CIOs wish for simpler ways to wrangle data and experiment with business models – but change remains hard to scale. Also, it may be time to stop chasing “alignment.”
3 things people get wrong about analytics
Did you use that report data as intended? Let's be blunt: Did you even read it?
I’m a firm believer that when discussions and decisions are based on actual facts, life is better. Companies are better, society is better. Opinions, gut feelings, and pre-existing biases are all forces that compete with real data and facts. One would assume that by empowering users in our companies with the data analytics they need, they would eagerly engage with and consume anything we give them, right?
But although analytic insight is at people’s fingertips, many still don’t utilize valuable and available data resources. Throughout my career of designing, developing, and deploying business intelligence and analytics solutions, I have observed a spectrum of users that embrace and engage these solutions differently.
Inevitably, there are the data geeks – those who can’t wait to get their hands on the solution. They will reconcile the data, tell you all the things that are wrong with it, but ultimately use it for what it was intended – to identify insights that can help drive better decisions. But in my experience, only a small percentage of people actually think this way.
Many more are likely to be indifferent to a new solution – they may log in occasionally, consume some pre-built reports – but more often than not will remain infrequent users who gain little value from the new system.
This situation has not really changed that much in the last couple of decades. Interacting with analytics inside a company doesn’t feel radically different than it did ten or even twenty years ago. Sure, we have better-looking tools, we can process larger data volumes and so on, but data users are still looking at reports, spreadsheets, and dashboards.
We’re drilling, pivoting, filtering, exporting, and maybe even printing. The basic units of interaction, and the tasks that we perform, remain largely the same.
That’s a report, not a solution
Despite our best intentions, we continue to try to solve the problem in the same way, and we continue to achieve largely the same outcomes.
First, we assume a report, chart, or dashboard is the solution to everyone’s problem – and that by arming them with a variety of these, we are empowering them with everything they could possibly need.
What I have actually observed in my user constituencies is something quite different. I see people who neither have the time nor make the time, to look at their data. They view themselves as too busy or don’t prioritize this highly enough. Yet they will happily respond to emails, important or otherwise, all day long.
Many companies solve this problem by hiring dedicated and expensive analysts to look at the performance of different parts of the business and distill insights down into monthly reporting packs, replete with commentary and analysis. More often than not, the reporting packs themselves are as unused and unread as the dashboard solutions they are designed to supplement.
Second, we assume the people we hire for the solution have the skills to effectively interact with the data and tools. This can manifest at a number of levels, such as understanding the nature of the data they are reviewing and what a metric truly measures (customer churn, for example); understanding how that data is presented to them (such as interpreting a scatterplot chart); and the ability to interrogate the system further if the answers they need are not immediately obvious (using techniques such as filtering, drilling, etc). While modern BI tools are easier to use than Vlookups in Excel, to most untrained users, the user experience is not as intuitive as it could be.
Third, we assume all the answers are in the data, and the viewers of that insight will all interpret it consistently and use that information to make the best business decisions. The reality is, an insight contained in a dashboard or report rarely has all of the context required. Often that context comes from the addition of insight that is external to the system; the interpretation of a subject matter expert; or through discussion, debate, and agreement between colleagues.
Automation and machine learning bring changes
[ What's coming next? Read AI in 2019: 8 trends to watch. ]
Let’s face it: Data users want to be told only what they need to know, as too much unstructured data is painfully difficult to get our heads around. The automation of the discovery of data insights promises just that: The system analyzes the data for you and simply tells you what you need to know.
BI and analytic tools will start to feature automated discovery as a core part of their platforms. They will improve in other ways, too, such as the creation of advanced capabilities to collaborate around data insights with layered contributing data. The ability to share powerful and engaging narratives will become easier to embed, allowing analytics to be delivered at the most relevant point of consumption, and provide continuity for other business groups and processes.
Visualizing numbers on a dashboard offers limited information for making intelligent decisions, when what you really need to know is how those numbers relate and are impacted by cumulative data. New analytics platforms are being designed to rethink the entire user experience, not just for savvy data analysts, but for all levels of users. By building a compelling story about the numbers, users become more engaged with the analytics and can better understand how to leverage the context and narrative around the data to improve business functions and customer interactions.
[ How can automation free up more staff time for innovation? Get the free e-book: Managing IT with Automation. ]