A number of businesses are sharpening up their algorithms so as to better serve customers and understand their individual needs
Time-poor consumers can’t find what they’re looking for online, so instead they’re leaving it in the hands of someone else
Information is easy to come by when shopping online; it’s expertise that’s lacking. Every product description on the internet is accompanied by a list of details designed to make sure customers know exactly what they are getting, but that’s rarely the problem. The frustration comes when a customer realizes they’ve bought the wrong tool for the job.
Perhaps the biggest problem for online retailers is their inability to offer customers insight into their purchases. The ten-second conversation that gives a shop assistant enough information to make a recommendation doesn’t exist online. Instead, customers are left to do their own research – which can easily end in a mistake. Increasingly, though, retailers are putting the humanity back into online shopping.
An example of this trend in action is the subscription box. A business picks a theme, and customers subscribe to receive regular deliveries of a curated selection of goods. The range available today is staggering. Boxes can be as specific as international candy, moustache-themed stationary or different kinds of crystals. The exact content of a box is often a surprise until it’s delivered. Many allow consumers to set preferences, but whether a Bocandy Box contains a green tea flavored Kit-Kat from Japan or a Hershey’s bar from America is up to the curator. It’s not always a small-scale operation. Loot Crate, founded in 2012, has a subscriber base of 500,000 ‘looters’ who receive an assortment of themed pop-culture paraphernalia each month, chosen by the company.
Grumpy subscribers didn’t kill Columbia House, but they gave it a real battering
There needs to be a great deal of trust between the company and the consumer for these transactions to work – something subscription sellers are acutely aware of. Trunk Club assigns customers a personal stylist who, through interviews, makes selections based on individual tastes. It’s a pricey service, though it cuts out bad selections and shows that there is still a premium on expertise.
For bigger online retailers the game is very different. The amount of transactions they deal with makes individual customer-service near impossible. Instead, they are crowd-sourcing knowledge.
The Filter is a UK-based company founded in 2004 that provides the algorithms and platform for companies to use data to curate content. One of The Filter’s board members and initial investors is Peter Gabriel. In an interview with CNET in 2008, he said that his interest in the company came from a lack of tools available to people trying to navigate the volume of content online. “Maybe there is a deeper yearning out there for freedom from choice,” he was quoted as saying.
Rather than leave decision-making in the hands of an individual, The Filter is driven by the nearly unlimited-data points that online companies have access to. It’s incredibly responsive to customers with the right information. The Filter CEO James Routley said it lends itself particularly well to music playlists.
“We can learn your taste profile, and your taste profile can vary depending on different contexts. It can vary based on time of day, based on location, it can vary based on the device you’re using. So we will build your general taste profile, but that taste profile can be segmented by device and time of day, and location as well. Actually, in fact by any other context we can feed in and segment.”
While initially The Filter focused purely on the numbers, increasingly it has combined raw data with human input to lever individual expertise. An equation can learn a new fashion trend, but it can’t yet predict it. In this regard at least, The Filter isn’t much different from traditional retail. A shop assistant might learn a trend, analyse it, and change their future recommendations. It’s the difference between offering last season’s fashion instead of next season’s. Amazon is often seen as an industry leader in this regard, but Routley said they are making some fundamental mistakes.
“Because of the size of their catalogue it’s very hard for them to get around every single possibility. For example, go add a pair of jeans to your shopping basket and they’ll say ‘okay, by the way, how about these five different pairs of jeans while you’re shopping.’ What that does is have the effect of making the customer less sure of their purchase. They’ve probably done the research, they’ve put a pair of jeans into the basket, and now Amazon is saying here are five other pairs you want to consider. That’s not an ‘as well’ purchase in a lot of cases. They’re introducing uncertainty, when really they should be trying to check the customer out or offer them a belt, a t-shirt and pair of shoes.”
Nonetheless, Amazon seems confident its algorithms could eventually lead to something akin to clairvoyance. In 2014 it filed a patent for a system that would allow it to ship items to customers pre-emptively. If you read the first Harry Potter book and reviewed it highly, the rest could be on your doorstep the next day. Routley said that this is a very long way off, and it would probably frustrate customers more than anything.
“It reminds me of the old record club model.”
Grumpy subscribers didn’t kill Columbia House, but they gave it a real battering. The mail-order music club offered incredibly generous introductory offers, such as 12 records for a penny, throughout the 70s, 80s and 90s. Subscribers agreed to purchase a certain number of mail order albums per year after the initial deal. Random selections were sent out if you didn’t express a preference and the bill came after you got your delivery, a process called negative-option billing. People were naturally frustrated when they had to pay music they didn’t want, and negative-option billing is now illegal in some parts of the world.
Like the score to a movie, good curation isn’t easily noticed, but bad curation is. Routley said there are common traps for businesses to avoid.
“I think customers do get fed up with seeing the same things over and over again, so one important thing that we do is we learn from feedback, negative feedback, so if somebody isn’t clicking on a recommendation then it shouldn’t keep coming back. We don’t respond well to being offered things we’ve already bought or already watched. So if you’ve just watched a film or a television program it is automatically removed from the recommendations for a period of time.”
People will make mistakes when shopping – it’s unavoidable. But it happens online more than it should. Individual expertise might be impossible on a large scale, but the virtual equivalent of ‘Ask the Audience’ might be the answer for creating more boutique online-interactions. It saves customers time they don’t have, and brings that bit of traditional retail-knowledge today’s online marketplaces sorely need.