Pinterest search architecture for a catalogue that compounds

Pinterest pays late. Across the catalogues I have run, a pin reaches its best month for outbound clicks around 90 to 120 days after it is published, and pins from two years back are still being saved today. Everything about how a catalogue should be built follows from that delay. Structure decided badly in March compounds the wrong way until the following spring, and no amount of publishing afterwards corrects it. This engagement sets the structure first: which parts of your catalogue get their own board, how a pin is built so the destination page is obvious before the click, and what the product feed has to send Pinterest for rich pins to stay valid. The decision this makes: Which products, recipes or rooms deserve to exist as a searchable destination on Pinterest, and which should never be pinned on their own. Most catalogues carry items that will never be searched for by name and only ever found through a category or a use case. Treating those as individual destinations spreads saves across pins that each stay below the threshold where Pinterest starts distributing them. Deciding what not to give a board to is the part clients find hardest and the part that changes the numbers. Sequence of work: First, the feed and the markup, because rich pins fail silently and a catalogue running on broken product data has been wasting its own distribution. Second, board architecture mapped to how people actually phrase the need in Pinterest search, which is rarely how your merchandising team names a collection. Third, pin templates: two or three layouts with fixed rules on where the text sits, what the overlay must say, and which image never works in the format regardless of how good the photography is. Fourth, the destination pages, since a save that lands on a page with no obvious next action is a save that never returns. What is measured: Saves and outbound clicks for the cohort of pins published inside the engagement window, read at 90 days from your own Pinterest Analytics, alongside rich pin validation status and referral sessions in your analytics platform. Impressions are reported but not used to judge the work, because impressions on Pinterest move with seasonality more than with anything I do. Who does the work: Rafael Costa. I have spent the last several years as the fractional search lead for consumer brands where visual discovery, not paid social, carries new customer acquisition. Positions I have taken and been paid to hold: telling a homeware brand that its best-performing board should be deleted because it was pulling in an audience that never bought; refusing a food client's request to pin every recipe on the site and building around eleven use-case boards instead, after which outbound clicks in the following season were higher than the year in which everything was pinned. I write the plan, I sit in the review, and I do not hand the account to anyone else. Scope boundary: I supply pin templates and the rules for using them. I do not produce pin graphics on an ongoing basis, and I do not run a scheduling tool on your behalf. No Pinterest Ads management, no influencer or creator sourcing, no product photography, no writing of the product copy on your own site. If your product feed is broken and nobody internally owns fixing it, that has to be resolved before the rest of this is worth buying, and I will say so before you pay rather than after. I turn this down when: The category has no visual search behaviour behind it. Pinterest is not a channel for developer tooling or B2B software, whatever the case studies claim, and I would rather lose the engagement than take money for that. When the buyer defines the deliverable as a monthly pin count, we want different things. When the brief is to recover a suspended account or to work around a policy decision by Pinterest, that is not work I do. And when the site cannot be changed within the engagement window, because the destination page is half of this and I will not pretend otherwise. Languages: English, Portuguese, Spanish, with catalogues run for United States, Canadian and Brazilian audiences.