Research & Data Science — GPASS: Evidence-based insights on digital loyalty for specialty coffee and local communities

Academic Research

What does the research say?

Peer-reviewed studies confirm that frictionless, app-free digital loyalty programs drive higher adoption, engagement, and retention.

2025

Bauat et al.

The Effectiveness of Loyalty Programs in Enhancing Customer Retention

Key Finding: Well-managed loyalty programs significantly improved repeat purchases and customer satisfaction.

Simple point earning and redemption systems 'significantly support long-term customer relationships and business success.'

Read study
2020

Son et al., Information Systems Research

When Loyalty Goes Mobile: Effects of Mobile Loyalty Apps

Key Finding: Digital loyalty platforms led to higher customer spending, more frequent purchases, and more active reward redemptions.

Digital loyalty platforms drive greater customer engagement by making participation more convenient and accessible.

Read study
2022

Naula, Aalto University

Customer Loyalty: Effects of Discount and Ease-of-Use

Key Finding: Straightforward, easy-to-use loyalty program interfaces led to significantly higher customer loyalty.

Simplicity in program design boosts engagement – customers are more loyal when the program is effortless to use.

Read study
2021

Wharton & Verde Group

Understanding the Boomerang Effect of Loyalty Programs

Key Finding: Shoppers who encountered friction were 35% less loyal than those with frictionless experiences.

Minimizing barriers – avoiding extra apps or complex processes – keeps loyalty programs seamless and customers returning.

Read study
2023

Myftaraj & Trebicka

Impact of Loyalty Card Programs on Customer Behavior

Key Finding: Customers who frequently used loyalty cards demonstrated significantly higher loyalty levels.

Easily accessible digital loyalty cards that require no additional app foster greater customer loyalty and repeat patronage.

Read study

Reducing friction and emphasizing simplicity in digital loyalty programs leads to higher customer adoption, ongoing engagement, and improved retention.

Loading community data...

AI Infrastructure

We're building the infrastructure to make data science work for you

Behind every simple toggle your store sees, there's a powerful, modular AI system designed to learn, adapt, and scale -- so your loyalty program gets smarter over time without you lifting a finger.

A layered system built for evolution

Feature Layer

Collects and processes customer behavior signals: recency, frequency, spending patterns, and more.

Policy Layer

Pluggable decision-making models that can evolve from simple rules to advanced reinforcement learning.

Decision Engine

Real-time engine that selects the optimal action for each customer at every interaction.

Merchant Surface

A clean, simple interface. Merchants see results, not complexity.

What merchants see

Your store dashboard stays effortlessly simple. One toggle, clear metrics, zero data science jargon.

Smart AI Mode
Customers recovered18
Repeat visits lift+11%
Incentives triggered7
Est. revenue impact+€640
AI Intensity
Conservative
Balanced
Aggressive

What powers it

A modular, version-controlled AI backend that can scale from a single store to thousands -- without rewriting a single line.

Policy AbstractionSwap models seamlessly -- from global bandits to per-customer RL -- without touching production.
Full Context LoggingEvery decision is logged with full context: action, probability, model version, and timestamp.
Hierarchical LearningShare learnings across stores while adapting locally. Low-volume stores benefit from the network.
Offline EvaluationReplay past decisions with new models to measure uplift before deploying anything.
Budget SafeguardsHard caps on incentive frequency, minimum reward intervals, and budget guardrails built-in.

Designed to scale

The system evolves as the data grows. Each phase unlocks new capabilities without disrupting what already works.

Phase 1

Global Learning

A single intelligent model learns what works across all stores.

Phase 2

Hierarchical Models

Store clusters emerge. The model adapts strategies by business type and context.

Phase 3

Per-Store Intelligence

High-volume stores get dedicated models. Lower-volume stores remain pooled for stability.

Phase 4

Per-Customer AI

True personalization. Uplift models, causal inference, and meta-learners for each individual.

Built with safeguards

Hard incentive frequency capsBudget guardrails per storeMinimum reward intervalsNo infinite explorationVolatility protection for small merchants