Internet sales account for a little over a quarter of all retail spending in Great Britain, according to the Office for National Statistics. Behind each of those transactions sits a stack of software for checkout, payments, fraud screening, fulfilment and reporting. Very few retailers own all of it. The question of which pieces to build and which to rent has become one of the defining strategic calls in digital commerce – and it is no longer confined to retail.
Ocado and the economics of licensing
Ocado offers the clearest British example of a retailer turning its infrastructure into a product. Through the Ocado Smart Platform, the group licensed its software and robotic warehouses to grocers including Morrisons, Groupe Casino, Sobeys, ICA and, in a 2018 deal, Kroger, which also took a 5% stake in Ocado for £183m.
For the partners, the attraction was speed: years of R&D bought rather than replicated. For Ocado, it was recurring fees on capital it had already sunk. But the arrangement cuts both ways. Kroger later shut three Ocado-powered sites and scrapped plans for another, with Ocado due to receive around $350m in compensation. As exclusivity deals with most partners expired, Ocado said it would take its technology back to market from 2026. Licensing reduces upfront cost; it does not remove dependency, on either side.
White labels, marketplaces and the shared back end
The high street has long understood white labelling through own-brand goods made by third-party manufacturers. The digital version works the same way: a brand controls the customer relationship while a supplier runs the machinery underneath. Marketplace models push this further. As Retail Times has reported, Regatta Great Outdoors joined the “Brands at M&S” platform, selling through M&S’s digital storefront rather than building that reach alone.
The common thread is that the front end differentiates while the back end standardises. Payment routing, identity checks and analytics look broadly similar from one business to the next, so it rarely pays to rebuild them.
Where regulation raises the stakes
The build-or-rent decision gets sharper in sectors where compliance is part of the product. Online gambling is a useful test case because the economics are public. The Gambling Commission’s industry statistics put gross gambling yield for the remote casino, betting and bingo sector at £7.8bn in the year to March 2025, up 13.1% on the previous year.
Margins, however, have just been squeezed. Under changes set out in a government policy paper, Remote Gaming Duty rose from 21% to 40% on 1 April 2026, with a new 25% rate for remote betting due in April 2027. When two-fifths of gross yield goes to the Exchequer, the cost of running proprietary infrastructure becomes harder to justify for anyone without scale. That is one reason many newer brands launch on a turnkey online casino solution that bundles game aggregation, payment integration and back-office reporting, putting their own budgets into marketing and customer service instead.
Payments are a particular pressure point: card issuers and acquirers often treat gambling as high-risk, so integrating several providers and alternative methods is slow, specialist work that few start-ups want to do from scratch. Renting the stack does not outsource accountability, though. The licence holder remains answerable to the regulator for anti-money-laundering controls, age verification and safer-gambling checks, whichever supplier wrote the code.
Questions worth asking before signing
Whether the platform runs groceries or games, the same due-diligence points recur:
- Data ownership. Who holds customer data, and in what format can it be exported?
- Exit terms. Ocado’s experience with Kroger shows that contracts need clear provisions for scaling back as well as scaling up.
- Compliance split. Which controls does the supplier operate, and which remain the client’s responsibility?
- Pricing model. Revenue share aligns incentives early on but can become expensive as volumes grow; fixed fees do the reverse.
The trade-off doesn’t go away
The platform model has matured to the point where almost any digital business can launch on rented infrastructure. What it cannot do is decide where a company’s edge lies. Kroger’s retreat showed that buying a proven system does not guarantee the demand to fill it. Operators in taxed, regulated markets face a version of the same calculation with less room for error: the parts customers notice are worth owning, and the parts they never see are usually worth renting.
