Google’s former HR Director and co-founder of a startup called Humu, Laszlo Bock, recently unveiled his stealthy startup’s first product. It has been a little over a year in the making. Humu was founded by Bock and his co-founder Wayne Crosby in May 2017, and soon afterwards raised a $40 million Series A from Index Ventures, which is a massive Series A even by today’s frothy standards. On Monday, we finally learned what they’re launching with all that cash: a Nudge Engine!
Without describing the product further, it’s obvious that Bock and his team are taking a page from the strand of behavioral economics popularized by economists Richard Thaler and Cass Sunstein in their book of the same name, “Nudge: Improving Decisions about Health, Wealth, and Happiness”. Sunstein characterizes a “nudge” as a “liberty-preserving approaches that steer people in particular directions, but that also allow them to go their own way.” The most famous example of a “nudge”, and supposedly Thaler’s favorite, are the urinals at the Amsterdam airport, which contain small pictures of flies in the correct place to “aim”, thereby dramatically reducing “spillage”. (Would you believe Thaler is a Nobel laureate? Believe it.)
“Nudging” is not without its critics, most vocally Nassim Taleb of “Black Swan” fame, who worries that “nudging” may take place without a complete understanding of why the underlying behavior exists in the first place. The result could be to nudge people in one direction, only to find that at the aggregate level, their behavior has changed the overall system in a negative way. (Non sequitur: hilariously, Taleb has taken to calling Thaler “Nudge Boy” on Twitter. For his part, Thaler insists he’s proud to be a person Taleb calls an “idiot”. Big egos at work!)
Regardless of whether “nudging” is science, or pseudoscience as Taleb claims, the fact remains that there are many examples where organizations have been able to push people in one direction or another with clever design.
Product…Good! Brand Name….¯\_(ツ)_/¯
Which brings us back to Humu, and their…interesting…new product name, “Nudge Engine”. What the heck is a “nudge engine” you might ask? I’ll tell you…
Every company’s HR function is ideally trying to improve employee productivity and retention. Happy, productive employees drive company revenue growth, by some estimates as much as twice as much as companies with average levels of employee satisfaction. Increased employee retention reduces training and staffing costs, which for highly trained employees is over 20% of their annual salary. Solving the problems of lacklustre revenue growth and high employee turnover is what Humu is all about.
To start with, Humu provides a real-time dashboard for a variety of HR performance indicators, and they condense these performance indicators into a “Happiness Index” for each team and organization (shown below):
The “Happiness Index” is a composite of other, proprietary variables, such as whether employees would recommend their company as a “great place to work”, or strong agreement with the statement that “I often experience positive emotions when I’m working”. Shown below is an example of Humu’s company wide measurement of one such variable, “team comfort”, over time:
What makes Humu’s software more than a fancy dashboard is the use of a “Nudge Engine”, also known as a Decision Machine. The decision machine is software that uses machine learning to decide when and how to provide a message to employees and managers to improve a company’s Happiness Index over time. These decisions are the “nudges” in the “Nudge Engine”. Each “nudge” hopefully leads to action on the part of employees and managers, which move the entire company closer to is HR goals. A example of a daily “nudge” to a manager and employee is shown below:
Like most decision machines, Humu’s software will A/B test different messages and “nudge” individual employees with different recommendations. It will then use the power of statistics to determine whether or not these “nudges” meaningfully impacted the survey responses provided by the employee over time. Some “nudges” will have no affect on employee happiness, while others will. Over time, Humu’s software hopes to reliably improve and maintain a company’s Happiness Index.
Where HR Meets Finance
Humu is likely hoping to replicate the success of one of the first startups inspired by Thaler’s work, called OPower, which was acquired by Oracle for $500M in 2016. OPower used the “nudge” concept to induce utility customers to lower their energy consumption.
OPower’s simple paper reports (shown above) have saved utilities in the U.S. over $2 billion. They work by “nudging” homeowners with above average energy use relative to their neighbors to conserve energy.
Humu has an even more difficult challenge than Opower: they must prove that they can not only increase employee happiness, but that doing so leads to a measurable increase in worker productivity, or a measurable decreases in employee turnover. Their healthy Series A could indicate that their customer data already shows such a linkage, and that the product effectively works out of the box. Alternatively, it could indicate that senior Xooglers can simply raise ridiculous amounts of venture capital, especially when their business plan includes the phrase “machine learning”.
If Humu’s software performs as advertised, expect every sales team on the planet to buy it, followed by every major retailer struggling with chronic turnover of low-skilled workers. In both cases, productivity and turnover data are likely tracked at the CFO level, and improvements in either will be rewarded by a hallowed line-item on the company’s software budget. Better yet, Humu will be rewarded by a lucrative acquisition or IPO.
On the other hand, if Bock and his team have merely raised gobs of money on Google’s good name, this may be a really lovely piece of software design with no staying power. The ultimate challenge for Humu is that humans are extremely complicated, and even behavioural economists with Nobel Prizes attached to their name have just scratched the surface. Is it therefore possible to affect someone’s underlying motivation and performance merely by reminding them to make changes to their behavior; changes that may conflict with years of habit and deeply ingrained personality traits?
It’s a fascinating experiment and I’ll be watching Humu’s progress closely. It would be incredible if they succeeded, but also, somewhat depressing to see that human behavior can be so easily hacked…excuse me…”nudged”…by a computer.