Where they started
When we started, we measured a full 90 day baseline before touching anything, from 30 April to 28 July 2026, with data across all 90 days. In that period the site averaged 1.88 clicks a day and 4.41 sessions a day, sitting at an average ranking position of 35.9 across the keywords we track, roughly the bottom of the third page of results. It picked up 41,412 impressions over those 90 days, so the pages were being surfaced in search fairly often. They just weren’t being clicked.
That’s the shape of the problem: visibility without conversion into visits. Against a backdrop of 432 pages managed, 497 keywords tracked, and 1,908 open issues logged on the site, the plan for this first stretch of work was not to chase rankings directly but to clear the on-page fundamentals that were likely holding position down: metadata and internal linking, the categories where the platform had the clearest, lowest-risk fixes ready to go.
Over the 30 days from 29 July to 27 August 2026, here’s what actually went out:
- 42 changes applied to the site, out of 171 proposed (the rest sit in the approval queue, not rejected)
- 38 of those applied changes were metadata fixes, 4 were internal linking changes
- 25 of the applied changes have been verified since going live
- none were rolled back
Against the 1,908 open issues, 106 resolved during this window. We can’t claim those as ours: the platform’s own count of issues it fixed directly this period is zero, so that resolution figure reflects the state of the site generally, not our work specifically.
What this first month produced, measured per day against the baseline: clicks held essentially flat, 2 a day against 1.88 before. Sessions per day rose, from 4.41 to 6.8. Impressions per day also rose, from roughly 460 to around 620, so more of the site’s pages are being shown, ahead of any of that translating into clicks. Average position sat at 36, against 35.9 in the baseline, which is not an improvement, just noise at this scale.
Of the 21 keywords we could compare directly, 9 improved and 9 declined, for a net average gain of 0.9 positions. One keyword moved into the top 10. None reached the top 3. That is a genuinely mixed result, not a clean win, and one month of metadata and linking work was never going to move 497 tracked keywords on its own. The next chapters will show whether clearing more of the 171-item queue starts to move clicks, not just impressions and position.