Perceived Variety

What it is

A shopper cannot count the items in a category and compare the total with a rival store's. The shopper forms an impression, and an impression can be driven by a few cues.

Susan Broniarczyk, Wayne Hoyer and Leigh McAlister studied two of them in 1998. The first is whether a favourite product is still available. The second is the amount of shelf space devoted to the category. Their question was how consumers form assortment perceptions when the number of stock-keeping units falls, and their abstract reports that retailers might be able to make substantive reductions without negatively affecting assortment perceptions and store choice, as long as only low-preference items are eliminated and category space is held constant.

The distinction matters because a retailer's costs and a shopper's judgement are measured on different scales. The retailer counts items. The shopper reads the shelf. A range can lose items and keep its look, or keep its items and lose its look if the cues change.

In effect

The practical reading is a rule about order. Remove the items shoppers least prefer first, keep the space the category occupies, and the cues the shopper relies on stay in place. Later work summarised the size of that room for manoeuvre as cuts of up to 54 per cent of low-selling items with no significant effect on perceived variety or sales.

The cues cut both ways. Removing an item that shoppers do prefer takes away the first cue, and cutting the shelf space takes away the second, so the same number of items removed can do different amounts of damage depending on whose favourites they were and what was put in their place.

Perceived variety also sits at the centre of the dispute that followed. Boatwright and Nunes (2001) found a sales rise after an online grocer cut items in 94 per cent of its categories. A 2005 study of the same grocer's records for every category found customers ordering less often, with less frequently bought categories hurt most. Whether the shopper's impression of variety held up is one thing, and whether the shopper kept coming back as often is another, and the 2005 loss showed up in the second.

What it does not say

It does not say that the number of items never matters. Boatwright and Nunes found that category sales depended on the total number of items offered, and on whether key attributes, brand and flavour above all, stayed available.

It does not say the 1998 conditions hold for every store or every category. The two studies were of grocery categories, and the conditions they name, low-preference items only and constant category space, are conditions, not guarantees.

It does not measure variety directly in the later field studies. They count sales and visits, so a range can look as varied as before and still lose customers for another reason.

And it says nothing about menus or product pages, where the shelf is a screen or a printed list and the space cue works differently.

How a shop owner can use it

The idea turns a range review into two lists. The first is the items to remove, ordered by how few shoppers prefer them. The second is the cues to protect, which on the 1998 evidence are the favourites and the space. A review that starts from the second list asks a different question from one that starts from a target number of items to cut.

It also changes what to watch afterwards. If the cues held, the range should still look as varied to the shopper, and any loss will show elsewhere. The 2005 grocery records suggest where, in how often households order, and in the categories that are bought less often. That is a suggestion drawn from one grocer's data and not a rule, and it applies with most force to shops that sell by category.

Why the term travels badly

The word variety is used for two things that need to be kept apart. One is the range as the stockroom sees it, a count of lines. The other is the range as the shopper sees it, an impression built from a handful of cues. The studies on this page measured the second, and the retailer's cost sheet measures the first. When a report says that a range was cut by half without anyone noticing, the useful question is which of the two it measured.

The same care applies to the number quoted from the 1998 work. Cuts of up to 54 per cent of low-selling items come from a later summary of earlier studies, and the conditions that came with it, low-preference items only and category space held constant, are part of the finding.

---