The structural shift: from points to precision
The old loyalty model was simple. Collect points, redeem rewards, repeat. It worked because shoppers valued predictability and retailers valued footfall. But the economics of that model have been eroding for several years, and in 2026 the erosion is accelerating.
The programmes that are pulling ahead are not the ones offering the deepest discounts. They are the ones using purchase history, basket composition, and behavioural timing to deliver offers that feel personally relevant rather than randomly promotional. Tesco Clubcard, Kroger Boost, and Albertsons for U represent the current reference set. Each has moved, at different speeds, toward AI-driven targeting at the SKU level. The ambition is not a better coupon. It is a better prediction of what a specific household will find genuinely useful on a specific shopping occasion.
That shift matters for your commercial planning because it changes who owns the customer insight. When loyalty runs on points, the brand and the retailer share roughly equal access to shopper behaviour. When loyalty runs on AI-generated individual offers, the retailer's data asset compounds faster than anything a brand trade fund can match. As the Opinion piece [O2] in this series argued directly, brands that participate in loyalty programmes as passive fee-payers are funding the infrastructure that makes them replaceable.
What the data says about shopper intentionality
The shopper environment in 2026 is not making loyalty easier to earn. Several overlapping pressures are making shoppers more deliberate and harder to influence through blunt promotional mechanics.
Food-at-home price inflation could exceed 4 percent in 2026, according to an FMI briefing covered by Food Business News. The April Consumer Price Index showed food-at-home prices rising 0.7 percent month to month and 2.9 percent year over year. The Producer Price Index, which tracks costs further up the supply chain, jumped 1.4 percent month to month and 6 percent year over year in April. When shoppers feel that kind of pressure, they do not become less loyal. They become more selective. They stay loyal to the retailers and brands that make them feel their money is well spent, and they switch away from the ones that feel indifferent to their situation.
The Kantar BrandZ analysis, reported by FoodNavigator-USA, adds important texture here. Brands that hold pricing power through inflation combine three things: they are seen as meaningfully different, they are emotionally connected to the consumer, and they are culturally relevant in the moment. None of those three factors is delivered by a points programme. All three can be reinforced by a well-designed personalisation engine that knows what a household cares about and reflects that back in the offer it makes.
How AI is changing the offer architecture
The most interesting development in loyalty right now is not the programme mechanics. It is the data infrastructure sitting underneath them. AI-driven flavour and preference segmentation, already being used in product development, is beginning to flow into promotional targeting as well.
Research covered by FoodNavigator shows that companies including PepsiCo and Nestlé are working with AI firm Foodpairing to link flavour preferences to demographic and geographic signals. The precision, as Bernard Lahousse, co-founder of Foodpairing, put it, has shifted "from a marketing story to a measurable engineering problem." Age is one of the clearest signals in that data: younger consumers show narrower, more distinct flavour preferences than older cohorts, which means a personalisation engine calibrated to household demographics can target promotional offers with a level of relevance that category-level coupons cannot approach.
The implication for loyalty programme design is direct. If a retailer knows a household skews young, buys high-fibre and functional products, and responds to novelty formats, an AI offer engine can surface the right SKU at the right moment. That is a fundamentally different value proposition than a ten-percent-off voucher on any cereal in the aisle.
The GLP-1 signal and what it means for personalisation
One demand shift that loyalty programme managers should be tracking closely is the GLP-1 effect on snacking behaviour. Research covered by Bakery and Snacks found that mood, specifically sadness and stress, was a major driver of indulgent snacking behaviour. GLP-1 medications appear to dampen that dopamine-driven impulse, which means a growing subset of shoppers is moving from emotionally-driven snack purchases toward more intentional, functional choices.
A separate study on an oral GLP-1 called orfoglipron, reported by FoodNavigator, found that patients who switched from injectable weight-loss drugs maintained significant weight loss for over a year. Those who had been on tirzepatide kept off 75 percent of the weight they had lost, compared with 49 percent in a placebo group. Those on semaglutide kept off 79 percent, compared with 38 percent in the placebo group. The appetite-suppression window is widening, not narrowing.
For loyalty programme design, this is not a category-level observation. It is a household-level signal. A programme that can identify GLP-1-adjacent behaviour in a basket, shifts toward high-protein and high-fibre products, smaller portion formats, fewer impulse confectionery items, and begin surfacing relevant offers in those functional subcategories will deepen loyalty with exactly the shopper cohort that is growing fastest. A programme still optimising for confectionery attach rates in that same household will feel tone-deaf and accelerate churn.
Where brands need to act this week
The strategic gap most brands face is not a lack of loyalty programme participation. It is a lack of data access within the programmes they already fund. You are likely contributing trade spend to Tesco Clubcard mechanics, Kroger Boost activations, or Albertsons for U targeted offers without receiving back the household-level insight that would let you evaluate what that spend is actually doing.
The brands closing that gap are the ones renegotiating their loyalty participation as a data partnership rather than a promotional placement. That means asking, specifically, for post-campaign basket analysis by household cohort, not aggregate redemption rates. It means requesting segmentation cuts that show you which households increased purchase frequency versus which ones redeemed once and did not return. And it means building joint business plan KPIs around loyalty data outputs, not just distribution and promotional compliance metrics.
The Kantar finding that emotional connection and cultural relevance, not discounts, protect margins in volatile markets is a useful frame here. A loyalty programme that makes a shopper feel understood is delivering emotional connection at scale. A programme that sends the same voucher to every Clubcard holder in a postal code is not.
What to watch in the next 90 days
Three things are worth monitoring closely as this landscape develops.
First, watch how Kroger Boost and Albertsons for U each evolve their AI offer cadence following their respective planning cycles. Both programmes have signalled moves toward more granular household targeting, and any public commentary from earnings calls or investor presentations will give you a read on how fast the personalisation engine is maturing.
Second, track the GLP-1 penetration curve in your core shopper base. The oral GLP-1 data reviewed above suggests appetite suppression is becoming a durable feature of the shopping population, not a temporary one. Loyalty programmes that are not yet segmenting on health-intentionality signals will be catching up to retailers that started this work 12 months ago.
Third, pay attention to the intersection of generative AI and loyalty personalisation. As FoodNavigator-USA reported, consumers are increasingly using tools like ChatGPT and Claude to find product recommendations. A shopper who asks an AI assistant which protein bar to buy at Kroger is effectively bypassing the shelf and the promotional mechanic entirely. Loyalty programmes that feed structured product data into those recommendation engines will have a channel that blunt promotional spend cannot replicate.
The programmes that win the next three years will not be the ones with the most points. They will be the ones that make the shopper feel, precisely and consistently, that the retailer knows them. Brands that help build that feeling rather than just buying space inside it are the ones that will still be in the basket when the dust settles.