An Insider’s Guide to Building Data Science Teams

How to build a data science team: It isn’t easy to recruit and retain a crackerjack data science team, especially now that data scientists know they have the best job in America. So it’s an opportune time to read “Doing Data Science Right,” by Instacart’s VP of data science Jeremy Stanley and former LinkedIn data leader Daniel Tunkelang, in First Round Review.

Along with scads of practical advice for getting the most from your company’s investments in data science, Stanley and Tunkelang suggest that you answer four questions at the get-go:

• Are you committed to using data science to either inform strategic decisions or build data products? If not, you don’t need data scientists.
• Can you collect the data you need and act on it? Data science requires data, and collecting data isn’t enough — data must drive action.
• Will you have enough signal in your data to derive meaningful insights? Big data is great, but without a high signal-to-noise ratio, it will be tough to tease out insights.
• Do you need an internal data science capability? If data science is solving problems that are critical to your success, then you can’t afford to outsource it. Otherwise, you should.

On-boarding in the cloud: The paper-based on-boarding process for all new employees at Brooks Brothers used to eat up hours, reports senior U.S. correspondent Katherine Noyes in Computer World. Now, the company simply sends out a link via email to newly hired employees prior to their starting dates. It enables them to complete the paperwork at home and provides them with access to the resources of the company’s cloud-based employee portal. “Not only has the process required on that first day been reduced from hours to minutes, but new hires begin to get acclimated before they even start work,” Brooks Bros. director of talent management and organizational effectiveness Justin Watras tells Noyes.

Cloud computing promises outsized benefits in HR because many of the function’s back-office processes, such as benefits, time and attendance, and other records, are mired in decades-old, on-premises legacy systems. “The result can be an almost impossibly intricate set of software and processes,” explains Noyes. Sure, a shift to simplified, cloud-based HR processes can cut costs and save time. But it can also help push ownership of HR information and analytics down to frontline managers, and better serve the ever-growing numbers of digital natives in the workforce.

What big data can teach you about communication: The connection between a speaker and her audience seems so intimate and fragile that it’s hard to see how big data might play a role. But don’t tell that to Noah Zandan, founder and CEO of Quantified Communications.

Zander and his team of data scientists analyzed more than 100,000 presentations from corporate executives, politicians, and keynote speakers. “They examined behaviors ranging from word choices and vocal cues to facial expressions and gesture frequency. They then used this data to rate and rank important communication variables such as persuasiveness, confidence, warmth, and clarity,” explains Matt Abrahams, a communications coach and organizational behavior lecturer at Stanford b-school, in an article for Insights by Stanford Business. The result is an analytics platform and app for measuring, evaluating, and improving corporate executives’ communication skills.

Not impressed? Abrahams reports that the language used in corporate earnings calls affects stock price movement by up to 2.5%. As little as a 10% increase in vocal elements, such as volume, rate, and cadence, can have a significant impact on audience reception levels. And enhanced authenticity, as evidenced by passion and warmth, boosts an audience’s perception of a speaker’s trustworthiness and persuasiveness by 1.3 times.