by whom? Everyone who landed, including people who cannot use the product, people who already finished last week, and people who opened the page from a status email?
We teach a slower fraction: completers of stage N divided by people who were allowed to attempt stage N. Allowed is a product rule, not a SQL convenience. If KYC is required before payout, visitors who never started KYC do not belong in the payout drop-off. Including them makes payout look catastrophic and KYC look fine.
Absolute counts still matter
Percentages also hide small cohorts. A 90% completion on a path used by twelve people a week is not a strategy. Path Density Workshop spends half a day just labelling which stages have enough volume to argue about. The atlas should show both the rate and the headcount. Executives skim rates; operators need headcount.
Time is the third vanity trap. “Average time on step” without a note on waiting for a human is fiction. In service journeys the clock is often someone else’s inbox. Report time-to-next-eligible-action, and say whose action it was.
If your current dashboard cannot express eligibility, write it in the atlas legend anyway. The legend is part of the map. Journey Signal Lab session two is almost entirely this unglamorous work. Teams who skip it keep shipping percentages that feel decisive and fail in critique.