Segmentation and profiles

Who your customers are, what each group is worth, and what to do about it

6,000 customers grouped into 5 segments on how they buy. Full-price regulars are 27% of customers and 57% of margin.

Prepared for a constructed retailer trading in GB, DE, FR, US. Every figure below was generated by the model from a constructed population when this report was produced — not any firm’s data or result.

How these figures were produced — the model

Your questions, answered

What a customer strategy needs to know, in a line each

Who are our most valuable customers?: Full-price regulars: 27% of customers, 57% of margin. Protect them.
Where is margin leaking?: Lapsing customers bring £40 each a year against a book average of £133. On average they last ordered 33 weeks ago.
How concentrated is our value?: Half of all margin comes from 16% of customers.
Does one strategy fit every market?: No. DE earns £24.09 a customer less than the book, mostly because of who it has, not how they behave.
What do we do first?: Full-price regulars: protect. Then Occasional big baskets: grow frequency.
Are the segments still true?: A later snapshot has moved: the stability index is 0.16, past the level the firm watches at — see how the segments have shifted.

Background and approach

Customer profiles built from what customers do

The objective: accurate, actionable profiles of the customer base, built from real transaction behaviour rather than assumption — and a way to place every new customer in one.

What the segmentation reads

Behaviour	Book average
Weeks since last order	10.6 weeks
Orders a year	10.0 a year
Average order value	£73.78
Margin rate	23%
Share of orders on promotion	35%
Share of value returned	13%
Share of spend in fashion	46%
Share of spend in home	33%
Share of orders online	52%

How the segments were found

Customers were grouped on these behaviours together. The number of segments was not chosen by eye: 7 different counts were tried, and 5 was the count whose segments came back most reliably when the data was redrawn, beyond what data with no structure in it produces. Every customer is then placed by a short rule a CRM can run, which agrees with the segmentation on 96% of the book.

Key findings

Value is concentrated, and it is not where revenue is

Customers6,000in the book

Margin a customer£133a year, after discounts and returns

Half of all marginfrom 16%of customers

Top tenth of customers36%of all margin

Average order£73.7810.0 orders a year

On promotion35%of orders

Returned13%of value

What it means

Half of all margin comes from 16% of customers. Losing a small number of them costs more than any acquisition campaign is likely to replace.
Full-price regulars are 27% of customers and bring 57% of margin — the group a loyalty programme exists for.
Lapsing customers bring £40 each a year against a book average of £133. On average they last ordered 33 weeks ago.
DE earns £24.09 a customer less than the book, and most of that is who it has — more customers in low-margin segments — not how they behave.
A later snapshot has moved: the stability index is 0.16, past the level the firm watches at — see how the segments have shifted.

The segments at a glance

Share of customers against share of margin

Where the bars differ, a segment brings more or less than its size. Shares are rounded so each set sums to exactly one hundred.

Share of customersShare of margin

Show as a table

	Share of customers	Share of margin
Full-price regulars	27%	57%
Discount-led	25%	9%
Occasional big baskets	15%	23%
Heavy returners	14%	5%
Lapsing customers	19%	6%

Full-price regulars

Customers: 27%
Margin: 57%
Role: Protect

Discount-led

Customers: 25%
Margin: 9%
Role: Re-price

Occasional big baskets

Customers: 15%
Margin: 23%
Role: Grow frequency

Heavy returners

Customers: 14%
Margin: 5%
Role: Fix returns

Lapsing customers

Customers: 19%
Margin: 6%
Role: Win back, selectively

Full-price regulars

Full-price regulars — 27% of customers, 57% of margin

Protect

The numbers

Customers: 1,613
Share of customers: 27%
Share of margin: 57%
Margin a customer a year: £281
Orders a year: 15.3
Average order: £57.67
Margin rate: 34%
Orders on promotion: 11%
Value returned: 7%
Orders online: 44%

How they buy

Orders about 15.3 times a year and spends £57.67 an order; the last order was 3 weeks ago on average, against 11 for the book. 11% of orders are on promotion (book: 35%) and 7% of what is bought comes back (book: 13%). Spend is split across categories much as the book’s is, and 44% of orders are online. After discounts and returns, each customer leaves £281 of margin a year.

Against the book

Each behaviour as an index: the book average is one hundred.

Show as a table

Behaviour	This segment	Book average	Index
Weeks since last order	3.3 weeks	10.6 weeks	31
Orders a year	15.3 a year	10.0 a year	152
Average order value	£57.67	£73.78	78
Margin rate	34%	23%	149
Share of orders on promotion	11%	35%	33
Share of value returned	7%	13%	50
Share of spend in fashion	40%	46%	86
Share of spend in home	35%	33%	107
Share of orders online	44%	52%	85

What to do

Recognise rather than discount: early access, service and a loyalty tier. A discount here gives margin away on orders that were coming anyway.

Occasional big baskets

Occasional big baskets — 15% of customers, 23% of margin

Grow frequency

The numbers

Customers: 876
Share of customers: 15%
Share of margin: 23%
Margin a customer a year: £211
Orders a year: 4.4
Average order: £170.12
Margin rate: 30%
Orders on promotion: 16%
Value returned: 6%
Orders online: 29%

How they buy

Orders about 4.4 times a year and spends £170.12 an order; the last order was 11 weeks ago on average, against 11 for the book. 16% of orders are on promotion (book: 35%) and 6% of what is bought comes back (book: 13%). Spend leans to home (69% of it), and 29% of orders are online. After discounts and returns, each customer leaves £211 of margin a year.

Against the book

Each behaviour as an index: the book average is one hundred.

Show as a table

Behaviour	This segment	Book average	Index
Weeks since last order	10.7 weeks	10.6 weeks	100
Orders a year	4.4 a year	10.0 a year	44
Average order value	£170.12	£73.78	231
Margin rate	30%	23%	132
Share of orders on promotion	16%	35%	44
Share of value returned	6%	13%	46
Share of spend in fashion	15%	46%	33
Share of spend in home	69%	33%	211
Share of orders online	29%	52%	57

What to do

Reminders timed to this group’s buying cycle and cross-sell within the category it already buys, to add a visit a year without discounting the basket.

Discount-led

Discount-led — 25% of customers, 9% of margin

Re-price

The numbers

Customers: 1,523
Share of customers: 25%
Share of margin: 9%
Margin a customer a year: £50
Orders a year: 10.8
Average order: £42.59
Margin rate: 12%
Orders on promotion: 75%
Value returned: 10%
Orders online: 60%

How they buy

Orders about 10.8 times a year and spends £42.59 an order; the last order was 5 weeks ago on average, against 11 for the book. 75% of orders are on promotion (book: 35%) and 10% of what is bought comes back (book: 13%). Spend is split across categories much as the book’s is, and 60% of orders are online. After discounts and returns, each customer leaves £50 of margin a year.

Against the book

Each behaviour as an index: the book average is one hundred.

Show as a table

Behaviour	This segment	Book average	Index
Weeks since last order	5.4 weeks	10.6 weeks	51
Orders a year	10.8 a year	10.0 a year	108
Average order value	£42.59	£73.78	58
Margin rate	12%	23%	53
Share of orders on promotion	75%	35%	214
Share of value returned	10%	13%	77
Share of spend in fashion	50%	46%	108
Share of spend in home	25%	33%	77
Share of orders online	60%	52%	115

What to do

Stop blanket discounts to this group. Test fewer, targeted offers and non-price incentives, and judge them on margin rather than orders.

Lapsing customers

Lapsing customers — 19% of customers, 6% of margin

Win back, selectively

The numbers

Customers: 1,122
Share of customers: 19%
Share of margin: 6%
Margin a customer a year: £40
Orders a year: 3.3
Average order: £48.07
Margin rate: 28%
Orders on promotion: 30%
Value returned: 8%
Orders online: 40%

How they buy

Orders about 3.3 times a year and spends £48.07 an order; the last order was 33 weeks ago on average, against 11 for the book. 30% of orders are on promotion (book: 35%) and 8% of what is bought comes back (book: 13%). Spend is split across categories much as the book’s is, and 40% of orders are online. After discounts and returns, each customer leaves £40 of margin a year.

Against the book

Each behaviour as an index: the book average is one hundred.

Show as a table

Behaviour	This segment	Book average	Index
Weeks since last order	33.0 weeks	10.6 weeks	311
Orders a year	3.3 a year	10.0 a year	33
Average order value	£48.07	£73.78	65
Margin rate	28%	23%	122
Share of orders on promotion	30%	35%	85
Share of value returned	8%	13%	62
Share of spend in fashion	45%	46%	98
Share of spend in home	31%	33%	93
Share of orders online	40%	52%	76

What to do

Contact only the customers a contact is shown to save — the retention model ranks them — rather than everyone who has gone quiet. Many of this group leave whatever is sent.

Heavy returners

Heavy returners — 14% of customers, 5% of margin

Fix returns

The numbers

Customers: 866
Share of customers: 14%
Share of margin: 5%
Margin a customer a year: £47
Orders a year: 13.2
Average order: £94.51
Margin rate: 7%
Orders on promotion: 35%
Value returned: 45%
Orders online: 91%

How they buy

Orders about 13.2 times a year and spends £94.51 an order; the last order was 4 weeks ago on average, against 11 for the book. 35% of orders are on promotion (book: 35%) and 45% of what is bought comes back (book: 13%). Spend leans to fashion (85% of it), and 91% of orders are online. After discounts and returns, each customer leaves £47 of margin a year.

Against the book

Each behaviour as an index: the book average is one hundred.

Show as a table

Behaviour	This segment	Book average	Index
Weeks since last order	4.3 weeks	10.6 weeks	41
Orders a year	13.2 a year	10.0 a year	132
Average order value	£94.51	£73.78	128
Margin rate	7%	23%	30
Share of orders on promotion	35%	35%	100
Share of value returned	45%	13%	339
Share of spend in fashion	85%	46%	183
Share of spend in home	8%	33%	24
Share of orders online	91%	52%	175

What to do

Returns erase most of what this group spends. Size and fit guidance, clearer product detail and a look at the returns policy come before any campaign aimed at them.

Prioritising the segments

Where the margin is, against how much of it is bought with discount

Up is more margin a customer; right is more of their orders on promotion. Bubble size is the number of customers. Top left is value to protect; bottom right is margin being given away.

Show as a table

	Orders on promotion	Margin a customer a year	Size
Full-price regulars	11%	£281	1,613 customers
Discount-led	75%	£50	1,523 customers
Occasional big baskets	16%	£211	876 customers
Heavy returners	35%	£47	866 customers
Lapsing customers	30%	£40	1,122 customers

In order of the margin at stake

Segment	Share of margin	Role	Why
Full-price regulars	57%	Protect	Recognise rather than discount: early access, service and a loyalty tier. A discount here gives margin away on orders that were coming anyway.
Occasional big baskets	23%	Grow frequency	Reminders timed to this group’s buying cycle and cross-sell within the category it already buys, to add a visit a year without discounting the basket.
Discount-led	9%	Re-price	Stop blanket discounts to this group. Test fewer, targeted offers and non-price incentives, and judge them on margin rather than orders.
Lapsing customers	6%	Win back, selectively	Contact only the customers a contact is shown to save — the retention model ranks them — rather than everyone who has gone quiet. Many of this group leave whatever is sent.
Heavy returners	5%	Fix returns	Returns erase most of what this group spends. Size and fit guidance, clearer product detail and a look at the returns policy come before any campaign aimed at them.

Variance by market

The same segments, in different proportions, behaving differently

Each market’s gap to the book’s margin a customer, split exactly into two: more customers in low-margin segments (mix), and the same segments bringing less (rate). They call for different fixes — acquisition and targeting for the first, the offer for the second.

Who the market has (mix)How they behave there (rate)

Margin a customer a year, against the book.

Show as a table

	Who the market has (mix)	How they behave there (rate)
GB	+£9.01	−£1.35
DE	−£23.45	−£0.64
FR	+£8.48	−£1.63
US	+£4.35	+£3.75

DE is furthest below the book, by £24.09 a customer, and the larger part is mix. The largest single rate effect is Heavy returners in FR, who bring £20.26 a customer less there than the same segment does across the book.

Segment mix by market

Full-price regularsDiscount-ledOccasional big basketsHeavy returnersLapsing customers

Show as a table

	Full-price regulars	Discount-led	Occasional big baskets	Heavy returners	Lapsing customers
GB	30%	21%	16%	14%	19%
DE	18%	39%	13%	13%	17%
FR	30%	22%	15%	16%	17%
US	29%	21%	15%	14%	21%

Is it still true

How the segments have moved since they were fitted

Segments describe the customers only as of the last time anyone checked. A later snapshot is placed by the same rule and compared. Both columns are placed by the written rule, so the first can differ by a point or two from the shares above, which come from the fit itself.

Fitted book, by the ruleLater snapshot

Show as a table

	Fitted book, by the rule	Later snapshot
Full-price regulars	29%	24%
Discount-led	25%	43%
Occasional big baskets	14%	10%
Heavy returners	14%	10%
Lapsing customers	18%	13%

watch

A later snapshot has moved: the stability index is 0.16, past the level the firm watches at — see how the segments have shifted. The firm watches at 0.10 and acts at 0.25.

Next steps

From segments to activation

Load the placement rule into the CRM so every customer, including new ones, carries a segment daily.
Protect — Full-price regulars: Recognise rather than discount: early access, service and a loyalty tier. A discount here gives margin away on orders that were coming anyway.
Grow frequency — Occasional big baskets: Reminders timed to this group’s buying cycle and cross-sell within the category it already buys, to add a visit a year without discounting the basket.
Re-price — Discount-led: Stop blanket discounts to this group. Test fewer, targeted offers and non-price incentives, and judge them on margin rather than orders.
Read every activation against a group that did not receive it, sized before launch. A campaign judged on before-and-after cannot tell itself from the season.
Re-run the segmentation when the stability index passes 0.10, and re-fit when it passes 0.25.

Methodology

How this was produced

Customers were clustered on the nine behaviours above, each standardised so that none dominates because of its unit.
The number of segments is the one whose segments return most reliably under resampling, beyond what structureless data of the same shape gives; the smallest count within one standard error of the best is kept.
Segment names come from the behaviour that most sets each segment apart from the book, by fixed rules. Profiles describe behaviour: the data holds no demographics, so no biography is offered.
Shares are rounded by largest remainder, so every set printed here sums to exactly one hundred.
The population is constructed. Every figure was generated by the model when this report was produced.

The model: fit, stability and validation

	Full-price regulars	Discount-led	Occasional big baskets	Heavy returners	Lapsing customers
GB	30%	21%	16%	14%	19%
DE	18%	39%	13%	13%	17%
FR	30%	22%	15%	16%	17%
US	29%	21%	15%	14%	21%

	Full-price regulars	Discount-led	Occasional big baskets	Heavy returners	Lapsing customers
GB	30%	21%	16%	14%	19%
DE	18%	39%	13%	13%	17%
FR	30%	22%	15%	16%	17%
US	29%	21%	15%	14%	21%

	Full-price regulars	Discount-led	Occasional big baskets	Heavy returners	Lapsing customers
GB	30%	21%	16%	14%	19%
DE	18%	39%	13%	13%	17%
FR	30%	22%	15%	16%	17%
US	29%	21%	15%	14%	21%

Customer segmentation — sample report