Hiring A Data Scientist? Do This First
It’s not exactly news to suggest that data analytics is an essential part of doing business. Almost all functions have, if begrudgingly, embraced it as way to understand customers and monitor business performance. Though companies have used data to drive decisions for over a century—think actuarial tables—new technologies have recently made it accessible to a wide range of industries.
Companies that have been around for a while can draw on many years of customer data, at least in theory. Unfortunately, making those data sets perform is often a challenge, thanks to factors that range from the skills of internal teams to the structure of the data itself. But complacency isn’t an option. Companies that can’t put legacy data to use in convenient and intuitive ways will lose ground to digital-first competitors that skip over legacy systems and vigorously gather new data, promising customers something of value in return for their information.
So … bring in the data scientists? If you work at a big company, you probably have one or more groups with “analytics” in their name. There are probably at least a few people with data science responsibilities embedded in your product and IT teams. (If you’ve been lucky enough to recruit them, they may even have been educated as data scientists.) To make the most of their expertise, however—and drive the kind of digital innovation that will prevent you from becoming the next Blockbuster—you’ll need to put the issue higher on your strategic agenda.
Here’s why that is—and how you can start extracting full value from the data goldmine you’re sitting on.
Big data or a big waste of time?
Though most companies hold themselves to increasingly rigorous standards, the most familiar presentation format for data is still business intelligence, which aims to turn information into more easily interpreted “insights.” It’s a worthy goal, but it’s rarely held to any sort of dependable test-and-learn process.
That means—beyond feeling like you must be making progress—there’s no way to evaluate the impact of your data or measure the marginal benefit of capturing it in the first place. It also means it’s difficult to make decisions on investing in platform upgrades or team capabilities.
And it won’t cut it when your competitors are establishing products and processes that combine customer actions with other relevant information in real-time.
As with any investment, you’ll need to calculate the value before you can assess the approach. Do your analytics support the front office? Then they’ll have to deliver top-line growth in the form of customer acquisition, retention, and cross-selling. Back-office activities will need to help you with the bottom line, perhaps by ensuring regulatory compliance, reducing fraud, or increasing cybersecurity. And all of it will need to live up to the lofty expectations of your customers, who are likely to go elsewhere if your mortgage calculators won’t bring them any meaningful information unless they complete a 6-step process that starts with a contact form and ends with the promise that a representative will be in touch.
Taking inspiration from the leaders
To get a clearer sense of what you might do with your data, think about how it might help you anticipate and react to your customers’ needs. Financial service companies have a lot of information about people’s spending, for example. Why not harness it to make better upsell and cross-sell recommendations, or simplify navigation on digital channels? If someone has recently paid for a few college application fees, chances are high that they—or a loved one—will soon head off to school. Can you get that information over to marketing in time to offer the customer a short-term loan?
You can make your data even more valuable by combining it with other information, whether it’s behavioral or from other institutions. The major UK bank NatWest has deployed technology from BioCatch that measures customers’ behavior while using digital services, using such parameters as reaction time or screen gestures as a way of detecting when a user is not legitimate. Meanwhile, portals like Mint and Acorns pull together information from disparate sources to bring consumers a comprehensive overview of their finances and make product recommendations that are independent of any service providers. Bots can grab this data from a bank’s website in a matter of milliseconds; if you’re not doing something meaningful with it, you’re behind the competition.
New data sources like wearables, smart homes, and autonomous vehicles will facilitate even more opportunities to have meaningful, data-driven interactions with your customers.
Understanding the challenges
So what’s holding back the would-be innovators? Many companies, especially in financial services and healthcare, depend on hosted core IT systems from a limited range of providers specialized in their respective sectors—and strict requirements when it comes to customer data protection. Adding customizations to these standard services is notably expensive and requires technical expertise that’s usually long gone after the platforms have been installed.
Legacy data on customer interactions, meanwhile, may be stored on different platforms and in formats that were more often were created by the platform manufacturer—with troubleshooting and maintenance in mind—rather than with the goal of creating a decades-long dataset. And migrating to the cloud is a challenge of its own, independent of any analytics efforts.
Team dynamics and incentive structures can also impact progress:
- Executives don’t always relish diving into data and may not know enough about what’s possible to set clear goals and priorities
- Data scientists don’t like inconsistent, messy legacy data and typically have not followed a career path that exposes them to general business goals
- Programmers are essential for making data extraction, AI and machine learning work, but need direction from data scientists
How to do more with your data
Data scientists are an integral part of the solution, but they’ll never be successful without support and understanding at the highest levels of your company. Here are some practical steps you can take to refresh your data strategy:
- Define success: Think about the metrics you could use to understand customer interactions, then link them to corporate-level goals like growth, profitability, revenue per customer, and attrition.
- Establish a baseline: You may be sitting on a treasure trove of data, but is it accurate? Is it in a consistent and easy-to-access format? If not, create a business model for evaluating the effort it will take to restore and reconcile imperfect data sets—then assess the tradeoffs, decide, and update.
- Prioritize your opportunities: Of all the areas where data analytics might make a difference to your business, where should you get started? Customer interactions offer one way to size opportunities. Have they requested services that draw on more data? Could data help you deliver the services they already use faster or more efficiently? Competitor products are also worth keeping track of for ideas and effectiveness.
- Think “real time”: Actively identify digital journeys that customers will increasingly expect to be personalized for them. This offers opportunities for analytics to be deployed ‘live’ and also gain valuable experience with test-and-learn algorithms along with architecting data stores for continuous use.
- Build the culture: Facilitate greater understanding and awareness for the power of analytics at your company by sharing your plans, your goals, and—especially—your success stories. You may also need to rethink team structures and incentives to give different functions a reason to collaborate on what are likely to be complex projects, and achieve a common understanding of the overall business goals.
Get the Skills You Need
Thousands of independent consultants, subject matter experts, project managers, and interim executives are ready to help address your biggest business opportunities.