A Winning Data Strategy: Capture It All and Keep It Governed

PinIt

Even the most advanced analytics tools remain useless if they’re not fed a complete, well-governed set of user data.

Businesses have been running on data for decades, but we’re on the cusp of a new era for analytics. Tools are moving from being fundamentally reactive – helping teams diagnose and respond to a suspected issue – to being fundamentally proactive: identifying important phenomena at their inception and locating problems before users communicate them to you. This is a major shift that constitutes a winning data strategy, which will save tremendous time and effort while helping people make quicker, more anticipatory decisions.

There is a catch: analytics tools are only as powerful as the datasets available to them. Even the most advanced tools remain useless if they’re not fed a complete, well-governed set of user data. It’s the equivalent of buying a Ferrari and filling it with low-quality gas: it may look nice, but it will never give you the performance you paid for.

All of this is especially true if you’re looking to your platform to provide proactive analysis or to identify patterns and make recommendations you wouldn’t have found on your own. But most companies won’t be able to access all this value, at least not yet, because they haven’t built the data foundation they’ll need to do so.

Maximize the return on your Snowflake investment with insight into performance,  quality, cost, and more. [Learn More]

To use data most effectively, you need a winning data strategy that looks at all of it for multiple reasons.

First, you can’t get an objective understanding of your users without a complete picture of their behavior. If you’re looking at data selectively, e.g., based on a tracking plan, you’re only paying attention to the information that you assume will matter in the future. In doing so, we encode our preconceptions and biases into the dataset itself. Capturing data selectively means that you’re actively choosing what data to not collect. It means you’re deciding – before any data even comes in! – what data won’t matter to you, your team, or your company in the future.

The idea that you can know in advance what data won’t matter to you is, of course, ludicrous. Game-changing insights often arise from unexpected sources. One might even argue that the point of data is to show you things you wouldn’t have found on your own. Yet many behavioral data tools, even those with great analysis capabilities, force you to decide in advance what data to capture and – as a consequence – what data not to capture.

It’s easy to see the problems with this approach. Say you want to improve the conversion rate of a purchase flow. You might notice that over half of people get to the final purchase page but don’t click “buy.” Let’s imagine that a huge chunk of those people are leaving your purchase page to visit your returns FAQ page. If you had that information, that might be a strong indication that your return policies are less clear than you had assumed. An easy fix would be to add more information about returns on your purchase page. If you’re not tracking behavior on your returns FAQ page – who knew that would be important? – you’d never know this was the problem. It’s an easy fix, but unless you’re tracking everything, you’ll miss it.

If completeness is one-half of the data foundation, you need to put proactive tools to work, and the other half is governance. A small dataset is easy to keep organized and consistent. But more complete datasets become unusable if they’re not governed. At many companies, even established brands, governance of behavioral data is conducted via spreadsheet. (This is often called a “tracking plan.”) This is usually a recipe for disaster. All it takes is one person forgetting to update the sheet to introduce inconsistencies, which have a tendency to compound. Enough inconsistencies, and the dataset becomes unusable.

We can do better than that. A winning data strategy includes automated governance capabilities that are built into the capture and naming process itself are more scalable and can be used to create a more trustworthy, understandable dataset. This allows for consistency in naming and organization without needing to tightly control a spreadsheet maintained outside of your analytics tool. It additionally makes it easy to maintain security and privacy, which is becoming increasingly important to consumers (84 percent of whom want more control on how their data is being used), and from a legal perspective (GDPR, CCPA.). This organized, secure dataset is crucial for proactive insights: if you deliver your data in a state of disarray, you’ll get results you can’t trust.

Proactive insights are coming, and they’ll be driven by advanced, smart tools that will surface the most important, most reliable signals about your users. But that’s only true if those tools have access to an organized, complete set of information, which is what effective strategies for capture and governance are all about. When proactive insights arrive, it’ll be a massive competitive edge for the companies that have the foundation to access them. Do the foundational work now: capture all of your data, and keep it governed. As these new technologies come online, these are the inputs you’ll need to make the most of them.

Dan Robinson

About Dan Robinson

Dan Robinson is Chief Technology Officer at Heap Inc.

Leave a Reply

Your email address will not be published.