Segments
Who is buying, how often, and what each group is actually worth.
Share of customers against share of revenue
Same segments, same colours, two bars. Where the bands do not line up is where the value is.
Segment detail
| Segment | Customers | % of base | Revenue · SAR | % of revenue | Value index | AOV · SAR | Orders / yr | Retained | Growth |
|---|---|---|---|---|---|---|---|---|---|
| High-value loyalistsThree or more orders, above-median basket, bought within 60 days. | 17.9k | 13.6% | 2.8M | 34.1% | 2.51× | 661 | 6.3 | 74.8% | 19.2% |
| ReplenishersRepeat the same SKU on a predictable cadence. | 19k | 14.4% | 2M | 23.9% | 1.66× | 478 | 4.7 | 69.1% | 15.5% |
| Discount-ledTwo thirds of their orders carry a promo code. | 24.5k | 18.6% | 1.5M | 17.5% | 0.94× | 378 | 3.1 | 36.3% | 15.7% |
| One-and-doneSingle order, no second purchase within 120 days. | 44.4k | 33.7% | 1.1M | 13.4% | 0.40× | 319 | 1 | 9.9% | 3.2% |
| Lapsing regularsWere on cadence, now 1.5× past their own interval. | 26.1k | 19.8% | 919.5k | 11.1% | 0.56× | 315 | 2.3 | 23.1% | 25.5% |
Segment transitions
Where every customer moved between 8 Jul 2026 and 7 Aug 2026. Ribbons take the colour of where the customer started; the faint ones held their segment.
Revenue per segment
Absolute contribution over last 30 days
High-value loyalists are worth 2.51× an average customer — and are 13.6% of the base.
131.6k active customers split across 5 tracked segments. The distribution is the finding: high-value loyalists are 13.6% of the base and 34.1% of revenue, while one-and-done are 33.7% of the base for 13.4%. Paid acquisition is currently priced against the blended average of the two, which is a price neither of them is worth.
The current source audience is every buyer, which trains the model toward the modal customer rather than the valuable one. Swapping the seed costs nothing and changes who the platforms find.
They absorb retargeting budget at a conversion rate below the prospecting pool. Suppressing them re-prices the whole retargeting line.
The median gap between first and second order is 18 days; the drop-off begins well before that. A single prompt placed ahead of the decay window is the highest-leverage lifecycle change available.