Returns and credits
The flip side of revenue: rates, outliers per item, credits on the timeline.
Which items come back conspicuously often – and does it actually cost us anything? In online retail that question decides margin and range. And it is surprisingly hard to answer as long as every credit note is treated as a return.
Not every credit note is a return
The first step was order. Pythia now distinguishes kinds of credit notes: return, credit without rebooking, address correction, settlements with partners, and wholesale. Only the first two count as customer credits – money that really goes back to customers.
Cancelled lines are out of every figure. What never happened should count neither as a sale nor as a return.
Credit notes on the timeline
Credit notes can be shown as a layer of their own – by credit date or by the date of the original sale. The second view is the more revealing one: it shows how much of a month actually stayed sold in the end.
In the candle the credit appears as a hatched cap. The lower edge of the hatching is the net value. The tooltip gives credit, rate and last year’s rate, and the key-figure tile calculates net throughout when the layer is switched on.
The returns page
A dedicated page shows the return rate by channel, range, product group, item, returns warehouse or transaction type – by value, units or lines, with year-on-year comparison and drill-down to the single item.
Underneath lies the real benefit: outliers per item, separated by quality, over a rolling twelve-month window. Pythia checks three signals:
- Does the item come back clearly more often than its own range?
- Does it come back clearly more often than last year?
- Does the second or third quality come back clearly more often than the first quality of the same item?
The list is sorted by excess returns in euros, not by percentage. A high rate on an item that barely sells is not a problem. A slightly raised rate on a bestseller is.
Every finding has a small trend line. Confirming a finding creates a quality event – visible on the timeline – and the finding leaves the list of open points. Factoring in the cost of a return itself is planned.
Next month: when the details have to be right – bundles, parcel carriers and an expectation that can explain itself.