Over the past two decades or so eCommerce has undergone a major change. The speed of this change has been, and continues to be, amazing.
A Short History
The world wide web got invented in 1989, became available on the Internet in 1991; first web sites humbly started as additional marketing channels, with corporate web sites giving information on the company and a product catalogue; 1994 the Netscape browser arrived. In 1995 Amazon got founded – and the story started to take off. In 2000 we saw US online shopping eclipsing the 25 billion dollar mark. 2003 we saw iTunes and in early 2005 there was already talk about ‘fully personalized shopping experience’. Read this!
Not to forget that Google got founded in 1998 and Facebook in 2004 – and with them data collection and its –use for advertising started to happen en masse.
In 2007 the iPhone arrived, followed by the iPad in 2010 – and a whole lot of additional demands got created. mCommerce arrived. Device friendly sites and eCommerce presences became mandatory.
Multi Channel needed (and still needs to) become omni-channel. About 2012 Big Data became mainstream although the term has been around for longer.
2016 then became the year of AI, conversational commerce, and bots. And these technologies may very well be the saving grace for eCommerce. Will they?
The Customer Side
In parallel, partly caused by what is possible and partly simply because of rising demands, customers developed an ‘I want it right now – and right here’ expectation; this is combined with a fading trust into institutions that to some extent got fuelled by too much scatter-gun advertisements. Add the meteoric rise of social media. This led to the (wrong) notion that the ‘customer is in control’ and to the need of analyzing lots – LOTS – of data and to rapidly deliver customers contextually relevant information – at the right time and into right channel.
Today, customers expect their experience to be flawless; they expect that relevant information be easily found, that recommendations coming from businesses are geared to their needs and that no irrelevant ‘clutter’ is displayed to them.
Of course they expect that help be in easy reach at any time, if they need it.
Where does AI come into the Picture?
Of course, the first thing that comes into mind is the provision of intelligent product recommendations. There are the Amazon’s and Netflix’s of this world who have built whole business models around delivering the right recommendation at the right time – and do this very well. And there are a number of solutions out in the market that do exactly this for eCommerce sites.
An important aspect is to combine this ability with other channels, in order to not only incentivize the cross- and upsell, but also repeat business. Intelligent product recommendations are a powerful tool to invite customers to become return customers.
AI supported predictive and prescriptive analytics, supported by an automated delivery of the determined content through the right channel helps to keep customers engaged and to then drive business. Delivery could be into the site in real-time to drive cross- and upsells, or at a time and through the channel that is likely to get the best possible engagement in an attempt to re-engage for further business or to convince a customer to re-engage with an abandoned basket.
Similarly, an eCommerce site can offer individualized, or personalized, pricing, which is a specialized form of dynamic pricing that determines a price based upon the customer and the company’s objectives. The company can, e.g. give a one-time price discount for an unknown shopper who seems to be on the edge of buying, or give just the little incentive to a known loyal customer as a reward. Knowing the customer this incentive could even be a free accessory that the company knows they are interested in, instead of a discount.
In parallel, and based upon a user’s current and past behavior the eCommerce site could offer a chatbot offering support to find a suitable product or configuration. This can be in the form of a support service or be a virtual agent as it is, e.g. offered by Dominos in Australia or New Zealand, where a customer can use a conversational interface, supported by Natural Language Processing to configure and order pizzas. This can, especially with a larger number of products, greatly increase the speed of ordering. Add an Alexa/Siri/Cortana-like voice system and things become really interesting.
Or how about having an AI-based configurator that helps creating a configurable product? Look at Bear Naked Granola. Using this site one can build an own flavor of Granola with some help of IBM Watson. On the B2B side CPQ is another example where AI is applied. Salesforce added Einstein functionality to their CPQ solution, while Apttus invested in AI for their CPQ solution. This can result in offering the best possible solution and quote for a customer.
A bit more on the esoteric side one finds solutions that offer AI-driven web site optimization, like Wacul, while research already looks into fully adaptive web sites – sites that automatically improve their organization and presentation for individual users by learning from the users access patterns of the site and probably additional ones.
Simpler solutions just translate the content of a web site. Try Google translate on the (Japanese) Wacul site. You will be amazed what this AI is capable of doing.
What Does It Mean?
There are many possibilities to use AI and machine learning to improve an eCommerce site, the most obvious ones being highly targeted product recommendations, pro-active decision support and the friction reducing possibilities that a conversational interface offers.
This in an integrated manner across channels (to avoid the term omni-channel). The overarching goal is to make the eCommerce experience as simple and frictionless as possible for the customer.
So, will AI be the future of eCommerce?
ECommerce does not need a savior and AI will surely contribute to providing new, more convenient, and exciting experiences for the customer that lead to better results for the business. Realtime and intelligent actions and reactions will make the difference in future and turn eCommerce into iCommerce.
Better be on the right side of it.
Working off an existing eCommerce presence business leaders therefore should:
- Proactively offer automated help in situations where the customer seems to become unsure about how to proceed, using an intelligent agent as opposed to pushing a chat window onto the user’s screen
- Have a deeper look at the possibilities that hyper-personalized recommendations offer, providing customers with exactly the recommendations they would need in support of their intended purchase
- Look into offering a conversational experience for customers that enables them to avoid browsing a category tree but give some ideas of their own and receive matching suggestions, also considering browsing and purchasing history. As in-store, let your customer ask for a ‘silky, blue, light summer dress, sized 38’ instead of browsing the catalog, studying the product description and then finding out that the size is currently out
- Consider personalized pricing as part of the customer acquisition- and loyalty strategies. There can be benefits for both, offering the hesitating unknown customer and the well known high value customer.
- Intelligently use their web site logs, having an intelligent agent continuously suggest improvements to the site that are geared towards further reduction of customer friction
And keep an eye open to offering customers an individually unique brand experience. The data is there – the automation intelligence (soon), too.
Great article, Thomas! One question: can you drill down for me on how IBM Watson comes into play in helping select granola? What information does Watson actually provide to customize the experience? I see Watson mentioned with various AI for customer service and am still foggy on where it comes into play.
Hi Jeremy, thanks heaps!
Watson does a seemingly simple thing: Suggest ‘matching’ ingredients based upon what flavours you already chose. It is doing this starting from a molecular level. The other thing it does is telling you how well the ingredients that you chose yourself will fit to each other, creating a tasty granola. Overall it is an application of science to food. The interesting thing about it is that tastes are different, so there needs to be a feedback loop, too. Still, it is a noteworthy application – especially if you are willing to be somewhat experimental 😉
Cheers
Thomas
@twieberneit
Thank you, Thomas! That’s fascinating. If Watson makes better granola I’m all for it 🙂
yes, it is interesting – although I am a little suspicious about Dr. Watson’s culinary taste when looking at some of its suggestions. But then I might be a bit conservative and not so much of a risk taker 😀