The science of teams

How Gigster designed a system to build the perfect team on demand

Building teams that work cohesively is arguably the hardest part of getting any job done. According to psychologist Bruce Tuckman’s theory on team development, an effective team can take months, even years, to go through the four essential stages of team bonding: form, storm, norm, perform.

Yet a new company called Gigster — a service employing a wide network of freelance developers, designers, and product managers — uses their own custom-built algorithm (called “Karma”) to bring these complete strangers together within minutes to launch complex projects.

Can an algorithm trump a human in designing the perfect team? With $1.1 billion in revenue earned from projects successfully delivered by these spontaneous, on-demand teams, they’re certainly making a case.

“This system makes it so people can form loyal bonds with each other so that, when they are on a team together, they don’t bail,” says Olaosebikan, “They feel like, ‘Oh, I know this person. This is my friend. I need to actually help this person out’.”

Here’s a quick breakdown of how Gigster (and, by extension, Karma) works: Say you have an idea for a new app but need help turning your napkin sketch into a working prototype. Enter your idea on the Gigster website and in no time, you have a full-fledged project proposal and quote generated by their custom pricing engine.

Once a customer signs on, Karma identifies the right freelancers (or Gigsters, as they like to call themselves) — out of the network of over one thousand remote freelancers — to do the job. The question remains: How does an algorithm suss out the “right” people to bring together for a project? Herein lies the magic (or the science, really).

Karma ranks each freelancer in the community based on three main criteria: performance (how many projects a Gigster has successfully completed), logistics (whether Gigsters are in the same time zone, which could affect speed of project delivery) and relationships (whether team members have worked together on past projects).

After scouring the network looking for people with the appropriate skill level, based on the size and the scope of the project, Karma then looks for people with similar rankings, since these rankings are an indicator that certain people might work better together over others.

Gigsters can also earn extra points, and up their ranking, by giving back to the community. For example, if a developer shares a block of useful code that could be applied to other projects, speeding up other developer’s processes and delivery times, then their Karma rank goes up, as do their prospects.

“The more you perform and the more you contribute to the community, the more you’re able to do,” says CTO and Founder Debo Olaosebikan, “That opportunity creates a huge incentive for Gigsters, and it helps us make sure we’re assigning work to people who are actually excited to do that work. It’s our way of making sure that the best people are on the job, which leads to much more favorable outcomes for our customers.”

Olaosebikan says Karma is modeled on star-rating systems used by Uber and Airbnb and community rating systems on HackerNews and Reddit. But to make Karma work, he and co-founder Roger Dickey knew they’d need to create a strong sense of community amongst Gigsters from the very beginning.

“It’s pretty clear that you have strength in numbers. If you have a system where everyone works for themselves, and they don’t work together and provide value to each other, then you have a much weaker system. When everyone is connected, they can ask questions, they can share code and resources so future projects can be delivered even faster.”

Despite their success so far, Olaosebikan cautions that any points or rating system like this is far from foolproof. You have to get the mechanics just right, and be careful not to value a certain behavior disproportionately over another or it throws off ratings and people start to distrust the system.

“It’s an interesting challenge,” he says, “we’re constantly adjusting the algorithm to make sure it’s fair and just, and that it’s motivating the right kinds of behavior.”

At first blush, it sounds like Moneyball for HR. Then again, who knows? Maybe one day we’ll all be part of a system that rates us based on our work history and performance, though it’s comforting to know there is at least one out there that equally recognizes the value of being good to each other.