How an algorithm can predict the success of start-ups based on the relationships between people in the team. After researching successful online dating matches and looking for patterns in personality and behaviour, Alistair created software that helps companies perform better, understand relationships, and create better teamwork through digital coaching. Want to make better hiring decisions or pull together the perfect team for your next project? Saberr takes away the guesswork.
WHAT'S THE BACKSTORY?
Alistair was at the Big Omaha conference in Nebraska when he met a woman called Sarah Prevette who shared her journey of launching a start-up. She presented a slide called the Start-up Lifecycle describing the descent from launch hype through to the first time you realise that you don’t have a clue, then the progression through ‘the humps of iteration’, ‘the plateau of disenchantment’ and ‘the sloth of ascent’. It’s a humorous illustration of what many entrepreneurs experience and charts Alistair’s own journey of starting Saberr.
Back in 2010, Alistair was studying under Noam Wasserman at Harvard Business School. Wasserman was writing a book called The Founder’s Dilemmas: Anticipating and Avoiding the Pitfalls That Can Sink a Start-up. He found that out of 1,500 US venture-backed firms, 80% fail. That’s not surprising in itself, but what’s interesting is that 65% of these companies fail because of team dynamics. One of the few areas that entrepreneurs have control over is who they hire. If this is the variable that is leading to failure then is it possible to quantify team dynamics? This was Alistair’s hypothesis.
Most of his early research focused on understanding individuals to understand the team. This wasn’t about identifying introverts and extroverts (the reality is that everyone gets on with a mix), but understanding the relationships between people. He decided to look at online dating as a data source. These were the days before app dating, so there were these information-rich profiles about what people were looking for in a partner. He then looked at patterns of successful matches – people who closed their account because they met someone.
One set of data taken from OkCupid revealed that men were most interested in profiles where there was a greater discrepancy in rating (from 1 to 5). In other words, if some find a person very attractive and others find the same person very unattractive, then more people were interested in that person. They also found that the less serious you are about religion, the most liked you are even by religious people. And that frequent tweeters have shorter real-life relationships than everyone else.
Alistair partnered with his friend Sam and started Saberr to put these findings into practice. Soon after, he was bragging about their ability to predict team performance to a professor at Bristol University, who the professor called his bluff and invited him to predict who would win at an event where eight teams of eight people were due to compete by pitching their business plans. Alistair went into panic mode and returned home, frantically writing the code to build their first algorithm. When the event came around, he sent out a list of questions to the teams and hoped for the best. Without knowing their skills, demographics or what they were working on, he successfully placed each of the teams in order of ranking. Saberr repeated this soon after at The Microsoft Imagine Cup, making correct predictions and beating odds of 1 in 3.6 million.
WHAT'S HE DOING ABOUT IT?
After he graduated, Alistair moved to London with about £500 left in his overdraft to get the business off the ground. He needed to meet the right people and find clients, but companies didn’t know how to pay for the service because the concept was new. This was when he reached rock bottom on Sarah Prevette’s Start-up Lifecycle. Just as funds were running low, he met Seedcamp while working in the basement at Google Campus. They agreed to invest in Saberr, allowing them to continue their experiments. But Alistair turned the microscope on Seedcamp, offering to predict how their investments would fare based on the teams within each start-up. Then he ranked a list of potential candidate companies from 1-21 and placed eight of them correctly. The people at Seedcamp were stunned.
One of the companies was failing because one person had a difficult relationship with the rest of the team. Seedcamp asked Alistair what they should do, and he told them to observe without taking any action. They waited to see how this would play out and sure enough, a week later, the company folded because of the team dynamic. That team member's name was Nik Brbora. After the company collapsed, Alistair contacted Nik and explained that they knew it would happen but that his profile was an exceptional match with the team at Saberr. He offered Nik a position at the company. After his initial anger subsided, he joined the team as CTO and turned out to be one of the best hires they ever made.
They now had a product, but this also led to a problem. Companies began asking them to identify low performers and rank employees in order of ‘best’ to ‘worst’. From a technical point of view, they could achieve this with accuracy. But they didn’t want to become an axeman, so they changed the focus to the ‘culture fit’ of the people within that team. Suddenly their aim was to improve performance within teams rather than simply predicting it. This was also a better approach for most companies, who could then incentivise employees by investing in their development.
A useful analogy to explain how this works is to imagine a flock of starlings. There are tens of thousands of birds flying in the same direction. How does this work without it looking like chaos? If you apply some simple rules, you can bring them together. You could begin by making sure each bird travels at the same speed. Then specify that each bird must travel close to its neighbour. Then, that each bird must avoid danger. In this model, it works best when each bird stays close to its nearest seven neighbours, so a flock of birds is actually a collection of sub-teams working together. It’s the same principle for humans. How many people have shared the Nobel Prize? In most categories except for literature, this usually requires people to work as a team. The biggest problems most companies face need effective teams to solve them.
Saber has now launched a second product for team development called CoachBot. It’s a digital coach, communicating with each individual in the team and bringing focus to areas such as goal setting, defining roles and improving teamwork. CoachBot is currently used by several large companies including the NHS. It's been shown in NHS acute care units that team effectiveness directly impacts patient mortality rates, which just goes to show the hidden impact of teamwork and why every company should be investing in digital coaching.