How much knowledge and institutional wisdom can your organization afford to lose? How successful could you continue to be if you decided to throw away lessons you’ve learned? And what if you’re doing that right now in your effort to become data-driven? This post describes what that looks like, with steps to ensure you’re not backsliding on your analytics transformation.
Losing IT — Both ancient and modern
Sometime between 200 BC and 87 BC, Greek scientists built a mechanical computer consisting of small, interlocking bronze gears. It is thought that the computer was used for astronomical calculations and could determine planetary locations 10 to 100 years in the future. In 1901, divers discovered the computer in a shipwreck off a Greek island, and the Antikythera mechanism received its name.
Researchers, scientists, and citizen scientists are amazed at the complexity of the gadget and incredulous at the prospect of a 2,000-year-old computing machine. There is no evidence of similarly advanced machines for many centuries afterward. How could information technology have progressed that far, then been abruptly stubbed with such a huge loss of knowledge?
That’s an example of losing ancient IT.
Now fast-forward a couple of millennia and consider a global IDC survey of 1,500 business leaders, 76% of whom see the business landscape changing faster than ever before. Furthermore, 73% of their organizations indicate that analytics spending will outpace other investments. But at the same time, only half base their business decisions on analytics, and 93% of them are not fully using the analytics skills of their employees.
IDC looked at four investment dimensions — comprehensiveness, flexibility, ubiquity, and usability — and observed that ubiquity lagged the others. In other words, the organizations surveyed are making progress in analytics but not giving everyone access to it.
That’s an example of losing modern IT. In another example, IDC breaks down the $90.4 billion of investments in big data and analytics software for 2021 as follows:
- Analytic and data management platform – $30.8B (34%)
- BI and analytic platforms – $20.3B (22.4%)
- Enterprise performance management and apps – $23.4B (25.9%)
- Other (AI and geospatial) – $17.7B (17.6%)
Notice that the investment in BI and analytic platforms lags the investment in the data management side.
Businesses typically invest heavily in data management, but the investment in BI and analytic platforms has always lagged. There are standard processes for data management, but for the most part, analytics is the wild west.
The point is that you can spend a lot of money yearly on expert teams of data engineers, data scientists, and ML engineering teams and departments. But it’s no guarantee that you’ll become a data-driven organization anytime soon.
Don’t get me wrong — plenty of data science teams use our products for developing advanced analytics and artificial intelligence (AI) models. I certainly don’t want to antagonize them. The problem is that it takes more than an investment in experts to get you on the path to a data-driven culture.
For that matter, in your quest to become data-driven, you may already have in place large quantities of the most vital element: the people. So, what is your organization to do? Your best hope is probably the citizen data scientists who are hiding in plain sight.
3 steps to accelerate your analytics transformation
We’ve written before about citizen data scientists and what makes a good one. A citizen data scientist is “a person who creates or generates models that use predictive or prescriptive analytics, but whose primary job function is outside of the field of statistics and analytics.” Here are three steps for accelerating your analytics transformation by investing in your citizen data scientists:
1. Identify the people, roles, and skills that make the business run.
Citizen data scientists work in the trenches day in and day out. They know what makes the business tick and they know which analytics and measurements business leaders need to steer a course to profitability. If — and that’s a big “if” — they have the right skills and tools, they can generate those analytics quickly, then crack on with higher-value tasks. But if they don’t have the right skills and tools, they spend their time tying together disparate data sources and cleaning up data in creaky spreadsheet models every month.
Another IDC Survey (The State of Data Science and Analytics) highlighted that knowledge workers use, on average, four to seven tools to perform analytic activities.
Figure 2: Workers Use Four to Seven Tools for Analytics
2. Ensure they are present and collaborating toward analytics maturity.
Some companies try to accelerate their analytics maturity by building on the model of a center of excellence. They’re trying to integrate the data science team to the fabric of the company by centralizing data and analytic skills.
But by nature, domain expertise and business knowledge aren’t centralized. They’re in all the corners of the organization, applied by people in various roles.
You’ve already made the big investment in your citizen data scientists — probably without even knowing it and certainly without calling them that. You’ve already invested years of salary, overhead, and training in their knowledge of and importance to your business. Now that they’re hiding in plain sight, it’s time to integrate them into your data science effort.
3. Invest in upskilling your citizen data scientists.
Alongside the money, effort, and time you invest in building your centralized data science team, carve out a more significant investment to upskill your citizen data scientists. That usually involves funding for training and for easily accessible, self-service tools they can use to mine insights from the data they work with every day.
The way to achieve that goal is to give them access to and training in low-code/no-code tools. Spreadsheets have their place in every company, but upskilling introduces your workers to higher-level tools that allow them to easily wrangle disparate data sources and huge numbers of rows.
The tools should promote the automation of analysts’ routine tasks. They should be powerful enough to make data science professionals more productive and accessible enough to convert line workers into citizen data scientists with just a little training. Smart companies, like Jones Lang LaSalle Incorporated (JLL), upskill workers by encouraging adoption through gamification. What an amazing idea!
The goal of upskilling is not to turn citizen data scientists and analysts into data science professionals. The goal is to free them up to derive insights and iterate quickly to solve all the small problems in the business that the data science team will never have the time for.
Isn’t that what they call “democratizing” data science?
It is. In a way, it represents a re-thinking of the full-speed-ahead approach to big data and the promised land of AI, for a few reasons:
- Data scientists have complained since the beginning of the era that they spend more time corralling data and cleaning it than they do developing algorithms. That has not changed much.
- The work of data scientists thrives in a data-driven culture, but it cannot create that culture in an organization where it does not yet exist.
- Successful data science projects yield predictive models around concrete business problems, like customer churn and inventory optimization. That requires collaboration between data science teams and knowledge workers in the business; unfortunately, there are too few effective collaboration tools suited to that.
Every organization has countless small-ish problems that a centralized data science team will never get to. Nevertheless, citizen data scientists understand those problems intimately and can make a dent in them — if they have the right tools and receive upskilling.
That leads to a dollars-and-cents perspective on democratizing data science. The nature of data science is that the professionals will likely work on complex problems, saving you large piles of money on a few problems. At the same time, once you’ve upskilled the citizen data scientists across your company, they can work on smaller projects and save you small piles of money. But because they outnumber your data scientists, the sum of those small piles will be much greater. And, as you continue to upskill your analysts, the value they capture will grow. As data science continues to evolve, its roles and skill sets will become more diverse and less centralized. A sensible approach to democratizing advanced analytics and AI, then, is for data science professionals to mentor citizen data scientists. Smart organizations will empower citizen data scientists with tools and training, then have data scientists ensure the resulting models are robust.
Invest in data science teams and in citizen data scientists
The role of a citizen data scientist is best suited to people with some affinity for information systems but also patience and skills in consultation and communication. They can translate between the business problem and the technology tools, then navigate the practical constraints of the IT infrastructure.
Citizen data scientists and data science teams are not mutually exclusive. In fact, according to Gartner, the former complements the latter. When data science expertise is scarce — as it will be for some time — the citizen data scientist hiding in plain sight could help accelerate your analytics maturity.
Don’t let your organization become the next Antikythera mechanism; another example of IT progress abruptly stubbed. Invest in your citizen data scientists now because they work at the most important levels in your business, and they will for years to come.