To drive sales through true relevance, you need true automation
Many e-commerce technology suppliers claim to offer ‘automated merchandising’ solutions. But most still rely on the automated application of hundreds of ‘rules’ – rules created and managed manually. Only true, intelligent automation, shaped by AI and machine learning, can deliver true relevance, and the transformative bottom line benefits it enables.
Imagine a world in which everyone uses self-driving cars – but self-driving cars without perception of the situation.
The cars know what their passengers want, and they have rules to follow in order to operate robotically in line with that individual desire. However, they have no external view, so they have no idea what other cars are doing.
It would be chaos. Even if individual cars were technically well driven – that’s the easy bit - they’d be constantly crashing into each other.
The problem, clearly, is that these cars have no way of knowing what other cars around are doing, what they are likely to do next or where the car is on the map.
Sounds crazy, right?
You wouldn’t run a car on that kind of tech, much less a business. But that is exactly what many online retailers the world over are doing.
Automation in name only
The vast majority of ‘so called’ e-commerce merchandising automation technologies work just like those crazy cars. They look at individual intent and use rules to try to deliver a superficially relevant experience.
Or they establish rules over the automatically generated results, in a belief that human intelligence still beats anything else, including computers.
As with those malfunctioning cars, what is missing is a view of crowd behaviour. Of context.
It simply does not work. It’s automation in name only. Rather than true relevance, you get a pale imitation - a guess at individual intent based on a gross simplification of human behaviour - which is taken entirely out of context with wider trends and the wisdom of the crowd.
Take a simple example of rule-based merchandising automation, which essentially draws on a series of ‘IF/THEN’ rules:
IF visitor = woman AND month = June THEN recommend beach fashion
Is that simple logic really going to be true of every female visitor? Clearly not. So how can the recommendations it delivers really be relevant to each visitor? Even that superficial relevance is an illusion – one that inevitably harms the customer experience and therefore sales.
But what is the solution to that theoretical car issue? More rules? If we were still in the 1930s and the roads were largely deserted, maybe. But the sheer scale of the real time data processing required means cars controlled by rules simply wouldn’t work on a car expected to be intelligent.
To function properly they must be adaptable, intelligent and critically be able to learn continuously in real time.
The same goes for e-commerce merchandising automation. Any solution that is not intelligent is not true automation, and will not deliver true relevance which – by its nature – must adapt in real time and reflect the full complexity of human behaviour.
'Dumb automation': The five key problems
Rule-based approaches, in essence, suffer from some major flaws:
- They over simplify – data points like gender and season alone, for example, are clearly insufficient when it comes to predicting and reacting to likely shopper intent
- They customise too much to a particular 'behaviour' – true relevance comes from understanding crowd behaviour, not that of one individual
- Rule-based algorithms rely on human intervention (for instance, to define the length of the summer season)
- Rule-based solutions also quickly grow out of control, with too many rules to manage effectively – change one rule and the consequences for other rules (good or bad) are unknown until it’s too late
- Rules and the people working with them simply cannot cope with the complexity of data quickly enough to react in real time to changing context and behaviour.
Apptus eSales: Intelligent automation
At Apptus, we specialise in intelligent automation. We help retailers to rip up the rulebook, completely removing the reliance on rules and replacing them with advanced, AI-driven automation that learns, reacts and adapts in real time to deliver true relevance.
It’s proven to transform sales performance and cut costs, all while freeing merchandisers from the thankless task of managing hundreds, even thousands of rules.
At its heart, Apptus eSales is about the relevance of digital merchandising – which products are shown to which shoppers, and when. The entire process is automated, with only high level human input required – specifically, decisions around whether digital merchandising should optimise for conversion, revenue or profit.
This is possible because Apptus eSales calculates product exposure not according to rigid rules, but according to a set of integrated AI-algorithms that account for individual visitor behaviour across the site, for all visitors, in real-time.
Made in Sweden: True Automation. True Relevance
Apptus eSales is the only solution that enables online retailers to optimise the exposure of every product in every context, across the entire product range – via clearly defined, holistic product exposure strategies and intelligent automation.
A mainstay of leading Swedish retailers, the automated product exposure approach has proven far more effective, efficient and adaptable than the ‘tip of the iceberg’ exposure delivered by manual or rule-based approaches to online merchandising.
Automated digital merchandising strategies draw on real-time visitor behaviour, context and crowd wisdom to expose the right products to the right people at the right time.
It’s proven to work. To find out more about intelligent automation from Apptus eSales, get in touch.