Organizations need to improve corporate adoption of analytics.
A June 2017, research study conducted by Forbes Insights and Dun & Bradstreet, revealed that 59% of more than 300 companies surveyed did not use predictive modeling or advanced analytics, and 23% still used spreadsheets for most of their data analytics work. Even more startling was the fact that 19% of respondents used no analytical tools more complicated than basic data models and regressions.
This news is not encouraging for big data and analytics champions and managers--nor is it good for their lagging companies, as evidenced by an MIT Sloan Management Review that found that 67% of companies that were aggressively using analytics achieved competitive advantage in their markets.
SEE: Quick glossary: Business intelligence and analytics (Tech Pro Research)
Sacrificing competitive advantage is reason enough for CIOs and CDOs to place analytics adoption by the company near the top of their priority lists.
What is slowing down meaningful big data and analytics adoption, and what can CIOs and CDOs do about it? See below five common problem scenarios and ways to overcome them.
Problem 1: The business is not seeing enough tangible impacts from analytics
There are still too many organizations running analytics as a series of "pilot projects" in test tube mode. While testing small pilots was a good initial concept for introducing analytics in companies, too much time has passed to continue this approach. Test tube lab projects suggest that analytics are not ready for prime time in businesses. It is impeding meaningful corporate analytics adoption because individuals at the C-level don't take these "lab" projects seriously.
Solution: CIOs and CDOs need to move analytics projects out of test mode and into active and meaningful production. They can accomplish this by collaborating with business managers who bring business cases for analytics. Together, they can insert analytics into active production workflows and measure results that either improve revenue or decrease operating costs. If analytics projects don't contribute to either of these bottom-line-influencing objectives, CIOs and CDOs could discontinue them.
Problem 2: Analytics are difficult to use
A supply chain manager who is used to conducting business in-person with handshake deals with suppliers won't willingly move to an analytics evaluation of suppliers. But if he or she sees enough late deliveries or quality issues for a favored supplier that impacts company revenue opportunities and operational costs, there may be a way to demonstrate how new analytics reports can help and not hinder the ways of business that have always worked.
Solution: Don't try to reinvent everyone's business processes by forcing analytics on them. Instead, find opportunities where analytics can contribute to what users already do. Then work continuously with users to revise processes and analytics reports for the best fit. The key is creating a collaborative process--not throwing handfuls of analytics reports over the fence and hoping that users catch them.
SEE: Quick glossary: Project management (Tech Pro Research)
Problem 3: IT and data science teams aren't working together
A majority of companies still keep the data science team in one departmental silo and IT in another. There are reasons for this. For a long time, the data science group was a "test lab," with few pressures to get projects into production. The time for pilot projects is over.
Solution: Organizations need CDOs and data science--but if data scientists and IT don't actively collaborate, analytics will have a hard time succeeding in production. We already know how important it is to bring data science unstructured data and IT structured data together in analytics. To facilitate this, CIOs and CDOs need to actively participate in projects together and eliminate silos.
Problem 4: IT infrastructure fails to address big data and analytics
As more transactional IT systems move to the cloud, IT is turning to cloud vendors for infrastructure management. This displaces some of the work that highly skilled system programmers and DBAs performed on in-house applications and data-processing. The risk is that highly trained personnel could leave the company.
Solution: As IT moves into new parallel processing and server clustering environments, big data and analytics workflows need to be managed. This provides the perfect opportunity to cross-train seasoned system veterans into managing and optimizing the workflows that run on these big data processing clusters.
SEE: Prescriptive analytics: A cheat sheet (TechRepublic)
Problem 5: Mission-critical systems need to be revisited
Corporate disaster recovery and business continuation plans continue to focus on transactional data systems.
Solution: CIOs need to meet with senior managers in business and IT to reevaluate, which systems are mission critical. As more analytics systems become integral to decision making and automated operations, these systems need to be included in updated DR plans. As analytics systems get classified as mission critical and moved into DR plans, CEOS and other C-level executives will take analytics and big data more seriously.
- 90% of companies are working on AI projects, but they're making one big mistake (TechRepublic)
- 85% of big data projects fail, but your developers can help yours succeed (TechRepublic)
- How to choose the right data analytics tools: 5 steps (TechRepublic)
- How to know if text-based analytics makes sense for your company (TechRepublic)
- How to select the right project management methodologies (ZDNet)
- Business analytics: The essentials of data-driven decision-making (ZDNet)