To answer these questions, we recently connected with LinkedIn’s DJ Patil. As the company’s Chief Scientist, a position that he created in 2008, Patil has been instrumental in transforming massive amounts of data into creative products like “People You May Know” and LinkedIn’s new “InMaps,” which allows you to create a visual graph of your entire professional network. (Patil’s own map is pictured above.)
Here’s 6 best practices for harnessing data and making it actionable, illustrated with excerpts from our interview with Patil:
1. Iterate quickly.
“We build all of our systems with the intention of being able to iterate very fast, so we’re able to modify and change them very quickly. Having that ability to iterate quickly is essential, because we never really know which aspect of a product is going to be the winner.”
“With our ‘Groups You Might Like’ module, the original product was built in less than a week. And we iterated on it very, very quickly; we tried a number of things to see what really worked. Then we were able to build out much bigger technology once we knew what was really winning – the aspects that people really liked – and translate it into a much more powerful product.”
2. Test, test, test.
“Sometimes we test internally, and let our employees try things out. We also do user testing, where we invite people into the company to test things. Once people have tried a new product out and we’ve taken their feedback, then we might roll it out in different versions to different people to see what they like, based on their usage. Then, once we find what’s good in there, we tend to roll it out on even bigger and bigger levels. Maybe we roll it out to 1% or 5% or 10% of our user base.”
“It’s essential that we’re always testing the product in different ways. And that’s not just with my team, that’s a cultural paradigm for all of LinkedIn. We really pride ourselves on a relentless pursuit of making sure we’re providing increasing value for the user.”
3. Use cheap solutions.
“There’s so much great technology that’s open source. You don’t need to pay 10 million dollars for these very sophisticated data systems anymore. You can start and try things on the cheap until you know that you’ve got a winner.”
4. Track your numbers very, very closely.
“You’ve got to be absolutely rigorous in how you track and how you monitor and are able to actually take that data and understand what the value is.”
“I think we’re seeing a unique time where there’s a lot of technology that we can apply to the data problem. At LinkedIn, we’re really focused on high-quality data. You can do tremendous things to make the product more useful when you have high-quality data.”
5. Empower your employees.
“We have a notion called: Leadership, Leverage, and Results. Leadership is the ability to inspire people and get them to drive towards a common, shared vision. Leverage is being able to really do something with a modicum of resources – so we don’t need to hire an outside company to work on new ideas that we have. Results is getting it done, moving the needle.”
“Then we empower our employees to do these things, believing that they are the best-suited people to drive the mission of the company forward. Everything from our lab sites, to internal rollouts, to product testing – there are lots and lots of ways to make those things real.”
6. Emphasize creative experiments.
“One of the things we have once a month is InDay. It’s on a Friday and there are no meetings allowed; it’s a day where you get to do whatever you want. We bring speakers in, people arrange classes, but you can also just decide that you want to work on a project and you plan it out. And you may get 3 or 4 people that say, ‘That’s a really cool idea; I want to help you with that.’” “So you get a bunch of guys, and you take a chance on it. And maybe it takes a couple of InDays to get it over the line. That’s how a lot of our innovation happens. A lot of great stuff is created by having time and flexibility to foster the creativity, come together to find solutions, and to make them real.”