close
close

How leaders can improve their strategic decision-making

Ever since Tesco Clubcard co-creator Clive Humby called data the ‘new oil’, there has been a huge emphasis on collecting as much as possible in the hope of making better, more informed decisions.

The consulting profession has been awash in articles and studies highlighting the value of data in guiding how we think and behave. For example, McKinsey claims that data-driven companies are nearly 19 times more likely to achieve above-average profitability than their instinct-driven counterparts, while EY reported that most executives today think data should underpin everything they do.

Researcher from HEC Paris believes that such an argument certainly makes sense. They investigated the use of artificial intelligence in the private equity and venture capital industries and found that the use of AI could not only increase efficiency but also change the way employees worked. One company, for example, marveled at how an AI-based decision support tool could quickly improve operational efficiency and provide intelligent recommendations to partners.

One would therefore be forgiven for thinking we were in a data-driven age, except that’s not really the case. For example, research from the University of London found that managers still like to trust their instincts more than data. This is even the case when making financial decisions, for which another study found that managers were actually less interested in using data to back up their guesses.

Research from INSEAD highlights how damaging this kind of instinct-driven decision-making can be, with investors found to have a positively rose-tinted view of their past financial performance. This distortion either made gains bigger and losses smaller, or saw investors selectively forget losses altogether. This tends to result in overconfidence, which can have disastrous financial consequences.

Actionable data

For data to be useful, it must be current, relevant, accurate, valid and complete, which too often is not the case – and this issue is compounded by the skills gap in interpreting the data. An attempt to lower the skills required to effectively use data is visualization. For example, researchers from MIT created a visualization to accompany their recent report on migration between Central America and the United States, arguing that visualization can help make complex root causes more accessible, both to policymakers and the public.

The use of visualization to better convey the message behind data had its most notable beginning during the Crimean War, when Florence Nightingale famously produced a series of “rose charts” to show the extent of improvements in British soldiers’ chances after she had implemented fairly basic hygiene improvements in field hospitals in Turkey .

“None but scientific men even look at the appendices to a report, and this is for the vulgar public…Now who is the vulgar public to have it?…The Queen…Prince Albert…all the crown heads in Europe, through the ambassadors or ministers of each … all commanders in the army … all regimental surgeons and medical officers … the chief sanitarians in both houses (of parliament) … all newspapers, reviews, and journals.”

Although the use of data visualization is more than 100 years old, it has come to the fore in our big data era, to the point that Accenture claims that the strongest players in the FTSE 350 were also the companies that mention visualization most frequently. Amid this somewhat thin evidence, Mark Kennedy of Imperial College London cautions that data visualization should not be seen as a panacea.

He argues that data visualization can indeed make data look more compelling and compelling, but if the underlying data behind the visualization remains of poor quality, it can encourage wrongdoing. Perhaps the simplest yet most powerful form of visualization is the humble checklist.

Consistent and repeatable

Checklists have been used for some time, but gained a lot of attention when Atul Gawande published his very successful The Checklist Manifesto. They provide valuable visualization thanks to their ability to convey often complex tasks in a series of manageable steps. Checklists are often used in situations where consistency is critical, even when tasks are performed by a large number of people, such as piloting an airplane or performing medical procedures. A number of digital tools have emerged that allow checklists to be integrated into more sophisticated task management workflows.

Gerd Gigerenzer argues that an equally crucial decision-making aid is the humble heuristic. These rules of thumb are easily dismissed in a big data era where analysis often takes precedence over intuition, but Gigerenzer believes that heuristics are fundamental to the kind of tacit knowledge that so often underpins what we do.

These enable the kind of quick decision-making that Daniel Kahneman calls System 1. Gigerenzer cites Warren Buffet’s famous mantra of never investing in companies he doesn’t understand as an example of a simple, but enormously powerful and effective heuristic that has allowed Buffett to outperform other investors who can use a dizzying array of analytical tools.

At a time when algorithms threaten to swallow the world, it can be tempting to hand over full control of our decision-making to machines, but many of the use cases for AI-based systems today have been extremely narrow. For example, AlphaGo beat the world at Go, but would have been surprised if it had been asked to play chess or Monopoly. Moreover, it would have struggled to provide the kinds of explanations behind its actions that are likely to be required as we regulate such systems in the coming years.

The noise that Kahneman and his colleagues Olivier Sibony and Cass Sunstein refer to in their book, Soundcould often be avoided if people better used checklists and heuristics to provide the consistency in their decision-making that is often lacking.

Recently research from the University of Notre Dame supports this view. The researchers studied customer service teams and found that our tendency to stick to tried and trusted practices can be hugely valuable in complex and difficult circumstances because they increase both our efficiency and effectiveness.

The key to managing the fusion of man and machine will be to better understand when tasks have relatively tight boundaries and plenty of data, and therefore may be well suited to automation, and when things are much messier and thrive more on the often undervalued the heuristics underlying this. so much of human behavior.

Adi Gaskell is the author of The 8-step guide to building a social workplace.

Image credit: Richard Drury via Getty Images.

Back To Top