(c)iStock.com/mbbirdy
There’s a perception in eCommerce that a focus on boosting selling merchandise is a solid strategy. That is perfectly logical, reasonable — and wrong.
Infact, boosting selling goods is a complete waste of time, effort and money, how it is different from the offline world and which springs from a fundamental misunderstanding of the challenge that is eCommerce.
Intelligent exposure
Item marketing via recommendation and search builds on one of retail techniques — the best seller list, which appeared in ‘The Bookman’ at London in 1891. The success of the seller list because context was built on three simple truths. To begin with, people purchase what is popular; second, showing people the best sellers makes them buy more; and third seller lists invite people to purchase from the top of the listing, not the tail.
Offline, you will find many reasons for this – space for the need to keep down production costs and stockimply the imperative is to streamline requirement. But eCommerce presents challenges, and needs a different perspective of marketing. So the imperative is easier — maximise sales, online, space is practically unlimited. As a result, marketing is not about promoting a focus on the sellers only; it is about maximising product vulnerability. Within this context, intelligent means related exposing more of the right goods more of this time, to the individuals.
In terms of the timeless top seller list, our proxy for product promotion here, those online priorities change the rules. Yes, the goal is to encourage people to buy more, however, the need to expose more goods, not fewer, necessitates that we add a key new ingredient to the mix – volatility.
Online, computing top vendors over a period of time of a month is erroneous. Rather recommendations and research must reflect both sellers and tendencies — these short lived, but sales spikes driven by external events that are hard-to-predict. Volatility is crucial to capitalising on these events, once the time is appropriate and for the right period of time.
Exploitation versus exploration
In essence online promotion, via recommendation and search is about striking the Ideal balance between exploitation and exploration:
- Exploiting the very best sellers to keep consistent sales and also to preserve their effect as the all important basket openers
- Encouraging exploration farther into the product catalogue, to drive additional sales while gathering the data that helps identify emerging trends and predict future best vendors.
Crucially, that exploration should be formed by a mixture of feedback to recognize trends and business metrics to guarantee advertising makes commercial sense.
The bandit problem
As it happens, this exploitation versus exploration conundrum is neatly encapsulated in a probability theory scenario referred to as the multi-armed bandit problem. Basically, the multi-armed bandit problem refers to a gambler or one-armed bandits, that must decide which machines to play.
Is involving the manipulation of the machine with the maximum anticipated pay-out, in order to maximise the opportunity and exploring the other machines.
In eCommerce, focusing on vendors is like ignoring all the other machines, and pulling on one lever. The challenge in this variant on the difficulty isn’t to find the slot machine, yet to find.
This keeps the upsides of marketing top vendors, but also empowers the benefits or adding volatility to promotion are discounted.
The winds of change
Manufacturers have yet come to terms with these differences between online and offline merchandising implemented approaches that use them.
It is only a matter of time. The rest will follow, once the results appreciated by the few that have taken the direct are frequently understood.
One such pioneer is adlibris. It took on the challenge of this transforming a essential component of online merchandising – using real time behavioural data and more duration sales volumes to forecast top vendor trends and therefore surface more relevant products, all the time, also for the correct period of time. The results were so extraordinary.
Since it drove a spike in her autobiography, the announcement of this 2014 Nobel Peace Prize given the test, I’m Malala. And Adlibris’ merchandising model passed with flying colours. It drove merchandising and efficient navigation across multiple classes — even autocomplete – and merchandising was optimised in terms of timing. The trend has been encouraged for exactly the time interval — which was reflected in A/B test results.
In search, this fresh way of merchandising in the internet context drove revenue increases of 6.4% (more than two successive AB-tests), while the amount of search refinements decreased by 41 percent. The solution delivered a return.
In an environment where retailers are questioning the worth of plug standalone and play merchandising solutions, it is only a matter of time until this kind of holistic approach is the norm, not the exception.
To read more about the theory behind adlibris’ victory, read the paper http://info.apptus.com/apptus_banditproblem on which this article is based.
