We used 541,909 public retail line items to move from a transaction dump to an operating picture: who really drives revenue, what the average hides, where demand concentrates and why a static forecast fails. This is a CADi business-intelligence exercise - not causal inference and not a claim about a client.
POSITIVE-SALE VALUE
£10.67m
20,002 observed orders
REPEAT PURCHASES
93.1%
of known-customer positive-sale value comes from repeat buyers
CUSTOMER CONCENTRATION
32%
of known-customer positive-sale value comes from the top 1%
DEMAND VARIANCE
16.6×
daily variance versus mean; homogeneous Poisson fails
THE EXECUTIVE READ
01 · REPEAT PURCHASES
Repeat means more than one observed order in this window, not a retention rate. Repeat buyers dominate known-customer positive-sale value; this does not establish lifetime value or a campaign effect.
02 · CONCENTRATION
That creates key-account upside and concentration risk at the same time. A useful dashboard should expose both, not celebrate one blended average.
03 · ORDER TAIL
The top 1% of orders contribute 19.3% of sales. One £168,470 bulk order can move a product leaderboard.
04 · PEAK SEASON
The September–November surge versus June–August is mainly a throughput problem. Capacity, fulfilment and stock availability matter more than a premium-basket story.
05 · IDENTITY BLIND SPOT
Identified-customer orders average £480. Anonymous orders are larger on average; the cause of missing identity is not observed.
06 · FORECASTING
In the separate UK arrival cohort, a Poisson 90% interval covers only 7.1% of later days. Seasonality and overdispersion are candidates for further testing, not validated fixes.
WHEN GROWTH ARRIVES
Positive-sale value rises 66.2% from the June–August baseline to September–November. Order count explains most of it: +52.1%, versus only +9.3% in average basket value. The operating question is therefore capacity under clustered demand.
Gross positive-sale value by source month. Both December 2010 and December 2011 are partial.
MARKET SHAPE
The UK supplies 84.6% of gross value, but order economics differ sharply abroad. Netherlands and Australia orders are several times larger than UK orders - small markets with a different operating profile, not merely smaller versions of home.
Average positive-sale order value
ASSORTMENT SIGNAL
Largest unit-volume item
UNITS
80,995
ORDERS
1
VALUE
£168.47k
One bulk order creates the top unit-volume product. A naïve “best seller” ranking would confuse a one-off wholesale event with broad customer demand. Frequency and concentration belong beside every product leaderboard.
THE DATA BLIND SPOT
IDENTIFIED ORDER AVERAGE
£480
ANONYMOUS ORDER AVERAGE
£1,219
Only 7.2% of orders lack a usable customer ID, yet they carry £1.76m of gross value. Customer-level retention and concentration metrics are therefore informative, but not a complete view of the commercial base.
WHAT THIS SUPPORTS
