GUEST COMMENT Online fashion: the three big issues only AI-enabled retailers can overcome
A recent IMRG report - Fixing Fashion
- included some facts and figures that will be all too familiar to a lot of people working in online fashion retail. In fact everyone from the merchandisers and the CRO guys through to the CEO in brands where crucial metrics are heading in the wrong direction.
For these fashion retailers, customer loyalty is on the decline, abandonment rates are up and conversion rates are falling fast – by 28% since 2010.
Much of this is self inflicted. Not every fashion retailer is suffering – just those that cling to an outdated way of selling online, a way that simply doesn’t work any more. These are the brands that still rely on a manual approach to merchandising. But this is being swept away by smarter, more agile approaches that draw on artificial intelligence and machine learning to transform sales performance.
The writing really is on the wall, and the issues that will increasingly impact on the late adopters are once again highlighted in that IMRG report.
The first is an inability to respond to trends anything like quickly enough. These days, socially-driven trends come and go in a matter of minutes or hours not days, weeks or months. The ability to ride these short-lived waves, by making on-trend products easy to find, is crucial.
Too few fashion retailers have this ability however. According to IMRG, fewer than a third can react to trends in real time, 63% take hours or days to respond and, incredibly one in 20 can’t respond at all.
Then there is the fact that hiring skilled people is getting harder and harder.
More than 90% of the retailers IMRG surveyed found it hard to recruit the people they need - and in a labour-intensive, manual merchandising world, this is a critical limiting factor.
It’s no coincidence that the people they found hardest to hire were those with the skills to understand customers (analytics) and then respond to that insight by delivering an outstanding customer experience (online marketing and UX).
Manual merchandising also promotes a focus on the wrong metrics. A retailer’s survival might depend on profit, but a merchandiser’s job is measured in conversion - a false proxy for retail success.
In my view, this focus on conversion is completely self-defeating. It creates an unhealthy focus on sales and promotions. Yes, slashing prices will help you convert more, but at what cost to profitability?
What really matters is profit, and that’s what should drive merchandising and, more specifically, product exposure strategies.
Man and machine
These three issues - too slow, too human, too conversion-focused - are absolutely behind online fashion retail’s declining performance.
The only solution is to rethink the role humans should play in merchandising - and how computers, specifically, artificial intelligence, can help shape a new kind of online retail.
This is only part of the evolution, in the same way tractors replaced horses - it is not about computers replacing people. It is about using computers where it makes sense; in gathering and analysing huge amounts of data, then testing and refining the merchandising response in real time.
There will always be the need for nuanced, creative, human input at a higher level - the longer term strategies that set brands apart and on which computer input should be based.
This is not a vision of some far off utopian future. Some retailers are already well on the way - they are adopting artificial intelligence-driven approaches to merchandising and ecommerce optimisation, and transforming performance as a result.
This is something brands like ShopDirect are talking about in the UK, and it’s something I have personal experience of from my team’s work with pioneering Swedish fashion brand, Stayhard
Stayhard has always been a pioneer when it comes to using technology to drive sales performance, so it was always likely to be early to the artificial intelligence party. In fact, it has been using Apptus eSales
- which is, to my knowledge, the only true artificial intelligence-enabled ecommerce optimisation solution on the market - since 2013.
It’s fair to say that it has been a critical component of the success that saw Stayhard from zero to multi-million euro annual turnover in less than a decade – all with a merchandising team of just three, supported by Distancify developers.
AI in action
It was the sheer volume of activity that delivered the real step change when we moved over to AI. For instance, before Apptus eSales Stayhard could only run two or three campaigns at a time and had to manually run A/B tests that took time to set up and lots of time to analyse. Now Stayhard runs 50 campaigns at the same time and gets automated performance reports for every one. As a result, it gets incredibly valuable insight that helps shape future campaigns.
With AI in place we could also help Stayhard move away from conversion as the metric. Instead, product exposure strategies are driven by business goals. In a sense, that’s because solutions like Apptus eSales don’t give you many controls - what would be the point of having lots of granular, manual controls in an AI solution?
What you get are the controls that matter, like the ability to select whether merchandising strategies are driven by high conversion goals or high revenue goals.
So, when Stayhard has a big sale it can live with having a lower profit. It can optimise its assortments for high conversion and it’s a strategic decision. The rest of the time, everything is optimised to deliver more profit.
The proof is in the numbers and there’s no doubt that Stayhard has bucked the downward trend. We ran a 60 day a test where we compared a strategy that exposed products to generate higher profit and a strategy to expose products that led to higher conversion- all managed by Apptus eSales. We could very quickly see that the profit strategy increased profitability by 3.5% – and that’s lots of money if you’re doing several million euros revenue per year.
It’s numbers like that, and the ease with which we delivered them, that mean it’s inevitable that online fashion will turn to AI. The innovators are already there so the question really is when the rest will take the plunge.Johan Davidsson is CEO at Distancify