Sergey Nivens - Fotolia
Even the best advanced analytics tools are doomed to failure if decision-makers on the business side of an organization don't use them.
That was the lesson Rod Moyse, head of analytics at British insurance company Aviva plc, learned after five years of implementing reports and data products aimed at simplifying and sharpening business processes but that ultimately went underutilized. This meant his team was wasting its time creating analytics tools and the business was missing out on an opportunity to gain a competitive advantage through data.
"People didn't really understand what analytics was," Moyse said in a presentation at the SAS Analytics Experience conference in Las Vegas. "It was considered one of the darkest of the dark arts. We needed to change this, and fast."
Between 2010 and 2014, Moyse's team created tools that predicted which customers were at risk of leaving based on customer service complaints, scored claims on the likelihood that they were fraudulent and predicted appropriate settlement figures for bodily injury claims.
Analytics tools adopted when embraced by employees
Moyse said each of these models had high predictive reliability and had the potential to save the company money. When his team presented these tools to management there was excitement about each one, but ultimately frontline workers did not embrace them.
Rod Moysehead of analytics, Aviva plc
It wasn't until 2015 when his team built a tool that predicted which vehicle accident claims would result in the car being declared a total loss and how much claims handlers should offer customers to settle that adoption took off.
The key, Moyse said, was that the tool simplified a part of the claims handlers' job without automating the job away entirely. It solved a problem for decision-makers while valuing their role in the process.
"This was the best thing we've done because it gives our users superpowers," Moyse said.
Keys to adopting analytics tools on the frontlines
We've all heard the advice telling analytics teams to find executive sponsors and tackle projects that are relevant to lines of business. Those things can be useful, but they need to be done in the right way.
Moyse said you need a "business sugar daddy," an executive who is interested and has a budget to fund a project, even if it doesn't deliver immediate payoff. The sponsor also needs a story to share about the project. This doesn't mean that you explain all the technical details of the predictive analytics algorithm, but it does mean that you make sure the sponsor can conceptually explain the benefit to other executives, both internally and externally.
Choosing the right projects is also critical. Moyse said advanced analytics tools should not only address a real business need but also should be fun and engaging to use.
For Kelly McGuire, vice president of advanced analytics at Wyndham Destination Network, keeping members of the analytics team close to lines of business helps accomplish some of these objectives. Her team has business engagement specialists who function as an interface between the analytics team and the business, which helps keep the analysts focused on delivering data products that frontline workers need.
For example, the team recently created a mobile application for agents in the company's U.K. cottage rental division. The app uses machine learning algorithms to predict the amount of money property owners could generate from their property by listing it through Wyndham. Agents can show property owners how that amount is likely to change if they make alterations to their listing, such as allowing pets or short-term stays.
Be proactive about solving problems
McGuire said this app has had good utilization because it neatly solved a problem for agents and drives business results. Without working closely with this line of business, the need may not have been apparent.
"What you're doing is demonstrating benefit to the business in a process that they understand," she said.
Another way analytics teams can prove their value is by increasingly taking on more advanced analytics projects. At this point, most companies have figured out basic business intelligence and there isn't much value in improving existing reporting projects.
Elisa Gois, chief analytics officer at MGM Resorts International, said the real value these days is in doing predictive analytics. These projects can open up new value streams that more reactive reporting can't. Also, she said the past isn't always the best predictor of the future, so knowing what happened six months ago doesn't really help a line of business that much. Having some idea of the future does.
She manages teams that predict labor expenses, optimize marketing campaigns and improve customer engagement. These predictive projects have all grown out of efforts to improve the value of existing reports and were developed in close consultation with lines of business.
"You really need to transition your company to focus on predictive analysis," she said. "You have to make sure that your priorities are aligned with [business] priorities."
Everything you need to know about building an advanced analytics program
Building advanced analytics tools is a little bit science, a little bit art
Advanced analytics can make your company more data driven