The Behaviour Business. Richard Chataway

Kahneman says: “A stupid decision that works out well becomes a brilliant decision in hindsight.”

Making search easy: Google

      In Google’s case, while behavioural insight drove the fundamentals of the product, part of their success was also sheer, dumb luck – it is a fairly open secret that the original Google interface was so clean and simple because Larry Page didn’t know how to code in HTML.

      Despite this, such was the effectiveness of the approach that the Google homepage remains largely unchanged in 20 years (aside from updated branding and the ‘Google doodles’48).

      The Google homepage in 1998 (top) and now (bottom)

      Source: Hooked by Eyal, N./Google

      One example of how Google makes the experience of searching as cognitively easy as possible is the choice architecture of providing two simple buttons: ‘Google Search’ and ‘I’m Feeling Lucky’.

      Have you ever actually used the ‘I’m Feeling Lucky’ button to conduct your Google search? Me neither. My friends at Google assure me this isn’t unusual. A 2007 analysis concluded that fewer than 1% of all search queries are conducted using that button.49 When users do click it, they go straight to the top result page, skipping the results listings. That means that Google shows zero ads (and therefore gets zero ad clicks) on 1% of all Google search queries. The analysis concluded that this button costs Google as much as $100m per year in lost ad revenue.50

      Yet the button still remains. If ‘I’m Feeling Lucky’ loses Google money when people use it, and hardly anyone uses it anyway, then why is it still there? Nostalgia? The anecdotal feedback from Google insiders is that this placebo choice remains because it subconsciously implies that Google will always give you the best possible result.51

      This subtle nudge, unchanged in 20 years, gives users faith in search results, and increases their self-efficacy (confidence and trust) in the brand to provide the right solution – as with the banking customers and the re-framing of security checks example in chapter 3. It is Google’s way of saying that no matter which button you press, it will deliver the result you want. Google is there (it implies) to serve your needs as a user – rather than its advertisers.

      Subsequent Google innovations, such as auto-completing queries, spelling corrections and so on, are also based on further improving the cognitive ease of the user experience. This includes using data on your previous search behaviour to make the results you see tailored to you (one change to the homepage since 1998 is the ability to sign in to your Google account). That personalised experience makes it even more effortless, as we like things more if we think they are unique to us.52 All of which serves to drive our Google addiction, and make it the go-to search engine for 85% of internet users.

Using data to leverage social proof: Netflix

      This personalisation effect has also been harnessed by Netflix to achieve market dominance using social proof – our innate desire to follow the herd. They have deployed machine-learning algorithms and customer data to create bespoke experiences for each user – the Netflix pages that you and I see will be different – and also to highlight what other people are doing.

      Much has been made of the binge-watching phenomenon, how it has changed the nature of TV viewing, and how specific content makes the site uniquely addictive rather than the site itself. Undoubtedly, if the content on Netflix was solely repeats of ‘Allo ‘Allo and Heartbeat (like some other channels I can mention), it would not have caught on quite so well with younger TV viewers.

      But far more relevant is the fact that more than 80% of all content watched on Netflix is actually based on its recommendation engine. That is, content recommended to a viewer according to a percentage-based ranking, based on things previously watched – effectively a numerical social proof value.

      Example of Netflix recommendations

      Source: Netflix

      No one outside Netflix knows exactly how that percentage matching works (is it based on what other people who liked that also watched? Or who stars in/writes/directs the show? Or genre?).53 The mechanism for it was copied by Netflix from dating sites, which provide matches based on compatibility with other users (such as shared interests).

      There is huge potential for this combination of data and behavioural insight to be applied in many other contexts to deliver the most compelling content.

      “Netflix are looking at the content itself, the storyline, the characters, a whole range of data points about you and the content,” says Steve Thompson, an experienced digital training consultant who works with media businesses across Europe. “And then using that insight to create future content. As a user you will also start to see recommendations down to your consumption of the content. They can easily use machine learning to nudge people into choices about what content they choose in the future.”

      The net effect is a kind of evolutionary natural selection of content – with only the strongest (i.e. most popular) content surviving, but maintaining sufficient variety to keep us interested. This keeps the content gene pool sufficiently broad – and making the best possible product overall.

      Netflix not only personalises recommendations, but how they recommend it. The artwork chosen to promote a particular show will be personalised to what is most appealing to each user, featuring different images or actors based on what is most likely to be clicked on.

      “We don’t have one product but over a 100 million different products,” say Netflix. “With one for each of our members with personalized recommendations and personalized visuals.”54

      As a result, users implicitly trust the mechanism simply because it takes the pain out of having to choose what to watch. In 2020, the paradox of choice of watching TV is that there is such a huge number of channels and content available it becomes an impossibly hard ‘choice-maximisation’ problem.

      Netflix solves it by providing a simple, ‘choice-satisficing’ solution: if you liked that, you’ll like this. And what do you know? Three out of four times that simple nudge works, along with others, like automatically starting the next episode and the ability to skip the credits, and keeps people watching.

      The paradox of choice

      The paradox of choice is a term coined by Professor Barry Schwartz (and the title of his 2004 book). Schwartz’s thesis is that in modern, westernised societies, we are now overwhelmed with abundant choice in almost every aspect of modern life: utilities, healthcare, pensions, beauty, work, love, religion, identity. But the paradox is, counter-intuitively, that increasing the choices available does not make us happier, nor make us more likely to choose the option we will like best.

      “Autonomy and freedom of choice are critical to our wellbeing, and choice is critical to freedom and autonomy,” says Schwartz. “Nonetheless, though modern Americans have more choice than any group of people ever has before, and thus, presumably, more freedom and autonomy, we don’t seem to be benefiting from it psychologically.”55