Machine Learning Times
Machine Learning Times
Why You Must Twist Your Data Scientist’s Arm To Estimate AI’s Value
 Originally published in Forbes, June 11, 2024. If you’ve...
3 Ways Predictive AI Delivers More Value Than Generative AI
 Originally published in Forbes, March 4, 2024. Which kind...
AI Success Depends On How You Choose This One Number
 Originally published in Forbes, March 25, 2024. To do...
Elon Musk Predicts Artificial General Intelligence In 2 Years. Here’s Why That’s Hype
 Originally published in Forbes, April 10, 2024 When OpenAI’s...

Five tips for bridging the business-IT gap

By Anne Robinson, Director of Supply Chain Strategy & Analytics, Verizon Wireless

Originally published in The Knowledge Exchange


Anne Robinson is the Director of Supply Chain Strategy and Analytics for Verizon Wireless and the President-Elect of INFORMS, the Institute for Operations Research and the Management Sciences. Before joining Verizon, she worked for Cisco, most recently leading business intelligence and advanced analytics. Throughout her career she’s been dedicated to bridging the gap between IT and the business side. It can be a tricky prospect that calls for understanding the concerns of IT and the needs of the business. As it becomes more important to incorporate analytics throughout the enterprise, she offers some solid tips for how to make the IT/business relationship stronger.

1) You can’t do enterprise-level analytics without the help of IT

Business units can maintain independent analytic solutions for small, discrete projects, but true enterprise-level analytics requires coordination with IT. Optimization or forecasting models, for instance, can be built within a business unit for localized or one-off uses. However, if the expectation is that these models will be connected with enterprise data or become part of a regular business process, IT has to integrate the analytics into the existing systems that run the organization – otherwise the models won’t deliver their intended value. Engage IT early.

– otherwise the models won’t deliver their intended value. Engage IT early.

2) Consider the 70/30 Rule

Communication is critical no matter where you sit in an organization. To ensure effective interactions between my team and my IT counterpart’s team we implemented the 70/30 rule. For my analytic and operations units, I told them 70 percent of their function was business-focused, while 30 percent was IT related. And my IT counterpart told his team the opposite, that their roles comprised 70 percent IT and 30 percent business. Team performance was incented and measured against this mix. Now both the business and IT teams can translate. They understand each other’s worlds. They don’t have to throw things over a fence.

3) Look for options that make IT comfortable

If IT is a little hesitant about analytics it’s because they don’t want anything bringing down their systems, particularly business critical ones. Customized programs and solutions are often at odds with the mantra of keeping things vanilla. However, it is important to understand that “customized” and “configurable” are different – the latter allowing a vanilla implementation while keeping things flexible for analytical modeling purposes. Consider framing your analytics as “configurable” solutions to ensure IT system continuity.

In addition to understanding IT concerns, you can help alleviate their fears by implementing checks and bounds to stop issues like runaway calculations. One great option for business and IT is to incorporate in-database analytics. It runs within the database in a way that doesn’t disrupt the day to day functioning of the database.

4) Develop a rapid prototyping model approach

Analytics efforts can get bogged down, especially when you’ve got to get IT and the business working together. Start with a small part of the problem, develop a goal or agree on a small milestone – and run with it. Then bring in the stakeholders and domain experts and have them validate it. Then do it again. Along the way, people begin to feel comfortable about what you’re doing. And even if the analytic efforts come up with different answers from what they’re expecting, in the end they begin to accept the solution.

If you come in with the approach of having the big, scary, black box that will spit out all the answers – neither side will accept it.

5) You will never reach the automation state of analytics without strong IT buy-in

Being able to automate decision making is a big part of making enterprise analytics work. In a small company, it’s easy enough to have a solution that sits on the CEO’s desk so that he or she can see what decisions are being recommended. But if you want to automate decision making in larger companies, you will need an enterprise data warehouse and some enterprise business intelligence tools. Either you need to work with IT to help embed those into what’s currently there, or build a new infrastructure to support it. Either way, you won’t get it done without partnering with IT.

Originally published in The Knowledge Exchange

Comments are closed.