CAMBRIDGE, Mass. -- For multibillion-dollar industry giants like Airbnb and Univision Communications, best practices...
in BI help drive growth.
BI at Airbnb
Data streams in constantly at Airbnb, and top executives want insights daily, said Theresa Johnson, a data scientist and products manager for financial infrastructure at the San Francisco-based company, during a panel at the Real Business Intelligence 2018 conference.
To handle demands, Johnson helped develop an easier system for creating and linking metrics, and she fosters a workplace environment that has her data science team focusing on collaboration.
Airbnb largely uses its own set of tools for sorting data, Johnson said. But she said one of the best practices in BI she used was to create a scalable system, with linked metrics about the online market for the private vacation housing rentals that could be looked at and edited on the go with "slice-and-dice" tactics.
The system has its faults, as new metrics that harbor errors could inadvertently shut down large chunks of the system when it was introduced. Recognizing that, Johnson said she established a developer environment in which new metrics can be tested safely before linking with the system.
Johnson also developed scalable and repeatable tools to easily link new metrics without having to do the bulk of the work manually, which she said saves time and energy.
"One thing I learned from fellow designers is that you want to preference familiarity," Johnson said. "You don't want people to come to your tool and have to struggle [with] it."
For these types of tools, "the more work that it is to use, the less likely I am to use it," she said.
Data at Airbnb is also transparent, according to Johnson.
A version of final reports on the data is accessible by anyone in the company, and it's used weekly by certain key stakeholders.
Airbnb has about 4,000 employees, and most of them look at company reports, Johnson said.
Daily active users are far fewer, "but everyone in the company cares; everyone in the company should be informed," she said, highlighting one of the best practices in BI.
Using an open and transparent platform inspires care and accountability, she noted. So, "don't be secretive."
Making data accessible
Univision Communications has a similar philosophy, noted Simone Knight, vice president of marketing and media intelligence at the New York-based media giant, which targets Hispanic Americans, in her panel talk.
Knight also cited the need for an easy-to-use and open system.
Listing some of the best practices in BI, she suggested setting up a dashboard on an "always-on display," preferably on a large touchscreen. With a setup like that, she said, people will be more willing to look at the data and will be able to more easily dig into it.
In meetings, she said, "stop PowerPoints." Use a live dashboard instead, which will be more interactive and feature the freshest information.
Storytelling as a best practice in BI
Noting the importance of telling a story with the data, especially in the marketing world, Knight said it's best to try to weave stories into new products or tools when they launch, to better hook a customer and establish a user base. Storytelling can be important, too, when trying to explain data insights to management.
Knight said she often talks her 7-year-old daughter through the data first before presenting something during a meeting.
"If I can find a way to simplify it with her, I can find a way to simplify it with anyone," she said.
She also noted the growing relevance of automated BI systems, which can save employees valuable time and cut costs.
Theresa Johnsondata scientist and products manager for financial infrastructure at Airbnb
In addition to BI, Univision has begun using predictive analytics for determining the potential ratings of upcoming television shows. Using an algorithm, the company can predict ratings on premieres, taking into account the time of year, other shows currently being run at that time and how well past premieres have done.
For its part, Airbnb has its own "fully automated" framework in place for sorting through and digging into data, Johnson said in her own take on best practices in BI.
"We've set up a framework to make sure that we're ingesting the right data, that we're combining the data in ways that are meaningful, and that we're surfacing regularly and reliably," Johnson said.
It took a "bit of infrastructure" to make it automated, she said, but now saves on time, cost and energy.