16.01.2016 Views

Tips for Building a Data Science Capability

WH4vS

WH4vS

SHOW MORE
SHOW LESS

You also want an ePaper? Increase the reach of your titles

YUMPU automatically turns print PDFs into web optimized ePapers that Google loves.

THE DEPLOYED MODEL<br />

As with the diffused model, data science teams are<br />

embedded in the business units. The difference is<br />

that the embedded teams in the deployed model<br />

report to a single chief data scientist as opposed to<br />

business unit leaders. In this model, also called the<br />

matrixed approach, teams are generally assigned to<br />

individual business units, though they are sometimes<br />

also assigned to broader product lines, or to mission<br />

sets comprised of members from several business<br />

units.<br />

This model often works best in organizations with<br />

medium-sized data science capabilities— ones<br />

that have a sufficient number of teams to handle<br />

multiple projects, but must still carefully target their<br />

resources. This model has many of the advantages<br />

of both the centralized and the deployed models;<br />

the data science capability is more of an organic<br />

whole, yet the embedded teams are close to the<br />

business units.<br />

CHIEF DATA<br />

SCIENTIST<br />

DATA SCIENCE<br />

TEAMS<br />

BUSINESS UNIT<br />

LEADS<br />

<strong>Data</strong> science teams are overseen by a chief data<br />

scientist and <strong>for</strong>ward deploy to business units.<br />

Because the deployed model is often seen as the<br />

best of both worlds, organizations may be quick to<br />

adopt this approach. But it is also the model with the<br />

THE DEPLOYED MODEL<br />

ADVANTAGES CHALLENGES PLACES EXTRA FOCUS ON…<br />

+ Shared benefits of both the<br />

centralized and diffused model<br />

+ <strong>Data</strong> science teams collectively<br />

develop knowledge across<br />

business units, with central<br />

leadership as a bridging<br />

mechanism <strong>for</strong> addressing<br />

organization-wide issues<br />

+ Access to data science is<br />

organization-wide, and close<br />

integration with business units<br />

promotes analytics adoption<br />

+ Project diversity both motivates<br />

data science teams and<br />

improves recruiting and<br />

retention<br />

+ Central leadership streamlines<br />

career management approaches,<br />

tool selection, and business<br />

processes/approaches<br />

+ Deployed teams are responsible to<br />

two bosses—staff may become<br />

uncertain about to whom they are<br />

ultimately accountable<br />

+ <strong>Data</strong> science teams may face<br />

difficulty being accepted into<br />

business units, where long-time<br />

relationships have been established<br />

+ Access to analytics-resources<br />

may still feel competitive between<br />

business units, and data science<br />

units risk alienating business<br />

units whose proposed projects<br />

are not selected<br />

+ Conflict Management. The chief<br />

data scientist should proactively<br />

engage business unit leaders to<br />

prevent competing priorities from<br />

becoming the data science teams’<br />

responsibility to resolve<br />

+ Formal Per<strong>for</strong>mance Feedback.<br />

Agree to per<strong>for</strong>mance goals at the<br />

onset of each project, and collect<br />

feedback during the life of project,<br />

including at its conclusion<br />

+ Rotation. Allow data science teams<br />

to work on projects across different<br />

business units, rather than within a<br />

single business unit—take<br />

advantage of one of the main<br />

benefits this model af<strong>for</strong>ds<br />

+ Pipeline. Regularly communicate the<br />

data science project pipeline,<br />

allowing business units to see how<br />

their priorities are positioned<br />

Aligning <strong>Data</strong> <strong>Science</strong> | 25

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!