Loyalty and customer value · FLO Group · 2025
Loyalty that looks past the receipt
A points-for-purchases programme that measured spending and ignored engagement.
In short
Situation: The loyalty programme of an omnichannel retailer worked like most: points per purchase, rewards per threshold. It measured transactions and said nothing about why customers came back, or why they stopped moving between app, web and store. Decision: Redefine the programme as an engagement system: dynamic tier-based segmentation instead of fixed spending bands, incentives built on behaviour (bounded rationality, nudges) instead of flat discounts, and interactions triggered by engagement signals across channels. Result: Implemented in production across channels and published as a conference paper in April 2026. The paper reports that customer value was more closely associated with sustained engagement and cross-channel continuity than with transaction volume alone (observational, no control group). My role: Product owner: framework design, segmentation and incentive rules, rollout with CRM, marketing and engineering teams.
- My role
- Product owner
- Period
- 2025
- Scale
- Omnichannel: app, web, stores
- Key figure
- 3 mechanisms, one framework
Situation and complexity
Loyalty programmes are easy to build and hard to make matter. Points per purchase reward people who were going to buy anyway and say nothing to the customer who browses in the app, buys in the store and never connects the two.
The question we worked on was which behaviour we actually wanted, and why a customer would choose it.
Options on the table
- Raise the reward rate
Rejected. It moves margin to customers who already buy and changes no behaviour.
- Buy a loyalty platform and run campaigns
Rejected. Campaigns are one-off; the goal was a system that reacts to behaviour continuously.
- A behaviour-based framework built into the platformChosen
Segmentation, incentives and interactions shaped by what customers do across channels.
The decision and how it was organised
The framework has three parts. Dynamic tier-based segmentation moves customers between tiers on engagement as well as spend. Incentive design borrows from behavioural economics, using bounded rationality, defaults and nudges in place of flat discounts. Engagement-triggered interactions reach the customer when a signal appears (a store visit, a returned item, a lapsed session), not when a campaign calendar says so.
Because the mechanisms run as system rules, they scale. The same logic serves every channel and is tuned from data instead of being rebuilt for each campaign.
Result
Commercial figures are confidential. The published paper reports directional outcomes from a before-and-after comparison without a control group.
Public record: Paper: 10th International Mediterranean Scientific Research and Innovation Congress, April 2026, pp. 193–204
What went wrong, what I learned
Intent: Switch every channel to the new mechanics in one release.
What went wrong: Tier movement was harder to explain to customers than to build, and the first messages had to be simplified.
What changed afterwards: Tier communication came down to one rule a customer could repeat, and the rollout was staged channel by channel.