20,495
vs avg +13%
01
Getting Started
How this report reads.
Before the numbers: why this report exists, what it is built on, where the RADAR framework comes from, and how to read the three stages ahead.
1 of 5 · why this exists
The premise
why this exists
This report exists for one reason: to improve decisions. Most companies don't struggle because they lack data; they struggle because they don't know where to focus or what to do next. Decisions end up driven by urgency, averages, or intuition instead of customer reality. What's missing isn't more data — it's data placed in context: the same transactional history, read against the value each customer represents, stops being a record and becomes a basis for decision.
Before acting on any customer, two things have to be clear — without them, execution is just activity:
- Where they are in the relationship — starting, growing, stabilizing, or declining. the lifecycle path
- How much value they represent — their economic contribution, and how it shifts over time. the value path
The two are related, but not the same: a customer can carry high value while their engagement fades. Acting well means reading both at once — and managing both. Manage only the lifecycle, and effort lands on the wrong customers; manage only value, and the right priorities never get executed.
Value-Cycle Management
Ineffective
Effective
Lifecycle Management
Effective
Efficiency without direction
Strong execution applied to the wrong priorities.
Sustainable growth
Clear value priorities, supported by effective lifecycle execution.
Ineffective
Value erosion
No clear priorities and no effective execution.
Potential wasted
The right priorities, but weak execution to realize them.
RADAR Insights is a decision operating system for a specific, high-impact set of recurring choices — which customers a business acquires, retains, and develops. These choices matter because the resources to act on them are scarce and costly: no company can invest in every customer at once, so the question is never whether to act, but where.
These decisions repeat across retail, services, and subscriptions; each traces back to a measurable effect on the portfolio; and most are still made by guesswork rather than evidence. That gap is where a structured system earns its keep.
2 of 5
The foundation
what it rests on
The foundation
Customer centricity
Running the business around the value of individual customer relationships — not around products, and not around an average buyer. Every decision RADAR supports flows from this single idea.
A business draws revenue from two sources: the customers it already has, and the ones it has yet to win. Both are under pressure — existing customers expect to be treated as individuals, and acquiring new ones keeps getting more expensive. Customer-centricity is the response, and RADAR's purpose follows directly: to help you identify your highest-potential customers, maximize the value they create, and attract more like them.
i.
The average customer doesn't exist.
Two customers with the same monthly spend can be on opposite trajectories — one compounding, one leaving. Reading the base as an average hides exactly the differences that decide the next twelve months.
ii.
Value is the lens.
Customers are measured by what they have already delivered and what they are projected to deliver. Value is the one metric that connects a customer's past with their future — used consistently from the first chart to the last cluster.
iii.
Growth has three forms.
A base grows by acquiring new customers, retaining those at the top, and developing those climbing the value ladder — each a distinct decision, with its own audience and economics.
3 of 5
The framework
what RADAR does
RADAR connects an intention to a practice, and supplies what's missing in between:
- The objective — customer-centricity. Treating customers as distinct individuals, not one average.
- The method — Customer Value Management. Diagnosing each customer's value, prioritizing it, and acting on it.
- The bridge — a system of AI agents. Each carries the analytics, methodology, and decision support most small businesses can't staff, so the discipline runs without a data team behind it.
Where it comes from: customer-centricity became a working method through six ideas from customer science, each answering the limit of the one before. RADAR is their synthesis.
RADthe purpose
Retention, Acquisition, Development — the three ways a business grows its customers. It gave focus, but no method to act on it.
STPthe structure
Segmentation, Targeting, Positioning gave that focus a strategic backbone. Yet classic STP is static — blind to how customers actually move.
RFMthe lens
Recency, Frequency, Monetary made segmentation behavioral and readable straight from transactions. But it ranks customers without saying what they're worth.
CLVthe north star
Customer Lifetime Value supplied the missing figure: what each customer is worth over time, not just how they rank. Alone, though, it stays a reporting number.
CVMthe discipline
Customer Value Management turns that figure into a system of decisions — diagnose, prioritize, act. But enterprise CVM is heavy and costly.
RADARthe synthesis
Brings it all together — enterprise-grade CVM made light enough for an SMB, built for non-contractual commerce and powered by transactional data alone.
4 of 5
The audience
who this is for
RADAR exists to support decisions — strategic and tactical — and the same customer-value lens speaks to every seat at the table. It turns customer behavior into one shared language for where to focus and what to do next.
CEOChief Executive
Growth quality. Whether expansion rests on durable, high-value customers or short-term volume — shaping customer mix, focus, and the value story behind the company. strategic
CFOChief Finance
Where capital earns its return. Which customers are profitable, whether spend on acquisition, retention, or service yields sustainable value, and how to forecast revenue with more confidence. strategic
CMOChief Marketing
Where effort goes. Which segments deserve attention, which channels bring high-value customers, and where retention and development move the most value. tactical
CPOChief Product
What the product prioritizes. Which features reinforce long-term relationships, and which parts of the portfolio create friction or fail to support growth. tactical
COOChief Operations
Where to align operations. Matching service load, demand, and process complexity to the value customers actually generate. tactical
Read this way, RADAR isn't a report to file — it's a shared decision lens, so every function acts on the same question: how customer behavior becomes long-term value.
5 of 5
The map
how to read this
RADAR is a method before it is a report. It runs in three stages, each building on the one before:
- Understand — read the portfolio as it really is: its value, its shape, how it moved, and what drives it.
- Diagnose — turn the reading into judgment: what's protected, what's at risk, and where to grow.
- Target — act on the right customers, with the right priority, for the right reason.
Every section opens with the idea, then shows your base through it — the lens is universal, the reading is yours, and every number comes from your own data. This edition is a single snapshot: the method pays off in repetition. Each cycle re-reads the base, and the movement between snapshots — who climbed, who slipped, what the last decisions changed — is where growth is actually managed.
Stage 1
Understand your customers.
Four complementary lenses read the customer base, each building on the last: from total value, to how that value is distributed, to how it moved this cycle, to why it rises and falls.
- Aggregate Valuemeasure the baseWhat the portfolio is worth today, how it grew this cycle, and where the next twelve months are heading.
- Portfolio Mixopen the baseHow value distributes across customer tiers — where it concentrates, and what that means for the business.
- Portfolio Mobilityfollow the movementHow value moved between tiers this cycle — retained, developed, activated, recovered, or acquired.
- Value Driverswhy value rises and fallsThe behaviors and attributes behind the movement — recency, frequency, spend, and the categories, channels, and stores that distinguish each tier.
Stage 2
Diagnose your priorities.
Stage 1 described the portfolio; this stage reads it as an asset to be managed. Five readings turn the lenses into judgment — what held, what is slipping, where to grow, what constrains the base, and the priorities that follow.
- Protected Valuewhat heldThe value successfully retained at the top of the base — the core the business depends on.
- Value at Riskwhat is slippingValue showing early signs of erosion — customers cooling before the loss compounds.
- Growth Opportunitieswhere to expandWhere value can still be developed — customers with clear room to climb the ladder.
- Structural Weaknesseswhat constrainsPersistent patterns that cap the portfolio's potential, regardless of this cycle's movement.
- Management Prioritiesthe synthesisThe few priorities that deserve attention this cycle, drawn from the four readings above.
Stage 3
Target the right audiences.
The previous stages read the portfolio and set the priorities; this stage gives them an address. Every customer is mapped by value and behavior, and the clusters that matter are opened, profiled, and made reachable.
- Targeting Matrixplace every customerThe full grid mapping every customer by value position and behavioral state, with portfolio-level row and column views.
- Clusters at a Glancescan the fieldA compact read of every cluster at once — size, value, and where each sits on the grid.
- Cluster Anatomyread one cellHow a single cluster is built — the structure and logic behind one cell of the matrix.
- Cluster Profilesmeet the audiencesThe priority clusters opened one by one, each with its profile, behavioral signature, and value position.
- Cluster Affinitiesfind what identifies themThe attributes that distinguish each priority cluster from the base — the levers to recognize and reach them.
Two things this report deliberately does not contain. First, execution — the channel, the message, the offer. That is yours: the report points to where effort earns the most; how to make it land is the work that follows.
Second, a single-cycle answer. Each edition is a snapshot; the movements between snapshots are where the real story is told.
Getting Started · conceptual framing. Figures begin in Stage 1 · Understand.
02
Understand
Read the customer base, one lens at a time.
Four lenses describe the portfolio in sequence — from the total value it carries, to how that value distributes across tiers, to the directions it grew, to the behaviors behind that growth. This stage only describes; judgment comes in the next stage.
Lens 1 of 4
Aggregate Value
measure the base
Four measures anchor the reading of the customer base — how much value it carries, how much revenue it produced last month and over the trailing year, and how much it is projected to generate in the next twelve months. Each measure is read on its own, in turn.
1 · Customer Value
May 2025 · YoY May 2024 → May 2025
Customer Value is the worth of the entire base as an asset — the net present value (NPV) of all expected cash flows from current relationships, combining value already realized with value projected forward, discounted to today.
Realized
Value already delivered by customers through their purchase history to date.
Projected
Value expected from those same customers over the next twelve months.
Customer Value
net present value · realized + projected
UYU1,436.9 M
+29.0%nominal YoY+22.9%real+20.7%USD
Composition
Realized79.7%
Projected23.8%
Realized · value already delivered
Projected · forward, next 12 months
Δ from existing customers
Δ from new customers
Customer Value grew UYU 323.4 M over the year — almost entirely from within.
- Existing customers drove it (+29.1%). New customers acquired in the period added just UYU 1.0 M — under 1% of the year-on-year gain.
- The mix tilted toward realized value. Projected value fell from 29.2% to 23.8% of the total as realized value caught up.
2 · Monthly Revenue
May 2025 · YoY May 2024 → May 2025
Monthly Revenue is the total billed in the reference month — the most immediate, concrete measure of activity in the base.
Repurchase
Revenue from customers who had already bought before — the recurring core.
Conversion
Revenue from customers making their first-ever purchase in the month.
Monthly Revenue
billed in the reference month
UYU38.0 M
+9.4%nominal YoY+4.1%real+2.3%USD
Composition
Repurchase88.8%
Conversion11.2%
Repurchase
Conversion
Monthly revenue grew +9.4% nominally, both engines pulling the same way.
- Repurchase +8.9% (UYU 30.1 M → 32.7 M) — the larger source.
- Conversion +12.2% (UYU 4.7 M → 5.2 M) — the fastest-growing, but the smallest in absolute terms.
3 · Run-rate Revenue
trailing 12 months · YoY May 2024 → May 2025
Run-rate Revenue is the total billed over the previous twelve months — a smoothed view that removes the seasonality of any single month.
From existing customers
Revenue from customers already in the base at the start of the window.
From new customers
Revenue from customers acquired within the trailing-12-month window.
Run-rate Revenue
billed over the previous 12 months
UYU439.9 M
+23.7%nominal YoY+17.8%real+15.6%USD
Composition
From existing customers98.6%
From new customers1.4%
From existing customers
From new customers
Run-rate revenue climbed steadily, finishing at UYU 439.9 M — +23.7% nominal over the year.
- Existing customers +23.7% (UYU 350.4 M → 433.6 M).
- New entrants +19.7% (UYU 5.2 M → 6.3 M) — both moving in step.
4 · Forecast
next 12 months · issued at May 2025
Forecast is the revenue the base is projected to generate over the next twelve months. Both of its components are projections — what differs is whether the customer exists yet.
Current Customers
Projected revenue from customers already in the base today.
Future Customers
Projected revenue from customers expected to be acquired in the coming year.
Forecast
revenue projected for the next 12 months
UYU425.3 M
+1.3%nominal YoY−3.6%real−5.3%USD
Composition · both projected
Current Customers80.5%
Future Customers19.5%
From current customers (projected)
From future customers (projected)
The forecast grew nominally but contracted in real terms.
- Current customers +5.3% (UYU 325.0 M → 342.4 M); future customers −12.6% (UYU 94.9 M → 82.9 M).
- The mix shifted to the known: current customers now carry 80.5% of the projection, up from 77.4%.
5 · Forecast Accuracy
how the model has tracked reality
Forecast Accuracy measures how close a past forecast came to what actually happened. It is the credibility check on every projection above — only a forecast whose 12-month window has already closed can be verified.
Current Customers
Accuracy of the projection for customers already in the base — the easier forecast.
Future Customers
Accuracy of the projection for not-yet-acquired customers — the harder forecast.
Latest verified accuracy
forecast issued May 2024 · realized through May 2025
95%
total revenue · UYU 419.9 M forecast vs UYU 439.9 M realized
By cohort
Current Customers99%
Future Customers81%
Verification — May 2024 forecast vs May 2024 → May 2025 realitylatest closed window
CohortForecastRealizedAccuracy
Current CustomersUYU 325.0 MUYU 323.1 M99%
Future CustomersUYU 94.9 MUYU 116.7 M81%
Total revenueUYU 419.9 MUYU 439.9 M95%
Five-cycle history
Across the five most recent verified cycles, the current-customers projection has tracked realized revenue within ±1% every time; the future-customers projection has averaged 92%, with the latest the weakest at 81%.
The model is sure of what it can see, less of what it can’t.
- Accurate on current-customer revenue, looser on future customers — the harder forecast, since it projects acquisitions that haven’t happened yet.
- This cycle, realized future-customer revenue came in 23% above projection.
All values expressed in revenue terms. Real growth applies Uruguay general CPI (+5.0% YoY); USD growth applies UYU/USD FX (+6.9% YoY).
Customer Value expressed as net present value (NPV) — projected cash flows discounted to today.
Lens 2 of 4
Portfolio Mix
open the base
The first lens measured the value of the base as a whole. That total moves through two levers — the number of customers and the value each one carries. But the moment we look at value per customer, a problem appears: there is no average customer. The base is a mix of customers worth wildly different amounts. This lens opens the aggregate and reads how size and value sit across it.
RADAR's first and simplest cut sorts every customer into value tiers — Platinum, Gold, Silver, and Bronze — ranked only by the value each one carries. Tiers are the coarse lens here; the fuller behavioural reading comes later, in Stage 3.
1 · The asymmetry of the base
May 2025
A few customers hold most of the value.
Each tier's slice of the customer base (bars), set against what an average customer in that tier is worth (line). The two move in opposite directions — the more customers a tier holds, the less each one is worth.
Share of base (bars)
Avg. customer value (line)
Base average (the "average customer")
The two curves cross like scissors — as a tier’s share of the base rises, its average customer’s value falls.
- A 59× gap. Bronze holds 71.8% of customers at UYU 2,451 each; Platinum holds 8.6% at UYU 144,738.
- The average describes no one. The base average of UYU 18,158 sits between Gold and Silver — about 85% of customers fall under it.
The same asymmetry, across every monetary measure
Customers
Customer
Value
Value
Monthly
Revenue
Revenue
Run-rate
Rev.
Rev.
Forecast
Platinum
8.6%
68.7%
68.1%
71.1%
65.3%
Gold
5.9%
11.1%
12.2%
11.1%
9.4%
Silver
13.6%
10.5%
10.3%
10.3%
7.1%
Bronze
71.8%
9.7%
9.3%
7.6%
3.5%
Platinum fills every monetary square; Bronze fills only the customer one.
- Platinum: 68.7% of value, 68.1% of this month’s revenue, 71.1% of run-rate, 65.3% of forecast — on 8.6% of customers.
- Bronze: 71.8% of customers, fading across the rest (9.7% of value down to 3.5% of forecast).
- One loosening: in this month’s revenue Bronze (9.3%) edges above its run-rate weight — the long tail buys a little more in the moment than its lasting value suggests.
2 · The two motors, tier by tier
13 months · YoY May 2024 → May 2025
Each tier grows by its own mix of size and value.
A tier's value moves through two motors: how many customers it holds (share of base) and how much each one is worth (average value). Read per tier over thirteen months, the two motors rarely move together — and average value is shown in both nominal and real (inflation-adjusted) terms.
Platinum
Gold
Silver
Bronze
Share of base
Average customer value
The two motors — size and value — move differently in every tier.
- Bronze: the only tier to gain share (+0.3pp) and the only one to lose average value (−3.7% nom, −8.3% real).
- Platinum: the only tier to grow average value in real terms (+2.7%), even as its share edged down.
- Gold & Silver: nearly still on size, slipping in real value. Headcount drifted toward Bronze while real value per customer climbed only at the top.
3 · What moved the value
contribution to ΔCLV · YoY May 2024 → May 2025
More customers, not more value per customer.
Each tier's value grew by some mix of the two motors. This splits every tier's year-on-year value gain into the part that came from more customers (size) and the part that came from more value per customer (value) — and sums them back to the portfolio total from Lens 1.
From more customers (size)
From more value per customer (value)
Value per customer fell
The portfolio added UYU 323.3 M — and 81% came from more customers, not more value per customer.
- Platinum is the exception: of its +240.6 M, about a quarter (64.7 M) came from each customer growing more valuable.
- Everywhere else, value per customer barely moved or fell — Gold and Silver grew almost entirely on headcount; Bronze grew on volume while value per customer slipped.
Lens 3 of 4
Portfolio Mobility
follow the movement
The last lens showed that 81% of the year's value growth came from more customers, not more value per customer. The base grew mostly by size. This lens follows that motor in motion: how customers moved between tiers over the year — who climbed, who slipped, who entered — and what that movement was worth.
Growth has three generic moves: protect the customers you have, develop the ones with potential, and bring in new ones. Simple to name — too coarse to act on.
The moment a base is organized by value tiers, each of those three moves splits in two. Protecting a top customer is a different job from catching one who is slipping. Developing a promising customer is different from consolidating a standard one. Acquiring a brand-new customer is different from winning back a dormant one. Three generic strategies become six concrete vectors of movement — this is RADAR.
Retentionprotect value
RetentionHigh-value customers staying high-value — protect the top.
ActionTop customers slipping down — intervene before value is lost.
Developmentgrow value
DevelopmentPromising mid-tier customers climbing toward a higher tier.
ActivationStandard customers consolidating into steady value.
Acquisitionadd value
AcquisitionNew customers entering the base for the first time.
ReacquisitionLow-value customers — dormant returners and new low-value entrants.
Two further states — customers who simply stayed in place, and those who slipped a tier without being flagged — are not growth vectors. They are not read as strategy; they appear only later, when the value of all movement is reconciled to the total.
1 · How customers moved between tiers
year on year · May 2024 → May 2025
A base that mostly stays where it is.
Before reading the vectors one by one, the overview of movement: of the customers in each tier a year ago, where are they now? The diagonal is what stayed; above it, customers who climbed; below it, customers who slipped. The bottom row is the new customers who entered the base.
Where each tier's customers ended up
where last year’s customers are today · % of origin (n) · ΔYoY (pp · custs)
From ↓ May’24
To → May’25
To → May’25
Platinum
Gold
Silver
Bronze
Platinum
97.5%6,532↑ 0.8pp · 210
2.5%169↓ 0.8pp · 64
—
—
Gold
4.9%224↑ 0.6pp · 33
89.7%4,087↓ 1.2pp · 88
5.4%248↑ 0.7pp · 41
—
Silver
0.4%37↑ 0.1pp · 6
3.1%324↓ 0.4pp · 22
93.2%9,888↑ 0.5pp · 540
3.4%356↓ 0.2pp · 18
Bronze
0.0%13↑ 0.0pp · 2
0.1%84↑ 0.0pp · 9
1.0%560↑ 0.2pp · 120
98.8%55,188↑ 0.3pp · 2,340
New
0.6%9↑ 0.1pp · 3
1.7%24↑ 0.4pp · 8
6.9%98↓ 1.1pp · 30
90.8%1,291↑ 0.6pp · 180
New customers who entered the base over the year, by the tier they landed in (1,422 total).
Stayed in tier
Moved up
Moved down
Entered the base
The base is strongly self-retaining.
- Every tier keeps the vast majority across the year: Bronze 98.8%, Platinum 97.5%, Silver 93.2%, Gold 89.7%.
- Movement is real but narrow: the widest upward flow is Gold→Platinum (224, 4.9% of Gold); the widest downward, Gold→Silver (248, 5.4%).
- New customers land almost entirely in Bronze (90.8%). Growth came less from climbing than from new arrivals at the bottom and the base holding what it had.
2 · The intensity of each vector
12-month rate · YoY to May 2025
How active each vector was.
For each growth vector: the current movement rate (share of its eligible base that moved this way), how it shifted versus a year ago, and its month-by-month trajectory. Each sparkline is scaled to its own range — the shapes show rhythm, not relative size.
Retention · protect value
RetentionTop customers staying on top
97%
±0 pp YoY
Jun '246,532 of 6,701 retainedMay '25
ActionTop customers slipping — intervene
rare
Structurally small — a consequence of high retention. See note below.
22 customers flagged
Development · grow value
DevelopmentMid-tier climbing higher
6%
−2 pp YoY
Jun '24274 of 4,555 developedMay '25
ActivationStandard customers consolidating
4%
−1 pp YoY
Jun '24406 of 10,592 activatedMay '25
Acquisition · add value
AcquisitionNew customers entering
7%
−2 pp YoY
Jun '241,422 entered the baseMay '25
ReacquisitionDormant customers returning
1%
−1 pp YoY
Jun '24560 of 55,845 recoveredMay '25
A note on Action
The Action vector — top customers slipping and needing intervention — involved only 22 customers this cycle. This is a direct consequence of the 97% retention at the top: when almost no high-value customer slips, there is almost nothing for the Action vector to catch. Its scarcity is a sign of stability, not of a missing motion.
Every active vector cooled slightly over the year.
- Most eased: Development to 6% (−2pp), Acquisition to 7% (−2pp), Activation to 4% and Reacquisition to 1% (−1pp each); Retention held firm at 97%.
- The base became no more mobile — if anything, marginally less. Growth came from steady retention at the top and steady inflow at the bottom, not from customers climbing between tiers.
3 · What the movement was worth
value uplift by vector · 12 months
Every month, movement adds and subtracts value.
Each customer movement carries value with it. This matrix reads, month by month, how much customer value each vector added — and how the two non-strategic states (customers who stayed in place, customers who slipped) close the balance. Every column sums to the net change in portfolio value for that month.
Jun’24
Jul’24
Aug’24
Sep’24
Oct’24
Nov’24
Dec’24
Jan’25
Feb’25
Mar’25
Apr’25
May’25
12-mo
Retention
17.3
6.7
12.3
7.7
7.7
17.8
17.1
7.8
8.3
17.7
10.9
9.5
141
Action
—
—
—
—
—
—
—
—
0.0
0.0
0.1
—
0.1
Development
8.4
4.8
9.4
7.1
6.9
6.4
6.5
4.5
2.7
6.5
12.1
7.3
83
Activation
5.8
4.9
6.2
5.8
3.8
6.8
5.7
3.3
3.0
5.1
6.4
6.2
63
Acquisition
4.9
4.3
4.1
3.9
4.0
5.7
7.2
6.2
6.1
7.6
5.7
6.2
66
Reacquisition
5.2
3.3
5.2
4.2
3.7
4.8
4.3
2.8
2.2
3.8
4.9
4.1
48
Not growth vectors — they close the balance to 100% of the monthly change
No tier change
−1.3
−7.1
−4.6
−5.4
−7.2
−0.7
−3.7
−7.4
−6.1
−1.4
−0.9
−4.6
−50
Moved to lower tier
−2.8
−3.2
−2.5
−2.9
−3.2
−1.8
−2.4
−2.8
−1.6
−1.0
−1.5
−1.8
−28
Total value uplift
38
14
30
20
16
39
35
14
15
38
38
27
323
Values in UYU millions · customer value uplift per vector per month
Value added (growth vector)
Value subtracted (non-vector)
12-month total per vector
Retention carried the year.
- Retention: UYU 141 M of value uplift — more than the next two vectors combined; brightest in June, November, December and March.
- A steady middle: Development (83 M), Acquisition (66 M), Activation (63 M); Reacquisition (49 M) thinner but consistent; Action barely registers (0.1 M).
- Net UYU 323 M: the six vectors generated UYU 401 M gross; flat and slipping customers subtracted 78 M. Value arrives in waves, peaking when retention is strongest.
So the size motor resolves into a clear shape. The base grew by holding its top — retention is both the widest movement and the largest source of value — and by a steady inflow at the bottom, while customers climbing between tiers stayed a narrow, if reliable, contribution. The next lens turns from how customers moved to why each one became worth more or less: the drivers of value per customer.
Lens 4 of 4
Value Drivers
why value rises and falls
The previous lens followed customers moving between tiers. But customers move for only one reason — their value changed. This lens turns from the movement to its cause: what lifts a customer’s worth over time, and what quietly wears it away.
A customer’s value isn’t declared — it’s revealed in how they buy.
Three behaviours capture almost all of it:
- Recency — how recently they last bought. Fresh activity keeps the relationship alive; long silence lets it fade.
- Frequency — how often they buy. A steady rhythm compounds into value; widening gaps erode it.
- Monetary — how much they spend each time. Larger, richer baskets lift the whole curve.
Value climbs when these move in the right direction, and slips when they don’t. The real question is less which tier a customer is in than which behaviour is changing their worth — and a handful of distinct drivers capture every meaningful shift:
Reengaged
A customer who had gone quiet comes back — recency resets, the relationship reopens.
Onboarded
A recent arrival places a second order — the fragile first step into a lasting relationship.
Cadenced
A customer settles into a steady rhythm of purchases — frequency becomes a habit.
Expanded
A customer spends more per order — the basket grows in size or value.
Sustained
A healthy customer simply stays healthy — holding their value rather than changing it.
These levers are scored across the whole base, independently of tier — in principle, any customer can pull any lever. But read tier by tier, the picture is not uniform at all. That is what this lens makes visible.
1 · The same lever is worth more at the top
success moments × tier · May 2025
Where each lever happens — and what it's worth.
Each cell reads two things at once: how often a lever is pulled in that tier (the rate), and how much value it adds per customer when it is (the impact, in average customer value). The levers are scored the same way everywhere — but their frequency and their worth shift sharply across tiers.
Impact per customer · rate within tier
impact in UYU · rate = % of tier · Portfolio = % of base
Lever ↓ / Tier →
Platinum
Gold
Silver
Bronze
Portfolio
Reengagedcame back
28,7032% · 15 custfrom 1% · +1pp
15,3893% · 32 custfrom 1% · +2pp
7,4583% · 114 custfrom 2% · +1pp
2,621<1% · 73 custfrom <1% · +0.3pp
8,3951% · 234 custfrom 1% · 0pp
Onboarded2nd order
70,32733% · 12 custfrom 30% · +3pp
25,01341% · 49 custfrom 37% · +4pp
10,39423% · 124 custfrom 21% · +2pp
3,1471% · 61 custfrom 1.5% · -0.5pp
14,4321% · 246 custfrom 1% · 0pp
Cadencedsteady rhythm
25,40324% · 105 custfrom 22% · +2pp
8,0288% · 82 custfrom 7% · +1pp
3,2721% · 28 custfrom 2% · -1pp
1,882<1% · 2 custfrom <1% · -0.3pp
15,7651% · 217 custfrom 1% · 0pp
Expandedbigger basket
8,89821% · 458 custfrom 19% · +2pp
6,5553% · 60 custfrom 2% · +1pp
—not in tier
—not in tier
8,6272% · 518 custfrom 2% · 0pp
Sustainedstayed healthy
1,24095% · 3301 custfrom 97% · -2pp
−1,20179% · 343 custfrom 82% · -3pp
−50584% · 882 custfrom 86% · -2pp
—not in tier
7156% · 4,526 custfrom 7% · -1pp
12,345 impact on average customer value (UYU)
24% · n cust how often, within the tier
value added
value eroded
The same behaviour is worth dramatically more at the top.
- An eleven-fold gap for the identical action: reengaging a Platinum customer adds UYU 28,700; a Bronze customer, just 2,600.
- Levers cluster by tier: onboarding and cadence concentrate up top; expansion barely exists below Gold.
- Sustained is the trap: the most common state (79–95% of mid and top tiers), yet in Gold and Silver its impact turns negative — behaviour held, real value quietly lost. Staying the same is not standing still.
2 · How each cohort matures
acquisition cohorts · by quarter of age
The customers who stay become worth more.
Each row is a cohort — customers acquired in one quarter — read left to right as it ages. Two columns track the value, two the loyalty behind it, two the buying intensity of those still active. Reading down a column compares cohorts at the same stage of life. The number is each cohort's latest quarter.
Financial · per customer
Loyalty · share of cohort
Intensity · among active
Cohort
age · size
age · size
Valuetotal per customer
Forecastexpected future
Repurchase% of cohort / month
Repeaterscum. % with 2+
Ordersper active / month
AOVper order
Revenueper active / month
2021-Q316Q · 16,752
3,159
vs avg +15%
6%
vs avg +25%
59%
vs avg +16%
1.6
vs avg +7%
3,579
vs avg -7%
5,726
vs avg +2%
2021-Q415Q · 14,175
21,081
vs avg +18%
2,969
vs avg +7%
4%
vs avg -7%
52%
vs avg +4%
1.6
vs avg +1%
3,726
vs avg -1%
5,961
vs avg +0%
2022-Q114Q · 18,537
11,982
vs avg -30%
2,090
vs avg -22%
4%
vs avg -18%
39%
vs avg -20%
1.5
vs avg -8%
4,136
vs avg +8%
6,204
vs avg -2%
2022-Q213Q · 14,445
13,825
vs avg -11%
1,954
vs avg -18%
4%
vs avg -8%
47%
vs avg -1%
1.6
vs avg -1%
3,651
vs avg -5%
5,841
vs avg -5%
2022-Q312Q · 11,565
16,850
vs avg +4%
3,025
vs avg +7%
6%
vs avg +3%
51%
vs avg +4%
1.6
vs avg -2%
3,809
vs avg +5%
6,094
vs avg +1%
2022-Q411Q · 16,242
15,152
vs avg -5%
2,904
vs avg +1%
5%
vs avg -11%
43%
vs avg -11%
1.6
vs avg -4%
4,087
vs avg -4%
6,539
vs avg -7%
2023-Q110Q · 14,988
15,284
vs avg +4%
3,420
vs avg +16%
4%
vs avg -14%
37%
vs avg -18%
1.6
vs avg +0%
3,986
vs avg +10%
6,378
vs avg +7%
2023-Q29Q · 13,212
14,284
vs avg +0%
3,042
vs avg +0%
4%
vs avg -19%
38%
vs avg -12%
1.6
vs avg -1%
4,609
vs avg +27%
7,375
vs avg +18%
2023-Q38Q · 8,013
13,508
vs avg -1%
3,279
vs avg +5%
6%
vs avg +12%
39%
vs avg -9%
1.4
vs avg -6%
4,515
vs avg +16%
6,321
vs avg +5%
2023-Q47Q · 9,393
11,601
vs avg -10%
2,949
vs avg -7%
5%
vs avg -10%
36%
vs avg -13%
1.5
vs avg -3%
3,629
vs avg -1%
5,444
vs avg -3%
2024-Q16Q · 12,015
13,356
vs avg +6%
3,962
vs avg +17%
6%
vs avg -7%
37%
vs avg -7%
1.4
vs avg -5%
4,114
vs avg +14%
5,759
vs avg +4%
2024-Q25Q · 13,995
11,077
vs avg -7%
2,824
vs avg -18%
6%
vs avg -16%
39%
vs avg -2%
1.5
vs avg -1%
3,734
vs avg +3%
5,601
vs avg +1%
2024-Q34Q · 9,831
14,333
vs avg +26%
4,715
vs avg +33%
9%
vs avg +23%
42%
vs avg +9%
1.5
vs avg -1%
3,665
vs avg +5%
5,497
vs avg +3%
2024-Q43Q · 11,286
13,234
vs avg +21%
4,829
vs avg +24%
9%
vs avg +8%
41%
vs avg +11%
1.5
vs avg +0%
3,354
vs avg -3%
5,031
vs avg -2%
2025-Q12Q · 12,783
11,876
vs avg +16%
4,433
vs avg +5%
10%
vs avg +3%
34%
vs avg +1%
1.6
vs avg +7%
3,718
vs avg +12%
5,948
vs avg +16%
Rows span each cohort's life, Q0 → latest. Scales shared within each column. Value & Repeaters = end-of-quarter; Repurchase, Orders, AOV, Revenue = quarter average. Repurchase is share of the whole cohort; Orders, AOV & Revenue are per active customer.number = latest quarter · dashed = average cohort · vs avg = gap at equal age
Three patterns run through every cohort:
- Loyalty thins, intensity deepens. Repurchase falls from 11–17% in the first quarter to 4–6% after a couple of years — but the ones who stay buy harder (orders climb from ~1.3–1.4 toward 1.6–1.7 a month, revenue with them). The base narrows and deepens; value concentrates in the loyal.
- New customers arrive stronger. First-quarter value rose from ~UYU 5,000–9,000 for the 2022–early-2023 cohorts to UYU 11,600–12,800 for 2024–2025, and the forecast confirms it at entry (~UYU 6,800–7,400 vs 2,000–5,500 for older vintages) — they enter buying more often and spending more.
- Time compounds value. 2021-Q3 reaches UYU 20,500 and 59% repeaters after four years, its forecast settling as value shifts from expected to already realised. The shape repeats across vintages — each new cohort now begins the climb from a higher base.
This closes the value motor. Customer value rises through behavioural levers worth most at the top of the base, and through the maturation of each cohort — a base that narrows in loyalty but deepens in intensity, with newer cohorts arriving stronger and projected to be worth more than those before them. Whether that improvement is fast enough, and where it is worth pressing hardest, is the question the next stage takes up.
03
Diagnose
From what the base shows to what it means.
Stage 1 described the portfolio. This stage reads it as an asset to be managed — what value is protected, what is at risk, where it can still grow, and what is weakening underneath — and turns that into the priorities for the cycle.
The reading this cycle
The portfolio's value grew — but the growth came from the top of the base compounding, not from new customers. The customers already at the top kept buying, deepening value they had already earned, while those who joined this cycle landed at the bottom and added headcount more than value. The decision this cycle is to protect that compounding engine at the top — and to revive the flow of customers climbing toward it, which is slowing.
01
Protected Value
where the business relies with confidence
The top of the base is where value compounds: these customers keep buying, turning continued loyalty into value already realized — not a projection, but value the business has banked.
- The Platinum tier holds 68.7% of portfolio value on 8.6% of customers, and its share of value still rose over the year — concentration is deepening at the top, not eroding.
- Almost all of the top tier's value gain this cycle came from continued buying by customers already there — value realized through repeat purchases, not from new arrivals. Platinum loyal customers buy roughly every 26 days, the highest-value rhythm in the base.
- Retention is the year's strongest force — it generated more value than the next two growth vectors combined. The base is strongly self-retaining (Platinum 97.5%, Bronze 98.8%). This is banked value, already earned, not a forecast that could fade.
Value pyramid — direction per tier
Platinum
Protect
Gold
Develop
Silver
Activate
Bronze
Recover
02
Value at Risk
where value erodes beneath the surface
The middle of the pyramid kept its loyal standing on paper while the behaviour behind it aged. The label held; the value receded with it.
- Gold and Silver customers who stayed loyal still lost value (−1,192 and −508 on average) — not because they left, but because their buying slowed.
- The mechanism is in the rhythm: recency runs ~100 days at Platinum, ~271 at Gold, ~285 at Silver, and the gap between purchases nearly doubles down the tiers. Value follows real behaviour, not the label.
- Customers who slipped a stage signalled it first — they were already taking ~149 days to a second order and buying every ~72 days before the drop. The risk was legible in behaviour before it became lost value.
Why loyal value receded — recency by tier (held-standing customers)
03
Growth Opportunities
where value can still expand
The next source of value is already inside the base. The highest-return move is not acquiring more — it is deepening the rhythm of customers already at the top.
- The same behavioural move is worth far more at the top: a customer gaining cadence in Platinum adds ~25,000 in value; onboarding one adds ~84,000 — the highest value-per-customer of any move in the base.
- Expansion barely exists below Gold — the lever to lift mid-tier value is cadence, not acquisition. The room to grow is in frequency among customers who already buy.
- New customers arrive stronger than before (newer cohorts start at higher value), so the entry is worth densifying rather than simply enlarging.
- But the flow of customers climbing into higher tiers is slowing — the rates of customers reaching Gold and Platinum, and of new customers arriving, all fell over the year. The top compounds value today, yet the movement that refills it is weakening: reviving that upward flow is the opportunity with the longest reach.
Value added per customer, by behavioural move (Platinum)
First purchase (converted)
109,000
Became occasional (onboarded)
84,000
Gained cadence
24,800
Reactivated
28,700
Became loyal (expanded)
9,050
04
Structural Weaknesses
where future value fails to form
Beneath the year's headline growth, the base is widening without densifying. The body of the portfolio is sitting still and going cold.
- By far the largest group — about 64,000 customers — stayed in the same stage, with an average recency near 475 days. The growth came from more customers, but the core of the base is dormant.
- The entry is not densifying: 91% of new customers land in Bronze, and most stall there rather than building frequency — the inflow adds count, not future value.
- The base narrows in loyalty but deepens in intensity (orders per active customer rise as cohorts age) — which shifts the question from "retain everyone" to "retain the right ones".
The cold body of the base
~64,000
customers stayed in the same stage — the largest single group in the portfolio
~475 days
average recency of that group — more than a year since their last purchase
05
Management Priorities
what to do with this reading
What the business should protect, improve, and accelerate now — direction, not yet audience.
Protect
Defend the compounding core at the top. The Platinum tier carries most of the portfolio's value and still buys with rhythm. Protecting its cadence preserves the value the business can most rely on.
Improve
Address the loyal-but-cooling middle. Gold and Silver kept their standing while their buying slowed and their value receded. Improving cadence here recovers value that is quietly leaking, not yet lost.
Accelerate
Densify the entry, don't just enlarge it. The inflow grows the count but stalls in Bronze. Accelerating the move from first purchase to a buying habit is where future value will form.
These priorities set where to act and how. The next stage gives them an address — the specific groups of customers behind each one — without prescribing channel, offer, or timing.
04
Target
From priorities to the customers behind them.
The previous stages read the portfolio and set the priorities. This stage gives them an address: it maps every customer cluster by value and lifecycle, highlights where value and forecast concentrate, and lets you take the priority clusters directly into action.
Stage 3 · Targetsection one
The targeting matrix
A customer base is never one pool. Every customer sits at the crossing of two readings:
- Value stage — where their worth to the business stands, and where it is heading.
- Lifecycle behaviour — how they are buying right now.
Cross the two — five by five — and the base resolves into twenty-five clusters, each a group sharing the same value position and the same behaviour: the smallest group specific enough to act on, with the exact customer list behind it. Read this way, the grid is a diagnostic snapshot — a heatmap of where customers concentrate, where value is generated today, and where dormant value waits, before any action is prescribed.
Targeting matrix — all customer clusters
Cell shade = value + forecast combined (darker = more). Each cell shows its cluster, value today, and forecast. Row and column totals (dark) are segment totals, not clusters. Tap any cluster to add it to the download.
Value Stage ↓ Lifecycle →
Loyal
CustomersREWARD
CustomersREWARD
Light
SpendersEXPAND
SpendersEXPAND
Occasional
BuyersCADENCE
BuyersCADENCE
Just
AcquiredONBOARD
AcquiredONBOARD
About
to SleepREENGAGE
to SleepREENGAGE
Total
value stage
value stage
High ValueRETENTION
★
Cluster 01662.7MFcst 144.6M3,589 cust
★
Cluster 02231.1MFcst 64.5M2,162 cust
★
Cluster 0327.7MFcst 13.7M347 cust
Cluster 041.5MFcst 0.4M19 cust
★
Cluster 0563.5MFcst 0.2M698 cust
High Value986.4MFcst 223.4M6,815 cust
Value at RiskACTION
Cluster 06no customers
Cluster 070.06MFcst 0.01M3 cust
Cluster 080.31MFcst 0.14M15 cust
Cluster 090.04MFcst 0.01M1 cust
Cluster 100.05MFcst 0.00M3 cust
Value at Risk0.5MFcst 0.2M22 cust
High PotentialDEVELOPMENT
Cluster 1115.3MFcst 2.1M409 cust
★
Cluster 1268.1MFcst 15.5M1,974 cust
Cluster 1331.8MFcst 13.2M963 cust
Cluster 142.7MFcst 0.7M82 cust
Cluster 1541.7MFcst 0.5M1,253 cust
High Potential159.7MFcst 32.0M4,681 cust
Early StageACTIVATION
Cluster 1616.5MFcst 2.8M1,066 cust
Cluster 1729.2MFcst 5.4M1,940 cust
Cluster 1839.9MFcst 12.8M2,860 cust
Cluster 195.5MFcst 1.3M444 cust
Cluster 2059.9MFcst 1.9M4,469 cust
Early Stage151.0MFcst 24.2M10,779 cust
Low ValueREACQUISITION
Cluster 21no customers
Cluster 227.3MFcst 1.2M1,517 cust
Cluster 2326.7MFcst 4.9M8,477 cust
Cluster 2415.5MFcst 2.8M5,786 cust
★
Cluster 2589.8MFcst 3.2M41,055 cust
Low Value139.3MFcst 12.1M56,835 cust
Total lifecycle
Loyal Customers694.5MFcst 149.5M5,064 cust
Potential Growth335.8MFcst 86.5M7,596 cust
Occasional Buyers126.5MFcst 44.8M12,662 cust
Just Acquired25.2MFcst 5.2M6,332 cust
About to Sleep254.9MFcst 5.9M47,478 cust
Portfolio1.44BFcst 291.8M79,132 cust
value + forecast, low → high
★priority cluster (top 80% of value or forecast)
segment totals — tap to select the whole row/column
no customers this cycle
Value and potential are concentrated, not spread: seven clusters out of twenty-five carry the top 80% of the portfolio — whether measured by value today or by forecast. Most sit at the top of the value pyramid, but the set also includes the large Low Value · About to Sleep cluster, where a high stock of past value has gone cold: little forecast, but real value to defend.
Stage 3 · Targetsection two
Clusters at a glance
Seven clusters carry the top 80% of value or forecast. Here they are side by side — how many customers each holds, and how much of the portfolio's value and forecast sits inside it. The bars make the relative weight visible.
| Cluster | Customers | Share of cust. | Value | Share of value | Forecast | Share of fcst. |
|---|---|---|---|---|---|---|
Cluster 01High Value · Loyal Customers | 3,589 | 4.5% | 662.7M | 46.1% | 144.6M | 49.5% |
Cluster 02High Value · Light Spenders | 2,162 | 2.7% | 231.1M | 16.1% | 64.5M | 22.1% |
Cluster 25Low Value · About to Sleep | 41,055 | 51.9% | 89.8M | 6.3% | 3.2M | 1.1% |
Cluster 12High Potential · Light Spenders | 1,974 | 2.5% | 68.1M | 4.7% | 15.5M | 5.3% |
Cluster 05High Value · About to Sleep | 698 | 0.9% | 63.5M | 4.4% | 0.2M | 0.1% |
Cluster 20Early Stage · About to Sleep | 4,469 | 5.6% | 59.9M | 4.2% | 1.9M | 0.7% |
Cluster 03High Value · Occasional Buyers | 347 | 0.4% | 27.7M | 1.9% | 13.7M | 4.7% |
Stage 3 · Targetsection three
Cluster anatomy
The previous table sized each cluster as a whole. This one looks inside it — at the typical customer. The averages below describe how a single member of each cluster behaves: what they're worth, how recently and how often they buy, and how much they spend per visit. This is the profile your team will recognise when they open the list.
| Cluster | Lifetime ValueUYU |
ForecastUYU | RevenueUYU | Monthly RevenueUYU |
Frequencyorders | Monthly Frequencyorders |
Order ValueUYU |
Basket Sizeitems |
Items PriceUYU |
Breadthcategs | Active Periodsmonths |
Recencydays | Tenuredays | Inter Purchasedays |
Second Orderdays |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Cluster 01High ValueLoyal Customers | 185k | 40k | 183k | 4.3k | 39.5 | 0.78 | 5.1k | 26.0 | 389 | 2.3 | 26 | 97 | 1.1k | 26 | 59 |
Cluster 02High ValueLight Spenders | 107k | 30k | 93k | 3.5k | 32.6 | 1.14 | 3.2k | 18.6 | 206 | 2.2 | 22 | 34 | 1.0k | 22 | 63 |
Cluster 25Low ValueAbout to Sleep | 2.2k | 78 | 3.1k | 0 | 1.5 | 0.00 | 2.3k | 13.1 | 374 | 1.1 | 3 | 986 | 1.0k | 56 | 110 |
Cluster 12High PotentialLight Spenders | 35k | 7.9k | 33k | 827 | 13.4 | 0.37 | 2.8k | 16.3 | 217 | 1.7 | 10 | 132 | 979 | 38 | 108 |
Cluster 05High ValueAbout to Sleep | 91k | 287 | 135k | 0 | 28.6 | 0.00 | 7.4k | 28.4 | 4.7k | 1.5 | 12 | 705 | 1.4k | 21 | 30 |
Cluster 20Early StageAbout to Sleep | 13k | 425 | 19k | 0 | 6.0 | 0.00 | 4.7k | 23.1 | 888 | 1.2 | 5 | 829 | 1.2k | 45 | 82 |
Cluster 03High ValueOccasional Buyers | 80k | 39k | 42k | 7.9k | 4.8 | 0.92 | 12k | 27.4 | 5.2k | 1.3 | 4 | 41 | 207 | 25 | 43 |
Read each column on its own: the font colour runs green (best value in that column across the seven clusters) to red (weakest). The two arrows in each header sort the table by that metric. All money in UYU; days are averages per customer. Lifetime Value = average lifetime value; Tenure = days as a customer; Recency = days since last order; Inter Purchase = typical gap between orders; Second Order = days from first to second order. Tap a cluster name to add it to the download.
Stage 3 · Targetsection four
Cluster profiles
One card per priority cluster: who they are, what's happening in their relationship with the business, and the recommended direction — the strategic intent for the group. Select the clusters you want and download the customer lists to take into action.
1
Cluster 01 · High Value · Loyal Customers
Loyal top-value customers
Top of the value pyramid · buying regularly · the protected core
662.7M
value · 46.1%
144.6M
forecast · 49.5%
These customers hold 46% of all the value in the portfolio on fewer than 4,000 people. They buy regularly, recently, and at a consistent ticket — the most valuable rhythm in the base. Their forecast alone is nearly half the portfolio's total.
Recommended direction
Sustain and reward the relationship. These customers carry the largest concentration of present value in the portfolio. Deepen the engagement before any decline appears in their behaviour.
Include Loyal top-value customers in the download
3,589 customers
2
Cluster 02 · High Value · Light Spenders
Top-value customers, growing their spend
Top of the value pyramid · buying often and recently · ticket still rising
231.1M
value · 16.1%
64.5M
forecast · 22.1%
Top-tier customers with the strongest recency in the whole base (34 days) and high frequency — but a ticket below their tier's potential. The relationship is healthy; the room to grow is in value per purchase.
Recommended direction
Sustain the relationship while expanding economic depth. The lifecycle is healthy; the value-per-purchase has room to grow. Deepen the value of each transaction.
Include Top-value, growing spend in the download
2,162 customers
3
Cluster 25 · Low Value · About to Sleep
The large dormant base
Lowest value tier · long inactive · the portfolio's biggest group by headcount
89.8M
value · 6.3%
3.2M
forecast · 1.1%
By far the largest group — 41,055 customers, over half the base — but each one bought rarely and last did so almost three years ago. Individually they are worth little and their forecast is near zero. Yet the accumulated past value is real (6.3% of the portfolio), which is why the group qualifies: not for its future, but for the value already there to be recovered.
Recommended direction
Test reactivation at low cost, expect little. This is a recovery play, not a growth one. The right move is a cheap, scalable win-back test — not heavy investment in a group whose forecast is minimal. Treat any return as upside.
Include The large dormant base in the download
41,055 customers
4
Cluster 12 · High Potential · Light Spenders
Promising customers, growing their spend
Rising value position · active buyers · ticket below potential
68.1M
value · 4.7%
15.5M
forecast · 5.3%
The growth engine: customers one tier below the top with healthy, active buying but a ticket below their potential. They are the most likely source of next year's top-tier customers if their per-purchase value rises.
Recommended direction
Compound value through deeper transactions. The relationship is healthy and growing; the missing dimension is per-purchase value. Expand the economic depth of each interaction.
Include Promising, growing spend in the download
1,974 customers
5
Cluster 05 · High Value · About to Sleep
Top-value customers going quiet
Top of the value pyramid · high lifetime value · stopped buying
63.5M
value · 4.4%
0.2M
forecast · 0.1%
A small but valuable alarm: 698 top-tier customers who built high lifetime value — 28 orders each, the highest tickets in their tier — but have now gone nearly two years without buying. The forecast has collapsed to near zero, yet the realised value is among the highest per customer in the base. These are the most valuable customers at risk of being lost outright.
Recommended direction
Win back with priority — this is defensible value. Unlike the dormant mass, each of these customers is individually worth recovering. A targeted, personal re-engagement effort is justified here before the relationship lapses permanently.
Include Top-value customers going quiet in the download
698 customers
6
Cluster 20 · Early Stage · About to Sleep
Early customers who drifted away
Lower-value tier · a handful of early orders · long inactive
59.9M
value · 4.2%
1.9M
forecast · 0.7%
4,469 customers who started a relationship — around six orders each — but stalled early and have been inactive for over two years. Mid-sized tickets while active, so the accumulated value is meaningful as a group, but the forecast is very low. They never reached the top of the pyramid and have since cooled.
Recommended direction
Reactivate selectively, at scale. Too many and too cold for individual attention, but the early buying signal makes them better win-back candidates than the fully dormant base. A low-cost, segmented campaign is the right test.
Include Early customers who drifted away in the download
4,469 customers
7
Cluster 03 · High Value · Occasional Buyers
Top-value customers, building cadence
Top of the value pyramid · recent · frequency slowing
27.7M
value · 1.9%
13.7M
forecast · 4.7%
A small group — only 347 customers — but with the highest average ticket of any priority cluster (12,295). Top-tier and recent, yet their purchase frequency has slowed. The earliest place where a top customer starts to cool, before value is lost.
Recommended direction
Watch and protect. A frequency slowdown precedes recency loss and tier demotion in this group. Re-establish purchase cadence before they drift dormant.
Include Top-value, building cadence in the download
347 customers
Each download is the full customer list for the selected group — customer ID, recency, frequency, value, and behavioural detail — so your team can act on the exact people behind the number. The recommended direction sets the intent; the channel, offer, and timing are yours to design.
Stage 3 · Targetsection five
The affinity matrix
Beyond how much they're worth, each cluster has its own tastes — the product categories, payment methods and stores its members reach for more often than the average customer. Affinity measures exactly that: how much more (or less) likely a cluster is to choose an attribute compared with the typical customer. A cluster that buys wine three times as often as the base has a strong affinity for wine. This is what tells you what to offer, how to let them pay, and where to reach them.
| Attribute | Cluster 01High ValueLoyal Customers | Cluster 02High ValueLight Spenders | Cluster 03High ValueOccasional Buyers | Cluster 05High ValueAbout to Sleep | Cluster 12High PotentialLight Spenders | Cluster 20Early StageAbout to Sleep | Cluster 25Low ValueAbout to Sleep |
|---|---|---|---|---|---|---|---|
| Categories bought | |||||||
| Útiles | 6.2× | 6.0× | 1.9× | 3.1× | 2.8× | 0.9× | — |
| Vinos Finos | 5.2× | 4.7× | 2.1× | 2.9× | 2.9× | 1.3× | — |
| Ropa Interior | 5.1× | 5.6× | 1.7× | 2.5× | 3.1× | — | — |
| Calzado | 5.1× | 4.6× | — | 2.4× | 2.8× | — | — |
| Remeras y Camisas | 4.4× | 4.9× | — | — | 2.9× | — | — |
| Remeras y Blusas | 4.1× | 4.2× | — | — | — | — | — |
| Ropa de Abrigo | 2.7× | 3.0× | — | — | 2.8× | — | — |
| Pantalones y Jeans | 3.1× | 3.1× | — | — | 3.0× | — | — |
| Ropa Bebés | 3.0× | 2.6× | — | — | 1.8× | — | — |
strong affinity (2.5×+)
moderate (1.5–2.5×)
mild (1.0–1.5×)
avoids (0.7–1.0×)
strongly avoids (<0.7×)
no signal
Each cell is a lift: how much more (or less) likely this cluster is to buy that category, use that payment method, or shop at that store, compared with the average customer. Green means the cluster is drawn to it; red means it avoids it; a value of 1.0× is exactly the base rate. Blank cells mean too few customers to measure. About to Sleep clusters show little distinctive signal — their members have bought too little, too long ago, to leave a clear pattern.