Nudgestock, Ogilvy’s annual festival of behavioural science, is highly unusual in its ambition and scope. The day is about developing creative solutions to human problems. We’re here to understand people in all their complexity and consider what unites them in their decision making.
Behavioural science is a young discipline, still mapping its boundaries. So instead of being spoon-fed a narrow diet of industry veterans re-presenting case studies, we’re invited to harvest our inspiration from thought-leaders from across domains of expertise. It’s a challenging approach, demanding your attention and sapping your brainpower.
A succession of anthropologists, economists, academics, marketers, brands and campaigners take the stage. I turn to the people either side of me to make sense of it all: I’m sat next to an evolutionary anthropologist and a food entrepreneur. I can’t even rely on industry jargon to bridge the gap in my understanding our as we chat through our impressions of each speaker.
I’m out of my comfort zone, and loving it. Here are a few things that stuck in my head:
1) Many problems are better solved using “psycho-logical” rather than logical thinking
Rory Sutherland anchors the event. Beaming from the stage in a Hawaiian shirt, his talk is a whistle stop tour of his new book, Alchemy. His argument: logical thinking is a straight-jacket when you are solving human problems. We should all consider “using psycho-logic to conjure up value from nowhere.” A host of examples show us how looking through a psychological lens enables magical creative solutions:
- Instead of spending £80bn on HS2 to reduce journey times to Birmingham by 30 minutes, spend a fraction of that making the experience more pleasurable and productive;
- Make paying tax less painful by keeping rates constant (we become accustomed to the rate we pay) and instead giving an annual rebate. You then have option of donating a proportion to causes of your choice, like the NHS (feelgood factor). Those sharing over 50% could have their names publicised (an opportunity both for kudos, and to signal identity).
2) There is nothing wrong with doing good by stealth
Huge improvements to human hygiene in the 20th century were in a large part caused by a growing urge to maintain the appearance of cleanliness, unconscious status-seeking. Soap was sold on its ability to increase your attractiveness, rather than its hygiene powers. Positive ends derived from selfish means.
There is much we can learn from “Scenting the soap”. If you can change the story, you change the meaning and value of the product. In the present day Beyond Meat is framed not around a rational benefit or hairshirted denial, but exceeding (not just imitating) what it is replacing.
3) Make rational benefits irrationally appealing & “Don’t mess with Texas”
These themes were echoed by Richard Wise, a “global brand anthropologist” with a nice line in dry humour. He galloped through “five legendary case histories which made rational benefits irrationally appealing” which were all for “utterly worthless and contemptible categories”.
In each example, rational benefits were red herrings. Previous communication efforts often failed due to logical linear thinking. Brand owners struggled to see past them to connect to the human truths.
Anti-littering in Texas, for example, relied on Keep Texas Beautiful as a message, which failed to change behaviour (or in Wise’s words, “didn’t do shit”). Who litters most? Young men aged 16-35: pick-up driving, beer-drinking wrestling fans. The solution for the Lone Star State was to appeal to a more rebel instinct in the tone of their message, provoking tribal pride. This was about speaking to a lower self about a higher message. You can imagine the arguments getting it signed off.
The result? A 72% reduction in littering on highways over 4 years: the single greatest anti-littering campaign on record, subsequently adopted into culture, emblazoned on t shirts, meme-ified and parodied.
4) The 3 psychological hacks which Uber rely on
Taxis aren’t new. You’ve been able to summon a cab to pick you up for over a hundred years. What makes Uber different – and crucially feel different – is the user interface:
- The dot on the map reduces the uncertainty of waiting, reducing frustration. You can see where you are, where your driver is, and feel reassured. None of this changes objective reality: it changes the way you feel waiting for a cab.
- No money changes hands physically at the end of the ride – payment is pre-arranged and intangible, reducing the pain of payment.
- Chunking the customer journey into granular steps. The Uber Pool customer journey is longer, meaning people drop out. Showing updates like “now finding a driver going in your direction” and “finding other riders going in your direction” along the way resulted in a 11% reduction in ride cancellations. Chunking helps.
5) Stand up to quantification bias – unconsciously valuing the measurable over the immeasurable
Tricia Wang bounded on stage with barely suppressed energy. Wang is a tech ethnographer – someone who watches what people do with technology for a living. She coined the term “thick data” (which essentially means qualitative data) – “the most direct, unmediated data from humans which captures the full context of their emotions and stories”. As she put it “you cannot put tears and smiles into spreadsheet.”
Without paying close attention to human experience you miss emergent behaviours, unseen opportunities and strategic shifts – meaning your strategy will be wrong.
Wang believes companies make bad decisions because of quantification bias. They fall into a trap that their huge datasets hold the answer to what they should do next, rather than understanding people. She is famous for advising Nokia to change their reliance on featurephones after observing how smartphones were transformational to the lives of people living in the developing world. Nokia rejected the advice as their models predicted otherwise. Big data gives you an artifice of certainty: her experience taught her that without the right model you are not listened to.
6) Re-balance marketing functions to be more holistic
In Wang’s view, there needs to be a culture change in organisations. Thick data – qualitative insight from real people:
- Rescues the context loss that results from relying on big data alone;
- Inserts the human back into the loop that has become invisible undetected;
- Changes your perspective, allowing you to focus on the right things.
You need a holistic approach, scale and depth, to drive businesses forward. It is interdisciplinary: we need to work together. The marketing function should balance thick data, big data and creativity. In her view this is why Apple outperform Nokia, Fender outperform Gibson and Zara outperform Gap.
7) Simple rules of thumb outperform complex prediction models
A lot of people are here today to hear Gerd Gigerenzer speak; he’s spent 40 years studying how people make decisions in the real world. In his view: less is more.
- The world we live in is complex: we have imperfect knowledge of future states of the world, their consequences and probabilities;
- It’s counter-intuitive, but your aim should be to make things simple;
- Rules of thumb are more effective than all other approaches.
He compared two models of predicting whether a customer will make a future purchase:
Strategy 1 – Less is more – experienced managers used a rule of thumb: ”if a customer has not bought in the past 9 months classify them as inactive” (the “hiatus heuristic”).
Strategy 2 – A complex problem needs a complex solution. Use a pareto negative binomial distribution model.
The former was a more accurate strategy than the latter, across many categories (see pic). As Gigerenzer put it: “…simple solutions are out there, we just need to look for them. Often just a single variable is important…”
A simple “hiatus heuristic” required less information, was fast and saved effort. The implications are clear: speak to the doers, people on the front line more likely to have skin in the game experience. We should aspire to fast and frugal heuristics, not be ashamed of them.
8) The sentence that can predict flu with more accuracy than Google
Another example related to a flu prediction project at Google:
- Google created an algorithm to predict spread of flu. Their hypothesis was the volume of searchers would correlate with the number of ill people. They used 50 million searches centring on 45 search terms, with a 4-year calibration period. The model failed to predict the swine flu breakout in April 2009 – a black swan event with no precursors in the data: big data is backward looking. If the future is not like the past, this is a problem.
- To improve the model, they re-engineered the algorithm, with even more data: 160 predictors. Again, it failed, even with another 4-year calibration.
- What turned out to be more accurate was a simple rule of thumb, based on recency: comparing the number of flu related doctor visits 2 weeks ago to the number of doctor visits one week before flu diagnosis, in the previous year.
9) A lot of management decision making is about “arse-covering”
Defensive decision making is the notion people choose the option which protects themselves, rather than the best decision for their organisation. Many organisations have a culture of fear and blame: why take a risk? Far better to “fail conventionally than to succeed unconventionally” as John Maynard Keynes put it.
Gigerenzer said that many CEOs admit to him privately they approach decision making this way. This has serious implications:
“Defensive decision making strangles us because people can no longer do their best for you”
When it comes to medicine this can be the difference between life and death. He cited research showing 93% of doctors admitted to defensive decision making. If your doctor is motivated by protecting themselves from litigation rather than healing you, they end up recommending a different – interventionist – course of treatment. His rule of thumb? Ask them what they would do if they were choosing a treatment for themselves.
10) Organisations are monitoring people to death – decentralise and empower to win
Professor Sir Paul Collier from Oxford University closed the day. Talking without notes, he covered the birth of capitalism, how it developed, diagnosing where it is failing, before prescribing solutions – all in 30 minutes.
So I’ll focus on a single point. The annual jobs and skills survey has shown a sustained drop in agreement to the statement “Do you have enough discretion in your job to do it properly?” Since the 1970s there has been a 40% drop, and now most people disagree. People don’t feel trusted – but are micromanaged, targeted, tested and controlled. This simply does not work, in business (GM vs Toyota) or in the public sector (teachers in the UK vs Finland).
As our world becomes more complex, the amount of behaviour that can be codified reduces. So much of what we do at work relies on tacit knowledge. The lesson? Decentralise decision making, empower people and trust them. Work provides us with dignity, respect and esteem – we shouldn’t forget it.
Reflecting on it all, the common thread is our obligation to argue for, justify and reassert human connection.
Abstraction and data have a role to play, but in isolation they feel cold, insufficient. The relentless logic of businesses scaling-up risks humans being engineered-out completely. As Sutherland writes in Alchemy:
“It is impossible to quantify many of the psychological factors which people care about… there are no S.I. units for what really matters.”
So here’s to using our judgement, in business and in life.