Skip to main content

Outcome verification

How often the #1 for you held up

When a reader saves a decision and confirms their email address, a follow-up is scheduled 30 days later. The rate below is the share of conclusive answers that reported signing up and being satisfied, and it is published only from 30 conclusive answers.

Loading outcomes…
How sure readers feel (the SURE test)

Four yes-or-no questions from the Ottawa SURE test of decisional conflict, asked at the result and again in the 30-day follow-up. The figure is the share answering yes to all four, published per category only from 30 answers.

What "accuracy" means here
  • Counts as right: the reader answered the follow-up that they signed up and are satisfied.
  • Counts as wrong: the reader answered that they signed up and are unsatisfied, or that they switched provider.
  • Excluded from the denominator: answers of "did not sign up" and "too early to tell".
  • Prior: none. The rate is the raw share of conclusive answers, with no smoothing toward a site-wide figure.
  • Window: every answer from the last 180 days counts equally; older answers drop out.
  • Interval: a Wilson score 95% interval, printed only from 30 conclusive answers.
  • Caching: the rate is computed from the outcomes table on request by /api/outcomes/stats, and the response is cached for up to an hour.
  • Rebuild it yourself: the same response publishes the anonymised counts per category and answer (dataset), their totals (distribution) and the recipe (method), so every figure above can be recomputed from what is published.
Why this matters

The rate is computed from the answers with no manual step, and the same floor withholds it for every category alike. For the methodology that produces the Fit rank, see methodology. For how often a saved #1 has changed, see the drift feed.