Sonian is working on ambitious data science solutions in the realm of extracting business value from unstructured, employee-generated content. It’s estimated that more than 80% of an enterprise’s valuable data assets live in email, files, documents and discussions. Sonian’s cloud platform reveals organizational insights in the same way a logging system like Splunk reveals operational insights. With analytics, we’re helping enterprise IT provide more value to their organizations. While down at the Chief Digital Officer Summit in New York this week, it was great to hear validation of many of the analytics best practices we’ve been following ourselves on our own data sciences journey.
Here are 4 keys to analytics success:
The data must tell a story that can be understood by all stakeholders
A big trend is publishing dashboards displaying KPI information. These are well intended efforts, with green, yellow and red beacons providing 30,000 foot situational awareness. But dashboards introduce the “whack-a-mole” problem and what is missing is providing context associated with the status. If an indicator is red, the reason why must be available. Data science needs to tell a story about the data with context, offering both “the why” along with “the what.”
A data science team needs buy-in from company leadership
Teams embarking on their data science journey need to be empowered by company leadership and not treated as a half-hearted measure. The motto should be “all in or not at all.” Avoid the trap of starting a data science initiative before the CEO and CxO team are ready to commit their respective departments to the cause. Data science projects must reach across company functional silos to provide the comprehensive context and tell the complete story (see #1 above.)
Treat your data science project like a startup
A data science journey begins with top-level management approval and internal funding. New technologies will need to be purchased and most likely new team members hired. Use the “Lean Startup Playbook” to guide the project. This means work backward from the customer (in this case the management team and department leaders), have a big vision but focus on quick deliverables and short iterations (something new must be given to customers at least every 2 weeks) and be prepared to pivot to please the customer.
Don’t create in a vacuum
And finally… don’t create in isolation. A data science project will only be successful when business users are active co-creators in the project. Many technologists have blind spots toward the “business side of the house.” Connect business users to data scientists in one team so there is cohesion and shared sense of responsibility for success.
— Greg Arnette is the founder and CTO of Sonian. Follow Greg on Twitter for more information about email, collaboration, big data management, cloud computing, start-ups and more.
Sonian preserves, protects and presents the world’s information.
More than 25,000 customers in 43 countries trust Sonian’s secure proprietary platform to retain, retrieve and surface critical data and protect intellectual property.
Founded in 2007, Sonian is the only pure public cloud information archiving company, providing services that are easy, flexible, actionable and reliable for OEM partners and their end customers. Sonian allows companies to preserve, analyze and access their electronic communications for legal, regulatory and continuity purposes while gaining organizational insights.
Sonian is building the future and solving big data problems for companies, all while managing more than 20 billion objects in the cloud.
Have questions? Want more info? Need help?