Skip to main content
Enactor Agora
Retail Knowledge > Customer Segmentation & Targeting Strategy
Retail Knowledge

Customer Segmentation & Targeting Strategy

The retailers generating the highest returns from their promotional budgets are not those spending the most - they are those spending on the right customers, at the right moment, with the right offer.
Last updated: 22 March 2026Author: Richard Dodd, Enactor Strategy & InnovationCategory: Retail Knowledge
customer-segmentationrfm-analysiscustomer-lifetime-valuetargeting-strategyretail-strategypersonalisation
40%
More revenue generated by companies excelling at personalisation
5.2×
Average ROI on loyalty programme investment
25-95%
Profit uplift from a 5% increase in customer retention

Executive Overview

Customer segmentation is the practice of dividing a retailer's customer base into groups that share meaningful behavioural or value characteristics, enabling promotional investment to be directed where it will generate the greatest incremental return. Without segmentation, every customer receives the same offer - an approach that simultaneously over-invests in customers who would have purchased anyway and under-invests in customers who could be moved by a well-timed incentive.

The economics are compelling. The Pareto principle manifests consistently in retail: roughly 20% of customers generate approximately 80% of revenue. A 5% increase in customer retention can boost profits by 25-95%1 (Bain & Company). Companies that excel at personalisation - the downstream output of effective segmentation - generate 40% more revenue from those activities than peers who do not2 (McKinsey, 2024). Loyalty programme members are 62% more likely to spend more and purchase more frequently than non-members.

Yet the most common point of failure in retail segmentation is not the analysis - it is the connection between the segment model and the promotional execution system. A segment can be identified, an offer designed, but unless the POS system can validate the customer's segment membership at the moment of transaction and apply the corresponding mechanic, the segmentation remains a marketing exercise rather than a commercial capability.

In this article:

  • The three core segmentation frameworks: , , and
  • How customer segments translate into actionable POS promotion mechanics
  • The role of CDPs and AI in automating segmentation at scale
  • UK GDPR, ICO consent obligations and the Data (Use and Access) Act 2025
  • How Enactor's loyalty group architecture bridges segmentation strategy and in-store execution

Market Context

The global () market was valued at $2.65 billion in 2024 and is projected to reach $12.96 billion by 2032, growing at a CAGR of 21.7%4 (Fortune Business Insights, 2025). The retail and e-commerce sector is among the leading adopters, driven by the need to unify transaction data from POS, web, app, and loyalty touchpoints into a single, actionable customer view.

Despite this investment, the gap between segmentation capability and segmentation execution remains large. Institute's 2024 member survey found that only 64% of deployed CDPs deliver significant value5. McKinsey research found that effectively none of approximately 50 senior marketing leaders at Fortune 500 companies could clearly measure the ROI on their marketing technology investments. The problem is rarely data - it is the connection between data and the systems that execute against it.

The personalisation imperative is clear in UK retail. Over 70% of consumers now expect personalised communication tailored to their preferences3 rather than demographics (Optimove, 2025). Yet the ICO's 2025 monitoring of the UK's top 200 websites found that 30% were setting advertising cookies without consent - a compliance failure rate that signals segmentation ambition running ahead of governance infrastructure.

The risk of poor segmentation is not neutral. Broad, undifferentiated promotional programmes drive promotion dependency, erode brand equity for premium retailers, and generate significant incremental cost with diminishing incremental return. Retailers who have moved to segment-specific targeting - knowing which customers to invest in, which to win back, and which to allow to lapse - consistently protect margin more effectively.


How It Works

Framework 1: Segmentation

RFM - Recency, Frequency, Monetary value - is the most widely used and most immediately actionable customer segmentation framework in retail. Each customer is scored 1-5 across each dimension, generating up to 125 unique score combinations that are grouped into operational segments. Its enduring appeal is simplicity: it can be applied to any transaction dataset, requires no specialist software, and produces segments that are directly actionable through promotion mechanics.

SegmentRFM profileCommercial strategy
ChampionsHigh R, High F, High MReward and co-create; use as social proof; protect at all costs
Loyal customersHigh F, Medium-High M, variable RMember-exclusive offers; tier benefits; no deep discounts
Potential loyalistsHigh R, Medium F, Medium MAccelerate frequency with points multipliers and near-miss alerts
Recent customersHigh R, Low F, Low MSecond purchase is the critical milestone - onboarding sequence
At-risk customersLow R, previously High F/M with loss-aversion framing and genuine urgency
Lapsed customersVery Low R, previously activeReactivation offer if cost is low, otherwise allow to lapse
Low value / inactiveLow across all dimensionsMinimal investment; watch for unexpected reactivation signals

The value of RFM lies not in the scoring arithmetic but in the segment-specific action it enables. A Champion customer who begins to lapse needs a qualitatively different intervention from a Recent customer who has not yet made a second purchase. Those interventions must be configured as discrete promotions, vouchers, or loyalty mechanics at POS level - not just identified as strategic priorities in a spreadsheet.

RFM's Limitation: What It Doesn't Capture

RFM is a backward-looking framework - it describes what customers have done, not why they did it, and not what they are likely to do next. It does not capture product affinity, channel preference, price sensitivity, or life stage signals. In mature programmes, RFM is enriched with behavioural signals - category purchase patterns, promotion response type, seasonal cadence - to improve targeting precision. The Recency dimension is particularly misleading in categories with long natural purchase cycles, where low recency may reflect the category rather than lapsing behaviour.

Framework 2: ()

is the total net profit expected from a customer over the full duration of their relationship with the retailer. It is the primary strategic lens for deciding how much to invest in acquiring, retaining, and developing individual customer segments.

Acquiring a new customer costs approximately five times more than retaining an existing one. Loyalty programme members report an average 5.2× ROI on programme investment (Rivo, 2025). The of 3:1 is the widely cited minimum viable benchmark for marketing investment.

CLV-based segmentation is forward-looking: it estimates future value and uses that estimate to prioritise promotional investment, rather than simply describing past behaviour.

CLV tierCharacteristicsInvestment strategy
Tier 1 - High lifetime valueFrequent, high-spend, long-tenureProtect aggressively; personalised service; no margin-diluting discounts
Tier 2 - Growth potentialRecent, engaged, lower current spendInvest to accelerate frequency and AOV; near-miss alerts, cross-sell
Tier 3 - TransactionalInfrequent, discount-responsivePromote cautiously; measure incrementality carefully
Tier 4 - At-riskDeclining frequency, previously valuableWin-back investment justified; strong urgency framing
Tier 5 - Lapsed / unprofitableInactive or generating negative marginMinimal investment; reactivation only at very low cost

AI-driven CLV prediction improves forecast accuracy by 25-40% over traditional rule-based models8, with deep learning approaches that incorporate behavioural segmentation outperforming simple regression significantly (AJHSSR, 2025).

The Measurement Gap: Most Retailers Cannot Accurately Calculate CLV

Research from SCAYLE (2024) found that 76% of companies cite CLV as an important metric, but only 42% feel equipped to measure it correctly7. The most common error is using gross revenue uplift rather than net incremental profit - which makes all promotions appear effective because it ignores customers who would have purchased regardless. A reliable CLV baseline requires at least 12 months of transaction data with seasonal normalisation applied.

Framework 3:

Behavioural micro-segmentation moves beyond transaction history to segment customers on observed patterns: category affinity, promotional response type, channel preference, purchase cadence, and lifecycle signals. It is the approach that most closely replicates what a skilled retailer's intuition does - at scale and in real time.

Common micro-segment archetypes in retail include:

ArchetypeSignalTargeted mechanic
Category loyalist80%+ of spend in one departmentCross-category voucher to introduce adjacencies
Discount hunterPurchases only on promotionShift to non-price mechanics; reduce discount depth
Lapsed seasonalActive only in one calendar windowPre-season reactivation trigger
Near-tier customerWithin 10% of next loyalty tier thresholdTier-acceleration targeted voucher
High-frequency low-AOVMany visits, low basket valueSpend threshold mechanics; bundle offers
New member, first purchaseEnrolled in last 30 daysSecond-visit incentive sequence

Micro-segmentation at scale requires either a CDP capable of building and refreshing these groups automatically, or an AI agent capable of identifying cohorts from transaction analytics and assigning group membership through the POS system's API.

The

For segmentation to generate commercial return, four things must be simultaneously true: the customer must be identified at POS (via loyalty card, app, or payment token); their segment membership must be current and held in the POS system; the promotion mechanic must be configured to activate based on that membership; and the offer must be appropriate for that segment.

The majority of UK retailers currently fail on at least two of these four dimensions - most commonly because their CRM segmentation and their POS promotion engine operate in separate systems with no live integration.

Anonymous vs. Identified Customers

A persistent challenge is the proportion of transactions completed by unidentified customers - those paying by cash or card with no loyalty involvement. In UK grocery this can be 40-60% of transactions. Without identification, these customers cannot be individually segmented or targeted.

Two approaches address this. PAR (Pseudonymous Analytics at the Register) assigns a probabilistic pseudonymous identifier to anonymous payment transactions, enabling cohort-level analysis without PII. Loyalty enrolment programmes incentivise identification at POS. The ICO permits PAR-style pseudonymous under where no individual is singled out; individual targeted communications require explicit consent under PECR.


Costs and Considerations

Cost layerNotesFrequency
Data infrastructureTransaction data warehouse, CDP, or analytics environment to build and maintain segment modelsAnnual licence + implementation
Segmentation modellingRFM is straightforward; CLV and ML-based micro-segmentation require data science resource or specialist toolingOngoing maintenance
Loyalty scheme operationMember identification, points management, tier maintenancePer transaction
GDPR / ICO complianceData Protection Impact Assessment; consent management for marketing communicationsPer programme and annually
Promotional configurationSegment-specific promotions must be individually configured and QA'd in the POS systemPer promotion
Measurement infrastructureWithout incrementality measurement, segment-targeted spend cannot be justifiedOngoing

What is free: RFM segmentation can be performed on any transaction dataset with no specialist software. The analytical framework is free; the cost is in data infrastructure and execution systems.

What is not optional: Under UK GDPR, personal data processed for marketing purposes requires a valid lawful basis. PECR consent is required for electronic direct marketing. A DPIA is mandatory for AI-driven segmentation. The ICO's updated guidance under the Data (Use and Access) Act 2025 is under active review and all programmes should be assessed against current ICO publications before activation.

UK GDPR, ICO and the Data (Use and Access) Act 2025

The ICO published updated 'consent or pay' guidance in January10 2025, and the - with the ICO's guidance subsequently placed under review12. For loyalty-based in-store personalisation, is often the applicable lawful basis; for email and SMS campaigns, PECR consent is required. AI-driven segmentation carries additional transparency and oversight obligations: organisations using AI for segmentation must document how it works, ensure human oversight, and collect customer consent before using personal data in AI systems. The ICO extended its monitoring from the top 200 to the top 1,000 UK websites for online tracking compliance in 2025.


The Business Case

Core Argument

Segmentation improves the return on every pound of promotional investment by concentrating spend where the marginal response is highest. A Champion customer does not need an incentive to buy - giving them one wastes margin. A Potential Loyalist with two purchases in three months, buying the right categories, can be converted to a high-frequency loyal customer at low cost if the right mechanic fires at the right moment.

Revenue Upside

Businesses leveraging AI to orchestrate customer journeys report 33% higher CLV on average3 (Optimove, 2025). Personalisation marketing boosts revenue by 5-15%, reduces by up to 50%, and improves marketing ROI by 10-30%2 (McKinsey). Loyalty programmes deliver an average 5.2× ROI6 (Rivo, 2025), with 85% of engaged customers participating in programmes that generate a 25% improvement in retention.

Cost Reduction

The primary cost reduction is promotional efficiency: fewer promotions required to achieve the same volume response, with each promotion going to customers most likely to respond and least likely to have purchased without it. CLV-based investment reallocation removes spend from Tier 3/4 customers where incrementality is low and concentrates it in Tier 1/2 where lifetime value justifies investment.

Risk of Inaction

Retailers operating without meaningful segmentation are vulnerable to two compounding risks: promotional escalation (each successive campaign must be deeper than the last to maintain response as customers habituate) and competitive targeting (analytically sophisticated competitors are identifying and investing in their most valuable customers - some of whom shop across multiple retailers).

Indicative Business Case Model

Illustrative only, based on published industry benchmarks. Actual results vary by retailer scale, data maturity, and execution quality.

InitiativeIllustrative investmentIllustrative benefitPayback period
RFM segmentation from existing data£5,000-£15,00010-20% improvement in promotional response rate3-6 months
Loyalty enrolment programme£20,000-£50,00020-40% reduction in unidentified transaction proportion6-12 months
CLV-based investment allocation£10,000-£25,0005-15% reduction in promotional cost for equivalent revenue outcome6-12 months
AI-driven micro-segmentation£30,000-£80,00033% CLV uplift for AI-orchestrated journeys (Optimove benchmark)12-24 months
Group-targeted POS promotions£5,000-£15,000 configurationDirect activation of segment strategy at POS without CDP dependency3-6 months

Key Risks and Mitigations

RiskLikelihoodMitigation
Segment model drift - groups become staleHigh without governanceMonthly refresh schedule; automated staleness alerting
ICO/GDPR enforcement action on consentMedium-High without DPIAsDPIA for all AI-driven segmentation; document lawful basis per programme
POS group membership stale or empty at activationMediumShannon group management: verify population before promotion goes live
CLV mismeasurement - over-investing in wrong segmentsMediumUse incrementality-adjusted CLV; require 12 months minimum data
Segmentation perceived as surveillance rather than serviceMediumMake segmentation visible through the loyalty programme - customers should feel known, not profiled

Enactor and This Topic

Enactor's loyalty and promotions architecture is designed to bridge the that most UK retailers currently experience. The platform separates customer identification (loyalty), segment assignment (group membership via REST API), and promotional execution (group-targeted mechanics at POS) into distinct, configurable layers.

🏆
Loyalty Programme with Tier Structure

Fully configurable loyalty schemes with tier progression strategies: manual, points balance, annual spend, or custom. Tier-based promotion mechanics activate automatically based on the customer's identified tier at POS, with no external intervention required for tier-linked offers.

Available
🏆
Loyalty Groups for Dynamic Micro-Segmentation

Groups are flat, programmatically assigned identifiers assigned to customers via the Enactor REST API by any external system - CRM, CDP, data warehouse, or AI agent. Unlike tiers, groups carry no hierarchy and can be assigned and revoked dynamically, enabling real-time behavioural micro-segmentation to drive POS promotion execution without rebuilding promotions.

Available
🔧
Group-Targeted Promotions

Promotions are configurable to activate only for customers holding a specified group membership, enabling segment-specific mechanics - percentage discounts, spend thresholds, near-miss alerts, targeted vouchers - to fire conditionally based on segment membership validated at the transaction moment.

Available
🤖
Shannon as Group Management Agent

Shannon acts as an active group management agent: identifying qualifying cohorts from Estate Manager transaction analytics, assigning group membership via the REST API, verifying group populations before promotion activation, and clearing stale memberships. This closes the loop between segmentation analysis and POS execution without manual CRM intervention.

In development · PRD018
📊
PAR - Pseudonymous Analytics at the Register

Assigns a probabilistic pseudonymous identifier to anonymous payment transactions, enabling tier-2 without PII. Enables segmentation-informed decision-making for the unidentified customer proportion, operating within the ICO's legitimate interests framework for pseudonymous data processing.

In development · PRD006
🎟
Targeted Voucher with Serial Tracking

Serialised single-use vouchers issued to specific customers or groups, enabling segment-specific personalised offers with full audit trail and redemption tracking.

Available
Start with Groups, Not a New Platform

Many retailers assume meaningful segmentation requires a CDP investment before they can begin. Enactor's group architecture allows retailers to execute segment-targeted promotions from existing CRM or analytics data immediately. The simplest starting point is assigning a group identifier to a list of at-risk customers from an existing RFM export, configuring a for that group in Enactor, and measuring incremental response. This generates measurable ROI from existing data while building the business case for deeper segmentation investment.

For scoping discussions on loyalty scheme configuration, group-based targeting, and Shannon's group management capability, contact Enactor Professional Services.


References

  1. Bain & Company / Reichheld, F. The Loyalty Effect. Harvard Business School Press, 1996. Widely cited: 5% retention increase drives 25-95% profit uplift.
  2. McKinsey & Company. The Value of Getting Personalization Right - or Wrong - Is Multiplying. 2024. https://www.mckinsey.com/capabilities/growth-marketing-and-sales/our-insights/the-value-of-getting-personalization-right-or-wrong-is-multiplying
  3. Optimove. 25 Customer Engagement Trends for 2025. 2025. https://www.optimove.com/blog/top-customer-engagement-trends
  4. Fortune Business Insights. Market Size, Share, Trends & Forecast 2024-2033. 2025. https://www.fortunebusinessinsights.com/industry-reports/customer-data-platform-market-100633
  5. CDP Institute. Customer Data Platform Market Predictions for 2025. 2025. https://www.cdpinstitute.org/cdp-institute/customer-data-platform-market-predictions-for-2025/
  6. Rivo. Customer Lifetime Value (CLV): 25 Ecommerce Calculation & Benchmark Statistics. 2025. https://www.rivo.io/blog/ecommerce-calculation-benchmark-statistics
  7. SCAYLE. Customer Lifetime Value (CLV): Boost Your Marketing ROI. 2024. https://www.scayle.com/library/blog/customer-lifetime-value/
  8. AJHSSR. Optimizing Customer Lifetime Value Prediction Using Deep Learning and Behavioural Segmentation. 2025. https://www.ajhssr.com/wp-content/uploads/2025/07/J25907123131.pdf
  9. Information Commissioner's Office. Marketing and Data Protection in Detail. 2025. https://ico.org.uk/for-organisations/advice-for-small-organisations/direct-marketing-and-data-protection/marketing-and-data-protection-in-detail/
  10. Information Commissioner's Office. Consent or Pay Guidance. January 2025. https://ico.org.uk/for-organisations/uk-gdpr-guidance-and-resources/online-tracking/consent-or-pay/about-this-guidance/
  11. Force24. GDPR, Still a Thing in 2025: What UK Marketers Need to Know. January 2025. https://force24.co.uk/gdpr-still-a-thing-in-2025-what-uk-marketers-need-to-know/
  12. Taylor Wessing. Taking Control of Online Tracking: The ICO's Focus for 2025. 2025. https://www.taylorwessing.com/en/global-data-hub/2025/spotlight-on-the-uk-data-landscape/gdh---taking-control-of-online-tracking
  13. Braze. Understanding RFM Segmentation: Marketers' Guide. Updated November 2025. https://www.braze.com/resources/articles/rfm-segmentation
  14. Saras Analytics. Ecommerce Customer Value: How to Calculate & Improve It. 2025. https://www.sarasanalytics.com/blog/ecommerce-customer-lifetime-value

Enactor Retail Knowledge - published March 2026. This article draws on publicly available research and platform documentation. Market statistics are sourced from named third-party publications and do not represent Enactor's own research. Pricing figures are indicative based on publicly available information at time of publication and should be verified directly with providers.