From the course: DJ Patil: Ask Me Anything

When does data leave holes?

From the course: DJ Patil: Ask Me Anything

Start my 1-month free trial

When does data leave holes?

“

(upbeat music) - The early days of LinkedIn, there was this question of which country to go in first? Should we go into France, where there's a strong competitor? Should we go into Australia, where we see good, decent numbers growing? And we're not sure, where are we going to get traction. Resources are incredibly scarce. So, we started looking at this and we're like, "Wow, we don't know." This was a metrics meeting, where you're kind of looking at it and we're talking about the numbers, white-collar jobs, blue-collar jobs, we're like, "Maybe we don't have enough data." So we go and we try to start doing some surveys. We also start looking at the census data from those countries, we start bringing that together. We pull all that together and we start looking at it, and it's like, "Well, you know what, the indication really looks like we should go into Australia." Like, that's what the bet should be. So now there's the metrics meeting, which kind of made this assessment. Now you separate this. Now we have the decisional meeting. Decisional meetings as, we should absolute go into France, but the data said go into Australia. Why, why, why the difference? Well, it turns out, there's lots of things you can't necessarily measure. So in this case what we're doing is, we said, "Look, if we don't get ahead of this competitor, we may never have another shot." So, okay, so our hypothesis is that we have to go in there, otherwise we're going to lose potential market, and we're spend all these calories. So, it's not that we now just say go into the market. How do we make this testable? How do we make this a process in which we can iterate very quickly to assess that? And if we're wrong, we pull back and go the other direction. Now, none of that sophistication in the answer, in the techniques and tactics that we took, could've been figured out without the metrics meeting lifting up everyone simultaneously. And if we just had a meeting of, should we go into France or Australia, we would've just been doing this the whole time rather than actually having this very nuanced, with the sophisticated backing of material that helped us actually tease it out. And, for the record, I was one of the people who actually said, "We should go into France," despite the data, despite being in charge of, responsible for the data, because you just, you have to sometimes go with your gut. One of the most important things that we forget when working with data is that first we're human. That means, over time, we have processed a ridiculous amount of data and have this wealth of information to drawn on, primarily to create intuition. So it's not that our intuition might be wrong. It's about how do we actually leverage our intuition to develop a testable theory and that can be iterated on very quickly to get to more optimal answers. It's what, it's about agility, at the end of the day. (upbeat music)

Contents