Four Principals For Building A Better Analytics Team

One of the most important elements of any effective analytics program is its guiding set of principles. I’m not talking about a mission statement or core values, although those should both play a guiding role in your analytics program, as well. What I mean by data principles is a solid framework that defines how your team is willing to gather data, how it can be used, which data needs to be treated differently than other information, and other types of issues that provide “bumpers” for your overall data protocol.

It’s no secret that an overwhelming majority of data projects fail. Data principles may help you avoid that fate. They are by no means stiff rules that leave little wiggle room for growth or creativity. These are just meant to be guideposts as you think about creating your team and collecting data. Even if you’re well into data collection in your company, now is the perfect time to press pause on your data programs to see what’s working and what isn’t. Let these four principles act as your guide.

Now more than ever, companies understand the need to be agile and pivot quickly. You can’t pivot quickly — or meaningfully — without data. For instance, companies that once sold primarily in brick-and-mortar stores may have paid little attention to the online habits of their customers. Now, they have little option not to. Being able to change the type of information you collect is part of growing your business in a changing world.

There are, however, certain elements of your data principles framework that should always remain the same. Some countries, for instance, even have data principles to guide their analytic work. These types of principles provide a moral foundation for data use throughout the country. In the United Kingdom, for instance, data principles include understanding that data sets are assets that must be managed throughout their lifecycle; that data re-use is important and will be created with common terms that encourage that concept; that data will be governed with clear rules regarding sensitivity; and that data will be used with openness and transparency. They’re simple rules but they offer a clear pathway to the use of data in essentially any context.

Regardless of your industry, there are a few types of data principles you might want to consider in ensuring meaningful impact with your data strategy. The following are my top four.

1. Data must be diverse. First and foremost, you will want to make a commitment to create diverse data sets that allow for well-rounded, complete profiles of your customers. Without diverse data, you’re gaining a glimpse into just one small part of your customers’ lives. Just as social media generally offers a “too good to be true” view of personal lives, looking solely at your customers’ buying patterns of your products, for instance, offers a weighted view of their commitment to your company. Instead, it’s important to find out more: where else do they shop, what types of issues do they have that your company can fix, why do they need to re-order a certain part or service at a certain time each month? The more you know, the more deeply you can understand your customer and create a sense of loyalty over time.
2. Data teams must be curious. Just as you want your data sets to be diverse, you want the people looking at your data to be diverse, as well. I’m not talking about cultural diversity here, although that is important, as well. I’m talking more about diverse perspectives — insights from the sales team, marketing team, finance team, etc. Insights from people in product development, those on the social media team, and those who just tend to think differently than everyone else in the group. The more curious your team can get regarding your data, the more surprising and complex your analytic insights will be.
   This just makes sense when you think about the accidental biases that are being built into our AI and analytic systems. If we start working with more people from different backgrounds within the company, those biases will likely start to disappear letting you get an unhindered view of your data.
3. Data environments must be collaborative. As noted by the UK’s list of data principles, data is meant for re-use. It is not a one-and-done thing. The more data you have at-hand, the more teams within your enterprise that can likely benefit from it—not to mention your larger industry. As such, ensure that your data environment is fully collaborative. Create common language to tag and sort. Build compatible systems that allow for seamless sharing. And invite teams to the table who can benefit your team and vice versa.
   Silos never work. And data silos are often the culprit for failed data projects, data swamps, and unused data. Empower your organization to collaborate more with your data. Use a CDP to hold data from multiple sources and make it accessible by all departments — it’ll make a difference.
4. Data strategies must empower decision-making and action. Without the ability to take swift action, data is pointless. We have seen this with many large companies that invest in AI and real-time analytics, only to find that old legacy-era bottlenecks prevent the quick use of data in increasing efficiency or decision-making quality. As such, before you even launch a data strategy, ensure that there is a commitment among all leaders to enable and empower decision-making and action across the board. And of course, when empowering your people to make decisions and take quick action, it’s essential to create an environment where it’s OK to be wrong, so long as you learn from it.

There is so much data available these days that launching a data program without a guiding set of principles would be like diving into a rabbit hole of data without leaving bread crumbs to guide you back to the surface. Every company needs the ability to return to center in this world of constant shifts and change. Your data principles should offer that, and if thoughtfully created, will create lots of value, as well.

Futurum Research provides industry research and analysis. These columns are for educational purposes only and should not be considered in any way investment advice.

The original version of this article was first published on Forbes.