Survivorship bias appears when analysis retains only still-visible units. A network evaluated without abandoned accounts can look artificially strong.
01STORY BRIEFThe idea in three screens.3 SCREENS
The missing accounts had results too.
Survivorship bias appears when analysis retains only still-visible units.
Keep exit reasons and last observed dates. Separate intentionally stopped, inaccessible and uncollected accounts. Portfolio analysis should begin with an initial cohort and explain what happened to it. Surviving assets answer a different question from the original investment.
Show a small cohort-status table: still active, intentionally stopped, inaccessible and unknown. Categories should reflect verified facts and dates. This explains why counts decline and prevents strong survivor outcomes from being presented as returns across every account originally launched. Keep unknown states separate instead of assigning a convenient reason for disappearance. The incomplete history is itself a limitation of what the final performance comparison can claim.
Keep the accounts that disappeared
The accounts visible today are not necessarily the accounts that started. Some stopped publishing, changed direction, lost access or never found a workable production routine. If the report includes only the active survivors, it can make a demanding process look easy to repeat. Build the cohort from the starting roster and preserve status changes over time. An inactive account is information about the operation, even when its final view count is awkward to collect. Record missing data as missing rather than replacing it with success or failure by assumption.
The sample you see versus the one you started
- 01Start
Keep the original roster and planned scope.
- 02Exit
Log stopped, missing and changed accounts.
- 03Report
Show outcomes without deleting inconvenient histories.
Editorial model: a way to reason about the process, not measured platform results.
The same problem appears in creative swipe files. Public examples are selected because someone noticed them, while the quiet attempts remain scattered across accounts and drafts. That makes them useful references for craft and poor evidence for expected results. When writing a case study, separate the visible mechanism from the probability of success. For internal decisions, use the complete attempt log and account for time spent on work that never shipped. The most useful denominator is often the one the attractive presentation was about to leave out.
“The missing accounts had results too.”
02FIELD KITFrom concept to practice.2 SCREENS
Freeze the initial cohort.
Adapt this to your audience, budget and test scope.
Put it to work.
- 01Freeze the initial cohort.
- 02Record exits and their costs.
Notes & sources.
The limit. Lost data must not be invented; report the coverage limitation.
Further reading is linked throughout this article. Worked examples explain a process; they are not campaign results.
THE ORGANIC CLUB is an editorial project initiated by TokPortal. That relationship is separate from the evidence behind claims. Editorial signatures and image credits are explained in our policy. Our editorial policy →
