The new science of management?

“Social Physics” can lead to incremental improvements, but risks marginalising strategy

Thinkers influenced by the Enlightenment have long dreamt of applying the rigour, precision, and predictive power of the physical sciences to the social world. Seduced by the elegant laws of Newtonian physics, David Hume, Adam Smith and Auguste Comte sought to uncover rules governing interactions in politics, society and commerce.

With big data, thinks MIT’s Sandy Pentland, these dreams may finally be realised. In his new book Social Physics, he argues that understanding the way information circulates in organisations and society (now possible because of the digital revolution) means we can re-engineer these systems for the better. In one experiment, his team used “sociometric badges” to monitor interactions between workers at the call center of a large bank in the United States. Just by changing the structure of coffee breaks so that the “flow of ideas and conversation was different”, the researchers boosted productivity substantially, saving the firm millions. So are businesses missing out by not diving into the pool of big data?

Pentland’s basic contention is that networks matter. By tracking the “digital breadcrumbs” we all leave behind (through social media, credit cards, or smartphones), business and political leaders can get an idea of how to optimise the flow of information through organisations. And the results of his research are impressive.

With a couple of assistants, Pentland analysed 10m transactions by currency traders on a social trading network. They looked at the flow of information, and found that individuals fell on a continuum from “isolated” to “hyperconnected”. Those in the middle (the “decision-making sweet spot”) had a return on investment 30 per cent higher than the rest, who were either too isolated and lacked a diversity of ideas, or were stuck in a noisy “echo chamber” of information and ended up just following the herd. In another example, Pentland helped reorganise the seating plan in the drug discovery unit of a pharmaceutical company, which by itself seemingly boosted performance significantly.

Pentland sometimes talks about such incremental improvements as if they’re free lunches – efficiency gains that businesses are foolish not make the most of. But when the choice is contextualised alongside the other spending decisions firms face, it’s less obvious that big data should take priority. Is better information circulation a more pressing need than a new IT system? Is it worth buying all employees wearable devices to track these interactions?

And there’s a worrying conceptual trend in Pentland’s ideas. In 2008, then editor of Wired magazine, Chris Anderson, argued that a “data deluge” would see correlation replace causation. Scientists and businesses would no longer need to test theories in experiments, going back to revise them in light of the results. Instead, they would just pour through reams of figures to spot relationships between variables. It’s an inherently statistical way of thinking, and Anderson argued that this could mean the death of theory.

But theory (call it “vision”) can be necessary in management. Pentland’s use of big data will no doubt lead to incremental improvements in business processes. But a myopic focus on tweaking the flow of information may lead some firms to become backward-looking. As the writer Nicholas Carr has argued, acting to streamline informational flows alone “will encourage us to optimise the status quo, rather than to challenge it.” To really break the mould, good old fashioned business vision is still necessary.

Track your own work habits
WhatPulse Free

WhatPulse, available as an app for desktops, tablets and smartphones, extends the quantified self even further. The app measures keyboard and mouse activity so you can monitor the time spent doing different tasks on a device. Information is uploaded onto a central website, where you can share with friends.