Buying the Best Plans

In the harsh economic climate that we are all now facing, fashion retailers are still wrestling with the perennial problems associated with selecting the right merchandise in the right quantities to avoid the twin pitfalls of markdown and stockouts. Optimising the merchandise planning process has never been more relevant.

In many cases this may be the only available method of improving margins – or even of survival. Computer programmes for planning can greatly increase buying efficiency.

The recurring problem facing merchandisers throughout the fashion industry, is to find a balance in buying quantities at style and colour level that accurately reflect the sales pattern that will be achieved. It is universally accepted that in any range there will be winners and losers, and it is therefore logical that the buying pattern should reflect this. How, though, can the best course be charted, given the sheer volume of decisions required in forecasting? The ubiquity of personal computers  makes sophisticated Merchandise Planning Decisions Support Systems accessible to all but the very smallest independent retailers.

These can produce buying forecasts at style and colour level that accurately reflect sales patterns. The key to the success of these systems is a dramatic reduction in the number and nature of decisions required of the merchandiser. The systems perform the majority of the number crunching that has historically acted as a drag on the whole process, and leave the buyer better information on which to base decisions.

Such systems typically consist of a set of basic inputs, including historical sales data, seasonal structure, product hierarchy structure, top-level budgets, percentage participation by season and category. There are also basic outputs such as buying quantities by style and colour.

The inputs and outputs can be built around a `rank curve’ which shows the percentage of sales achieved by a given percentage of styles. Based on empirical evidence, a similar percentage of sales can be expected for a given percentage of styles for each season for separate groups of merchandise.

Rank Curve
For example, it is common for the first 10 per cent of styles to achieve about 28 per cent of the sales. It is therefore possible to predict with reasonable accuracy the sales levels of the styles in order of performance. The question then facing merchandisers is which are the best sellers rather than looking at buying quantities in isolation.

An extension of the principle allows retailers to forecast how many styles should be made available in a given garment type. If the retailer expects to sell 100,000 units and has a minimum order quantity of 250, then the lowest selling style will account for at least one per cent of sales. By using the shape of the curve, it is possible to deduce how many styles this minimum will allow – say 15.

The system would then work out the quantities per style to be bought and the merchandiser would finish the task by ranking the sales potential of the top 15 styles available. A further refinement would allow the forecast to be edited by the buyer.

Being based on historical data such systems also become more accurate from year to year, the data being “cleaner” each time the process is followed.

Verified by MonsterInsights