A Linear Data Model. In a Circular Economy.
The circular economy argument in fashion has been made compellingly for years. Resale is growing faster than primary retail. Rental platforms are finding audiences that ownership models never reached. Repair, refurbishment, and take-back are moving from niche to standard practice. The direction is clear. The business case, in many segments, is proven.
What is less often discussed is what these models require to operate at scale — and where most organisations discover, partway into building them, that the foundation isn't there.
The data model that underpins most fashion organisations was built around a single assumption: the brand's relationship with a product ends at the point of sale. Data flows in one direction — from brief to sourcing to production to sale. Each stage captures what it needs and passes the product on. Nothing accumulates. Nothing follows the product once it leaves.
That assumption is not a technical oversight. It was the correct model for a linear business. It is architecturally incompatible with a circular one — not as a matter of degree, but of direction. Circularity requires a data model in which the product carries its own record: updated at each stage of its life, persistent across ownership changes and use cycles, accessible to whoever is making the next decision about it. Not a snapshot at manufacture. A living record.
Most organisations have not built this. Not because the concept is new, but because building it requires decisions that cut across the linear model's assumptions at every point — about who owns the record, who updates it, what system holds it when the product changes hands.
The gap becomes visible the moment a circular model tries to operate. A product re-enters the value chain — through resale, repair, or take-back. It arrives with a history. That history is the operational basis on which every decision about its next life depends.
What is it made of, exactly — not at category level, but specifically enough to determine whether it can enter a defined recycling stream? Has it been repaired, and if so with what materials? Does anything in its history affect a compliance declaration still attached to it?
These are not philosophical questions. For a resale platform, a repair service, a take-back programme — they are the questions that determine whether the model is viable or whether every product that re-enters the system has to be assessed from scratch. In most organisations, the data to answer them does not exist in any retrievable form. The product arrived. Its history did not.
The regulatory context sharpens this from aspiration to obligation. Extended Producer Responsibility schemes are placing lifecycle accountability on brands that the linear model was never built to satisfy. The PPWR, applying from August 2026, places EPR obligations on producers to track and demonstrate end-of-life outcomes for packaging. It is a structural precedent for what is coming for products themselves. The DPP's trajectory toward full circular economy integration by the mid-2030s makes the destination explicit: a product record that persists across the product's entire useful life, not just its first sale.
The brands building circular operations now are not waiting for that obligation. They are discovering, practically, that the data infrastructure circularity demands and the data infrastructure the DPP requires are the same thing viewed from different angles. A product record maintained across the full lifecycle — manufacture, resale, repair, end-of-life recovery — is both a passport and an operational asset. Built once, it serves both purposes. Built separately for each — a compliance record here, a circular operations database there — it doubles the cost and produces neither well.
The organisations making genuine progress are approaching this as an architecture decision, not a circular economy initiative. The question is not how to add data capability to an existing circular programme. It is how to build a data model that treats the product's full lifecycle as the unit of account from the outset — so that when a product re-enters the value chain, its history is already there.
That decision belongs earlier in the process than most organisations place it. At the brief stage, where material choices determine what will ever be knowable about the product's composition. At the sourcing stage, where supplier relationships determine how far back traceability will reach. At the product design stage, where construction choices determine whether disassembly is practical or theoretical.
Circular economy initiatives that start at the point of take-back are starting too late. The data that would make the product's second life viable was either built in at the beginning or it was not built at all.
The linear data model is not a technical legacy problem. It is a reflection of a business model assumption — that the product's story ends at the point of sale. That assumption is changing, driven by consumer behaviour, investor expectation, and regulatory obligation simultaneously.
The data model has to change with it. The organisations that make that change deliberately, as an investment in the infrastructure that both circularity and compliance require, will find the transition manageable. Those that arrive at circular operations with a linear data model intact will find themselves rebuilding under pressure — with circular business model revenues on one side of the ledger and data infrastructure costs on the other, compressing the margin that made the model worth pursuing.
The circular economy is not waiting for the data to be ready. The question is whether the data will be ready for it.
Michael Shea is a digital excellence advisor, non-executive director, and leadership coach working with organisations navigating the human and technical dimensions of digital transformation. He hosts The Aeolian Discourse and writes at The Aeolian.