Case study · Supplements Wise
12.1x search clicks in one quarter, from a store with no content team
3.9 clicks a day in late February. 371 a day in August. Every number below is from Google Search Console, exported 28 August 2026, covering 27 February 2026 to 26 August 2026.
Quarter on quarter: 1,170 clicks (27 February 2026 to 27 May 2026) became 14,123 (28 May 2026 to 26 August 2026). Peak day so far: 430 clicks on 25 August 2026.
* partial month: February has 2 days of data, August has 26.
The store
A supplements catalogue, and a blog nobody planned
Supplements Wise sells supplements for people and dogs in the UK. Until March 2026 its blog was written when there was time, about whatever seemed interesting, with no view of what Google was already showing the store for. Search traffic was a rounding error: under ten clicks a day.
From March the store ran the loop that became QueryClimb. Same products, no new hires, no paid traffic, no link buying. The site did get technical and design work in the same window, which the caveats below account for.
The loop plans, drafts, links and schedules. A named author still reviews and stands behind every article, and every citation is checked against its source before anything is published. For a health brand that is the only acceptable way to use the tooling, which is why those checks are built into the product rather than left to discipline.
What changed, and when
- March 2026Search Console becomes the brief. Every opportunity classified: new, refresh, rewrite, or leave alone. Drafts written in the store's voice, internally linked, published on a schedule.
- SpringCannibalisation preflight added after a new article knocked out a winning collection page. Citation verification added after a draft cited studies that did not exist.
- SummerScheduled publishing: the loop runs on a schedule, drafts land for review, and the morning brief reports what shipped. Internal-link rescue for orphaned articles.
- August 2026GEO fix batch: author page, linked citations, schema de-duplication. AI visibility score 32 to 60 out of 100 (+28 points).
Google began reporting this surface in Search Console in May 2026. The store went from 581 impressions a day to 12,697 in three months, 481,425 in total, with 97% of them on blog articles. Pages that answer questions directly with verified sources are what AI surfaces cite.
Where the clicks came from
Articles, not product pages
| Content type | Share of clicks | Avg position |
|---|---|---|
| Dosage guides | 35% | 7.2 |
| Other articles | 23% | 9.9 |
| Timing and duration guides | 15% | 7.0 |
| Product and collection pages | 12% | 19.4 |
| Comparison articles | 8% | 7.2 |
| Pet supplement guides | 8% | 9.7 |
824 of the store's top 1,000 queries now rank on page one; 87 in the top three. Most of them are dosage, timing and comparison questions that product pages cannot answer.
Caveats
What this does and does not prove
- It is a timeline, not a controlled experiment. The interventions are dated against the traffic curve; no holdout existed.
- Content volume grew alongside the tooling. The loop produced far more published articles than the store had before. Part of the lift is simply publishing consistently; the tool's job was making that consistency possible for one person.
- Low base. Multiples from six clicks a day are easier than multiples from six hundred.
- Content was not the only change. Theme and design work, breadcrumbs, schema de-duplication and speed fixes all landed inside the window. Content was the largest change by far, and the one the traffic curve tracks, but the lift is not attributable to it alone.
- Google changed too. Core updates and the arrival of AI features in Search Console fall inside the window.
- Clicks, not revenue. We report search metrics only.
Your search data already knows what to write next
The free audit runs the first two stages of this loop on your own store: your striking distance list and your AI visibility score, in minutes.
Join early access