Ask almost any executive at any product company today if they would classify their business as “data-driven,” and you’d get an enthusiastic “yes!” But dig deeper, and you may find more good intentions than you will actual best practices.
The advent and accessibility of software using advanced technology like artificial intelligence and machine learning has helped many businesses use data more efficiently to identify their product-market fit and meet customer demands, but much of the time, this software takes business close–but not to the finish line. Throw in the inherent challenges around communication, data quality control, and internal power structures, and the problem seems
insurmountable.
But it isn’t. Software continues to evolve to address these inherent challenges–three of which I have seen dominate businesses’ ideas of “data-driven” throughout my 20 years in the analytics industry. What are those three challenges, and what should businesses be doing instead to become truly data-driven?
Messy communication → Streamlined Communication
Using data to solve a big problem requires the involvement of multiple teams, from engineering to executives. Each of these teams uses multiple platforms to communicate information, from email to slack to in-person chats. The irony is that having so many options to communicate has given way to other issues: try finding that fifth version of that second data chart in a Slack conversation you had days ago (and be sure to skip past the messages about what’s for lunch or where the happy hour is). Or, you can sift through that endless email thread where more and more people have piled on and inserted various versions of various data reports.
This is built into the backbone of modern workflows, and businesses can avoid the traps these methods set by focusing their communications around data, and working from a single source of truth. Relegate conversations to one platform, ideally in the same place where the data lives. This is extremely beneficial when the platform uses AI and machine learning, as the conversations about the data will fuel better insights over time. (We call this augmented analytics, smart technology that improves over time to generate insights and solutions for users.)
Confined Data → Democratized Data
Data is precious and is therefore guarded with the utmost care. That’s important, but take it too far and it can be a hindrance to building a data-driven culture. Typically, the ability to build queries around data has been given to a select few within an organization–meaning all queries come through them, causing bottlenecks, misinterpretation of needs and goals, and doesn’t involve other teams in developing a data-driven mindset.
Businesses are waking up to the fact that non-technical employees can offer unique insights–and ask questions specific to their mission that others might not have thought of–that could materialize into a company’s “next big idea.” (Take, for example, the Flamin’ Hot Cheetos story .)
Choose a platform that makes data discovery possible for everyone, from marketing departments to business development executives. Because these employees aren’t always tech-savvy, it’s important to provide them with a framework that is user-friendly and simple–but still powerful. This will lighten the load of the few who handle any and all queries, remove the time spent pulling data (again) because the original request was unclear to the querier, and take advantage of the specialized background knowledge of different parts of a business–creating a truly 360-degree, data-driven culture.
Partial Context → Comprehensive Context
All the data in the world can’t do much for anyone if it’s examined in a vacuum. First, the allure of collecting every data point possible makes it easy to lose sight of the question businesses are trying to answer. And, even with all that data, there may still be places where businesses should be looking, but aren’t. Dig further, and you may realize you’re asking the wrong question.
We call this benchmark analytics, as it provides further points of reference to better understand what your data is telling you. Companies that want to be data-driven need to create a standard around this, which is something software can help execute. The right tools will help teams hone in on the KPIs that matter to them, and more easily work with customizable data that is most pertinent to a product or industry. Additionally, things like a more robust historical context can save huge amounts of time—I’ve seen days wasted on data investigation only to discover a supposed problem was just an anomaly.
Following this year’s Product World , I was excited to hear so many conversations focusing on this new wave of best practices around data–and I think if more businesses start adopting these mindsets and tools, the next few years will see some intense competition (and much happier customers).
About the Author
Alex Li is the founder and CEO of Kubit, a smart analytics platform that turns any data warehouse into an analytics powerhouse by bringing transparency and collaboration to the data journey. Previously, Alex was CTO of Smule, a mobile app developer with 50 million monthly active users; VP of Engineering at Booyah; Principal Architect of Jasper; and founding engineer of eBay Kijiji. He earned an MS in CS from Columbia University, BS from Shanghai Jiao Tong University.
Sign up for the free insideAI News newsletter.
Speak Your Mind