2016 was supposed to be the year of conversational commerce. Chatbots were heralded as the new interaction model through which interactions between consumers and brands would be transformed.
Fast-forward a few months and despite many impressive demonstrations on platforms like Facebook Messenger and Slack, it seems that chatbot technology has been looking for a problem to solve, rather than the other way around. Compounded by ‘many brands dropping their bots, saying they didn’t do what they were supposed to’¹.
It’s no secret that we are all spending more time on our mobile phones, but the data shows that we aren’t as engaged and we are spending less money via this channel compared to desktop².
This is a problem for both consumers and brands. If consumers are increasingly spending their time on mobile, then brands need to create experiences that engage them, and ideally make them spend money. Chatbots, so far, haven’t been the answer.
The problem for brands is real.
The reason I start with chatbots is because we began Qubit’s latest project with the hypothesis that a chatbot might be the answer to helping consumers buy on mobile. However, our research of the issue identified that this was not the case.
To understand why chatbots haven’t fully taken off, let’s start with the user experience. It’s important to say that some types of engagement with a brand are perfect for the early chatbot platforms - for example when the consumer knows exactly what they want and trust the brand to deliver it. In addition, customer service requests (i.e. Where’s my order?) where a chatbot can operate as a transaction mechanism, can be addressed efficiently and elegantly. In fact, the rise of Amazon Alexa and other voice platforms really demonstrates that there is a need for this type of interaction (the question here is, does a consumer really want to type on their smartphone?).
The flip side is when the consumer doesn’t know what they want. We call this ‘discovery’. The number of ways enabling discovery - presenting products in different categories or classes, in any number of combinations and permutations - could be infinite (although in reality constrained by the size and scope of the category and product catalogue in general). The reality is that not even a human can service the needs of a consumer when you’re talking several thousands SKU’s. The problem is compounded on mobile by the lack of screen real-estate which makes browsing through product pages even more challenging. This is what we call the big small-screen problem.
If we look at this history of content on the internet, the challenge of organising information was initially solved using lists. For anyone who remembers using the internet in the dial-up modem days: Yahoo, Altavista, Lycos and Excite organised information in directories. These were human-compiled lists of websites around a specific information area like Music or Tech. Then along came Google, who organised information in a far more useful way - understanding that all the information on the web was too varied and too broad for lists.
Search engines have ultimately come to define how we find content, even in commerce. We all know that Google is great for finding information when we know what we are looking for. But, what about when we don’t know what we want?
Everyday, we spend more time on social media platforms than eating or drinking³. Facebook, Instagram, Pinterest, and many other platforms are designed to show content that consumers did not know existed and organise it by its relevance to the end user.
These platforms are ultimately providing discovery and inspiration, and they keep customers coming back. In the case of Facebook, it’s helping consumers discover friends, news articles, and status updates. Instagram is focused on sharing photos and portfolios, and Pinterest focuses on hobbies through imagery. As Ben Evans, from venture capitalist firm Andreessen Horowitz, eloquently comments – “Facebook recommends stories it thinks (based on its machine learning model) you might like. There's not really an equivalent for products."*
On acknowledging this challenge, the team at Qubit set out ways to solve it. Along the way, we engaged both customers and consumers. Performed many, many iterations of design concepts and took inspiration from other industry-leaders who are seeking to solve discovery. Along the way, we soon realised that solving the discovery problem is vast and complex – with a number of different requirements.
Firstly, the most important input is data. Not just random data, but data that is structured, very purposefully, so that it can provide a level of understanding of the consumer’s interests, preferences, and intent - and then match that information with a brand’s products and services. Structuring the data so that it can be used is an incredibly complex first step.
Intentionally structuring the data means you can apply artificial intelligence (AI), the ‘magic’ that matches consumer desire to a brand’s products. The artificial intelligence required to create output that is relevant to the user is very specific, with generic AI models incapable of operating out of the box. AI models have to be applicable to the data that is being inputted so that you get the right kind of output.
Secondly, you need to look at the user experience. When customers come to your ecommerce website, in most cases, they probably already know what they want – and that is where Lists and Search work. As observed by many industry experts, including Jason Del Rey at Recode, ecommerce is not good at serendipity*, and increasingly, customers arrive needing inspiration and ideas – in the same way that a store window may be used to entice passing customers into store. The difference with what is possible with the data described above is that it is possible to adapt and change the store to every shopper based on their intent and past history. This level of personalization will be new to many consumers, and will feel mysterious at first to consumers, but if we get it right – and the early signs are very positive – the opportunity is significant, as has been commented by leading strategy consultants, Boston Consulting Group* and McKinsey*.
Today, we are launching an exclusive beta of Qubit Aura, a brand new product with multiple patents-pending. The goal is to provide consumers with inspiration and discovery throughout their ecommerce journey. Our initial tests have already been successful – but this is a complex space – so we are limiting the number of spaces available to perfect the product. We are looking for innovative, early adopters, to come aboard our beta program to be part of discovering the next step of the journey.
If you want to be first, click here to find out more.
* Boston Consulting Group https://www.bcg.com/publications/2017/retail-marketing-sales-profiting-personalization.aspx* McKinsey http://www.mckinsey.com/business-functions/digital-mckinsey/our-insights/marketings-holy-grail-digital-personalization-at-scale