The headline number is the kind that gets used in pitch decks: 4,290 impressions to over 117,000, roughly twenty-five-fold, for a pet products exporter. The work behind it was considerably less exciting than the number suggests, which is exactly why it is worth describing.
The starting position
A product business with a genuine catalogue, real customers, and a website that search engines could barely make sense of. Products existed in a flat structure with no meaningful category layer, internal linking was accidental, and the pages that could have ranked for commercial terms did not exist at all.
What was actually done
First, category architecture. In product businesses the category pages carry most of the ranking potential, and they were missing. Building them, with real content rather than a grid of thumbnails, created pages capable of competing for the terms buyers actually search.
Second, content hierarchy. Deciding which page owned which term, so the site stopped competing with itself, and making the relationships explicit through internal links from supporting content into commercial pages.
Third, the unglamorous technical layer: indexation, crawl paths, and making sure the pages that mattered were reachable in fewer clicks than the pages that did not.
Why the multiple looks so large
Honesty about base effects matters. Starting from 4,290 impressions means the site was nearly invisible, so proportional growth was always going to look dramatic. A site starting at 400,000 impressions cannot produce a 25x headline no matter how good the work is. The multiple is a function of the starting point as much as the execution, and anyone quoting multiples without base numbers is hoping you will not notice.
What transferred to other clients
The method, not the number. The same sequence, architecture before content, one owner page per term, internal links pointing at commercial pages, produced 2.87 million to 6.36 million quarterly impressions for a bank and 327,000 for a fertility clinic. Different industries, different scales, same order of operations.
The part people skip
Patience. Category pages built in month one did not rank in month one. The curve was shallow for a quarter and then steepened, which is what compounding looks like from the inside and why so many e-commerce SEO programmes are abandoned a month before they would have worked.