1 Comment

Summary:

Uber, LinkedIn and Airbnb take different approaches to integrating data into their businesses, but they all agree that data is a crucial signal of whether their products and services are serving customers effectively or not.

Riley Newman, Airbnb; Henry Lin, Uber; Yael Garten, LinkedIn - Structure data 2014

The fact that data is important to the running of most businesses, particularly technological ones, is now taken for granted, but how do you integrate the collection and understanding of that data into your company so that it makes a difference? Three fast-growing web companies — Uber, Airbnb and LinkedIn — talked about the different ways they do this at Gigaom’s Structure Data conference in New York on Thursday.

Riley Newman, the head of the data science group at Airbnb (which is rumored to be raising a new round of financing that could value the company at $10 billion, according to the Wall Street Journal) said that from his perspective “data is the lifeblood of our business — we think of it as the customer’s voice. It’s them telling us what works and what doesn’t work, so we always start with the data.”

Newman said that Airbnb approaches data and its use as if it’s similar to a software stack, with data engineers at the base who build the tools, then a layer above that which creates the pipelines to get data into a format that the rest of the company can use, and then data scientists on the upper layer do the “heavy lifting” and deep dives into specific data.

Yael Garten, manager of mobile data science at LinkedIn, said that everything the company does is informed in some way or driven by data. “If you don’t measure it, you can’t fix it,” she said. And so a data scientist is embedded with each of the product teams so that they can track and analyze what is happening with customer behavior in real time and make changes on the fly. Newman said that Airbnb also has data scientists embedded with product teams.

Uber, however, takes a different approach, according to senior data scientist Henry Lin: the car-service company keeps its data team completely separate from the product management side, he said, so that they can pursue whatever research projects or experiments they think might help make the service more efficient. And the data science team reports directly to CEO Travis Kalanick, he said.

 

Structure Data Ticker

Photo by Jakub Mosur

  1. Diego Schmunis Thursday, March 20, 2014

    The problem with all this data-driven decision making process is that 1) this are all lagging indicators, and 2) that without understanding the WHYs (why are the users taking the action that they are taking) it’s hard to put the data in the correct context to fully understand its meaning and derive ideas and solutions

    Share

Comments have been disabled for this post