Skip to main content

How to win online retail’s zero-sum game

For years, almost since the birth of ecommerce, online merchandisers have been playing a zero-sum game – a game of give and take in which the simple act of promoting one set of products has the effect of demoting others.

In essence, that is a consequence of three things.

Finite real estate

First, screen real estate is not only finite; when measured per visitor, it is tiny.

Compared to physical stores or paper catalogues, levels of product exposure online represent a fraction of most retailers’ actual offering. As a result, selling online isn’t like physical retail – the space available for merchandising is extraordinarily limited – and that demands a different approach, especially as channel and device proliferation render customer behaviours ever more fragmented.

But unlike in physical stores, there is no need to show the same things to every customer.

Infinite merchandising options

Second, given the scale of many retailers’ product catalogues these days, the range of options in terms of promoting and merchandising products is virtually limitless.

The maths here is straightforward, yet mind-boggling. Take even a relatively modest category of 500 products and the total possible display permutations are practically endless: a number running to 1134 digits. In the real world, no retailer even begin to scratch the surface of this vast range of possibility. For one thing, it is impossible to do so when employing a largely manual approach to merchandising. For another, it’s probably not desirable to do so: to simply try every possible combination is not a very smart approach.

The smart approach is to draw on contextual information to narrow down the options and, ultimately, find the optimum approaches to site-wide merchandising. The question is whether you rely on intuition and sales feedback, which inevitably limits merchandising in both scope and agility – or widen the possibilities through real time automation.

A Zero-Sum Game

Agility and scope are, however, crucial bearing in mind the nature of online merchandising’s zero-sum game:

  • The amount of screen real estate is finite, and the merchandiser’s ability to increase it is limited.
  • So when a merchandiser promotes one product or campaign, it inevitably leads to the demotion of other products or campaigns. One in, one out.

What’s more, online merchandising isn’t simply the digital equivalent of shelf stacking, where every product has a place and every place a product. It is a process of continuous optimisation, where the status quo is consistently challenged by alternative courses of action, and not just on a page-by-page basis, but customer-by-customer.

Out-dated Segmentation and the Role of Online Merchandising

Unfortunately, however, few retailers have recognised this need to optimise customer-by-customer and fewer still have the capacity to truly deliver such granular approaches to merchandising. As a result, most retailers are playing the zero-sum game the ‘old way’.

They are merchandising only a handful of products at a time, and only on very limited, high traffic real estate. They are effectively merchandising in silos – a campaign here, personalised recommendations there – and this disjointed approach makes it very difficult to know if one is ‘winning’. That is, whether merchandising is improving sales performance overall or, by demoting the wrong products, hampering it.

What’s more, in most cases customer-by-customer means a limited application of personalisation – for instance personalising only recommendations. There is no doubt that personalisation works, but the vast majority of product exposure happens within search and category lists, not recommendations panels.

The winning strategy

Winning at online merchandising’s zero-sum game requires a radically different approach – an approach with intelligent automation at its heart.

Indeed, it is an approach that is impossible without automation, because the sheer scope of merchandising executions and agility it requires cannot be delivered the old fashioned way.

The key question here is not how to win the game, but how many games to play. That is, do you play one game, look for the ideal mix of products for an imaginary average shopper, or play several, segmenting according to arbitrary typologies and personalise some areas of the real-estate?

Or do you do neither of these? Do you instead play almost as many zero-sum games as there are visitors to your site?

The truth is, in a fast-moving world of fragmented shopper behaviour, the right approach is to treat every shopper as their own segment - and serve each with their own highly relevant experience. The result is a merchandising operation that redefines the zero-sum game to optimise both product exposure and product relevance for every shopper.

  • Relevance: Personalising the retail experience down to ever more granular segments maximises the opportunity to deliver relevant experiences and merchandise relevant products
  • Exposure: Product exposure is maximised too, because individually personalised merchandising is no longer limited by screen real estate, but by the number of eyeballs looking at it.

Intelligent Automation

Clearly, however, this kind of approach presents some fairly significant challenges for ecommerce merchandising departments still labouring to maximise product exposure the old way - largely manually and with tools that do not support the flexibility, agility and scale required to deliver truly personalised experiences for every shopper. Not just that, but experiences that learn from and adapt to crowd and individual behaviour in real time.

There is only one way to deliver merchandising on that scale – one to one, whole site personalisation – and that is through automation. Not just automating basic workflow tasks, but automating entire merchandising operations, with human input focused where it should be – on high level, strategic issues rather than the day to day of which products are displayed where.

Today, very few retailers have this capability. Those that do have adopted a new generation of merchandising automation tools, which draw on big data, cutting edge machine learning and predictive analysis to deliver true personalisation – which means maximum product exposure, maximum individual relevance and the ability to adjust everything in real time.

It is only a matter of time before the rest of the market catches on.

Michael Mokhberi, CEO of Apptus

Image source: Shutterstock/mtkang