How Trove was put together.

There is a category of independent retail that has been quietly underserved by the recommendation industry for fifteen years. The big platforms were built for the unit-volume end of e-commerce. The small platforms were built to keep the lights on. Almost nothing was built for boutiques that have a catalogue, a point of view, and a customer who notices the difference.

We started Trove Data Ltd because we kept being told, by buyers and managing directors at independents we knew, that the choice on offer was "no recommendation system, or a recommendation system that ruins our edit." That binary did not need to exist; the modelling, the engineering and the editorial discipline were all available. What was missing was someone willing to put them together for the right audience.

The company was incorporated on the 6th of February 2025. The first atelier cohort went live in late summer 2025. Nineteen boutiques have joined the programme since. Three slots remain open for the next intake.

The bench

The firm is run by two directors, who between them combine modelling, editorial commerce, retail operations and engineering leadership. We are deliberately not maintaining a catalogue of staff biographies; the public-facing surface of a curation firm should be its work, not its CVs. Cohort clients are introduced to the analysts and engineers responsible for their account at the start of onboarding. Press enquiries should be directed to press@troveboutique.net.

How the firm is funded

Trove Data Ltd is privately held and funded from operating revenue. We have no external investors. The decision not to raise capital at this stage is deliberate: it preserves the freedom to make the editorial-discipline decisions — particularly around taste-versus-engagement modelling, and around refusing provenance claims that cannot be evidenced — that a venture-backed competitor would, in our view, struggle to make.

Where we sit, in the market

We are not a competitor to the enterprise recommendation platforms. We do not run merchandising for retailers with eighty thousand SKU and a dozen warehouses. The fit is most natural for boutiques in the three-hundred-to-thirty-thousand SKU range, with a clear point of view, that have so far either gone without a recommendation system or run one that flattens their edit. There are several thousand such boutiques across Europe. We can work usefully with several dozen of them at any one time. The intake is paced accordingly.

The flattest category page is the dropdown filter grid. The most successful one is the edit. Our work begins by taking the second seriously, at scale, without losing the texture that makes it work.— Trove Boutique, Editorial Commerce Playbook

What we will not do

We will not greenwash. We will not publish a provenance claim we cannot evidence — even when the merchant has requested it, and even when the alternative is to leave the field blank. We will not optimise for engagement metrics at the expense of post-purchase satisfaction. We will not run two-tier service for clients who pay more; the atelier cohort is one cohort and the standards are one set of standards. None of these positions are negotiable, and none of them require any explanation past the sentence in which they are stated.