The products weren't the problem. The catalog was.
Nearly forty product pages existed with no consistent keyword targeting, collection pages with nothing mapped to them, and dozens of keywords tracked against nothing. Before touching a word of copy, we needed to know what every page was supposed to win. For a year, I coached their marketing team twice a month, an hour at a time.
We let real buyer hesitation set the priorities.
The next decision mattered more than it looked: start with what shopping assistants actually said about the products. The traffic data showed where people landed; the secret shop showed what buyers still weren't getting answered.
- I read the data, coached the calls, and set direction.
- We decided together which products needed attention and which questions their pages should answer.
- Their team, working with their agency partner, wrote the specs and did the implementation themselves.
Diagnose: map the catalog before changing it.
With any client whose catalog has gone unmapped, the first thing I do is pull the whole page catalog into a structured record before any content work starts. The catalog held 46 live pages: 8 collections and 38 product pages. Every collection had zero mapped keywords, and 68 keyword rows were attached to no page at all. That's the part most brands skip: knowing exactly what exists and what each page is supposed to win.
Decide: ask the shopping assistants first.
I ran an AI secret shop on the three highest-revenue products and let the results show where the hesitation was. The tuna line was already strong on texture, flavor, and sustainability. Sardines looked healthy and sustainable, but disappeared when a buyer asked about taste or occasion. The same check caught AI engines crediting the brand with a certification it had chosen not to pursue.
A repeatable process their team could keep running.
The AI secret shop and product page process turns shopping assistants' answers into a plan for better page content. It checks what a page already says against what buyers actually ask, then closes the gaps with structured FAQs.
Build: turn each hesitation into an answer.
The work used a repeatable have-versus-want method, with seven FAQs per product in the first pass. Real questions became real answers: why the cans have no water to drain, the difference between tuna cuts, and ways to eat it beyond a sandwich. Mercury questions called for actual testing numbers against the federal limit. Every FAQ went into a structured database record, so the work stayed trackable instead of sitting only in a document.
Run: leave the team able to do it again.
Their team and agency partner owned the actual build on the live product pages. My role was coaching and setting direction. Twice a month, I walked their marketing team through what the AI models were saying and what the traffic data showed. The work extended beyond the first three products, with FAQs for another product in the database and two more drafted in August.
Grow: give every page a target.
Alongside the FAQ work, we closed a keyword map across all 46 collection and product pages in September. Every page now carries a tracked target keyword and a baseline for ranking reports. The work kept extending beyond the measured traffic window, with refreshed personas and a claims document correcting an overstated sustainability claim.
Product pages now pull more of the weight.
69%
product page organic traffic growth across 38 pages, January to August 2025 versus 2026
33%
of organic visits now land on a product page, up from 21%
2x a month: working sessions with their team, for a year.
Site-wide organic traffic grew about 9 percent over the same period. The timing lines up with the work, but wider interest in tinned fish helped too; this doesn't isolate what any single change caused.
What could this look like for your brand?
Let's talk about what your team wants to build and how I can help.
Let's Connect →