Retail analytics
Retail analytics shows which stores are growing once new ones are taken out, which lines to reorder or mark down, and what a customer spends across store and online. Stores are compared over the same trading weeks under a written estate rule, and store and online customers are joined on one ID.
Like-for-like sales, example
Sales up 5.0% in total, down 1.0% like-for-like
-
Total sales growth, whole estate
+5.0%
-
Take out 2 new stores
−18.0 points
Running total −13.0%
-
Add back 1 closed store
+8.0 points
Running total −5.0%
-
Add back 1 refitted store
+4.2 points
Running total −0.8%
Refit: shut four weeks in August, out of both years this quarter -
Rebase to the 10 like-for-like stores
−0.2 points
Running total −1.0%
Rebase: −£40K over the 10 stores' £4,000K, in place of £5,000K -
Like-for-like growth
−1.0%
| Stores | Last year | This year |
|---|---|---|
| 10 like-for-like stores | 4,000 | 3,960 |
| 2 new stores | 0 | 900 |
| 1 closed store | 400 | 0 |
| 1 refitted store | 600 | 390 |
| Whole estate | 5,000 | 5,250 |
Show the numbers
| Step | Points | Running total |
|---|---|---|
| Total sales growth, whole estate | +5.0% | |
| Take out 2 new stores | −18.0 | −13.0% |
| Add back 1 closed store | +8.0 | −5.0% |
| Add back 1 refitted store | +4.2 | −0.8% |
| Rebase to the 10 like-for-like stores | −0.2 | −1.0% |
| Like-for-like growth | −1.0% |
The numbers that run a retail business
Stock is read line by line, because a sales total hides both the lines that have run out and the ones heading for markdown. Sales in every row exclude VAT and are net of returns.
| KPI | Formula | A good reading | The common mistake |
|---|---|---|---|
| Like-for-like sales growth | (This period's sales − the same period last year) ÷ last year, counting only stores that traded through both periods under the written estate rule |
Positive with the estate rule printed under it, and with online sales in or out as stated | Comparing calendar months, where one year has an extra Saturday. Compare the same trading weeks. |
| Sell-through | Units sold ÷ (opening stock + units received), per line, over the weeks it has been on sale. This row counts opening stock. Some retailers divide by units received alone. |
On the line's planned curve for its week of the season, at full price, before any markdown starts | Reading it per category. A line that sold out in week two can sit beside one that hasn't moved and average to a fine-looking figure. |
| Weeks of cover | Units in stock ÷ average weekly units sold over the last four weeks, per line |
Longer than the line's reorder lead time and shorter than the weeks left in its season | Averaging over weeks when the line was out of stock. Sales in those weeks were near zero, so cover looks longer than it is. |
| Gross margin after markdown | (Net sales after markdowns and promotions − cost of goods sold) ÷ net sales |
Close to the margin planned at the buy, with the gap traced to markdown line by line | Reporting the intake margin set at the buy as if it had been achieved, before the season's markdowns have landed |
| Average transaction value | Net sales ÷ transactions, from the tills |
Rising without units per transaction falling | Counting a refund at the till as a transaction, which pulls the average down in the weeks returns come back |
| Units per transaction | Units sold ÷ transactions |
Read by store and by hour, against the staff on the floor | Counting carrier bags, gift wrap and other non-stock items as units |
| Store conversion | Transactions ÷ visitors counted at the door |
Read by hour, so a dip can be matched to staffing or queue length | Dividing by a door count that includes people walking out. If the counter counts both directions, use the inward count only. |
| Shrinkage | (Stock value on the books − stock value counted) ÷ sales over the same period, at cost |
Falling from one count to the next, read by store and category | Booking the whole loss in the month of the stock count, when it built up over every month since the last count |
GMROI, stock turn and open to buy are on retail KPIs, and there are more examples in the KPI library.
Two new stores can turn a shrinking estate into 5% growth.
Example data
Where the data for retail reporting usually lives
The ecommerce analytics page covers contribution per order for the online side on its own.
Analytics services for retailers
The trading view needs every store's till data and the merchandise system's stock to land each day, and data monitoring and support checks each morning that they did and names any store whose tills did not report.
We agree the like-for-like rule with you before the first report, covering refits, new stores and online sales, and the dashboard prints it under the figure.
AI in retail analytics
AI-led analytics answers "which lines have less cover than their reorder lead time?" once weeks of cover leaves out-of-stock weeks out of its average, and every answer shows its query. The AI for retail page works through lines with stock on the books that have stopped selling, and staffing against door counts.
Questions from retailers
Do you build dashboards for retailers?
Yes. The weekly trading view opens on the like-for-like bridge, then shows sell-through and weeks of cover for every line, with the lines running out first and each store's estate rule written on the dashboard.
Can you do retail predictive analytics?
Yes, starting with demand per line. Out-of-stock weeks are flagged first, because they record what was on the shelf, and every forecast is scored against last year's same week, on weeks it never saw, before you reorder on it.
How do you treat a store that was refitted?
On this page's rule, a store shut for a refit leaves like-for-like for any quarter it was shut in, in both years, and rejoins the quarter after it reopens. Taking out only the shut weeks also works, if the rule is written down.
Should online sales count in like-for-like?
It's your choice. Counting online sales reflects customers who browse in store and buy online, and leaving them out keeps the figure about the shops. If the board wants both, one bridge shows them as two lines.
Do you do customer analytics for retailers?
Yes, once store and online purchases are joined. A till sale ties to a person only through a loyalty card or emailed receipt, so lifetime value and segments are built on that matched share, with the share printed beside them.
Which of your stores belong in the comparison?
On the call we go through your stores with the dates each opened, closed or shut for a refit, and agree which ones your like-for-like should include.