Case study 08 · Classifier recovery · July 2025

Reading the Helpful Content resample curve across 12 sites.

Twelve sites under a sitewide classifier penalty, instrumented across the resample cycle. The recovery curve is real, measurable, and predictable.

Sector
Classifier recovery
Published
July 2025
Models tracked
14 frontier LLMs
Reference
Case 08
Sites instrumented
n = 12
Measured resample
60–90d
Median recovery
+148%

The situation

Twelve sites, all hit by the Helpful Content System’s sitewide, multiplicative downranking. The industry folklore said “wait and pray.” We instrumented instead.

The approach

We treated the classifier as an object to be measured, not a curse to be lifted. Each site was tracked through the resample cycle with a fixed remediation applied at a known date, so the recovery — when it came — could be attributed.

  • Sitewide quality floor. The classifier scores the site, so thin pages anywhere drag everything. We pruned, not padded.
  • Resample timing. Remediation landed early in the cycle so the next resample would read the improved corpus.
  • Holdout. Two sites were left unremediated as controls.

The result

The remediated cohort recovered on a 60–90 day cadence with a +148% median rebound; the controls did not move. The resample curve is not mystical — it is a measurable cadence you can plan a remediation against.

The lesson

Most “algorithm recovery” is guesswork dressed as expertise. With a holdout and a calendar, it becomes engineering. Measure the cadence, time the fix, attribute the result.

Gilad Sasson

Run this on your own domain.

Gilad Sasson, also known as Algoholic runs every audit personally. Every engagement opens with a thirty-minute call — and a candid read on whether it is worth doing at all.