Business stakeholders often claim they “don’t need complex solutions.” This assertion regularly comes from those leaders who have already suffered through unsuccessful AI or machine learning initiatives. Initiatives such as these tend to be costly and become cautionary tales for employees.
On other occasions, analytics leaders confidently claim they are “not interested in doing reporting.” These individuals likely have already made a few attempts at traditional data projects with difficult-to-measure key performance indicators (KPIs) and failed to impress their upper management.
Yet, companies today need a mix of modern and traditional solutions. And analytics leaders stand at the crossroads of often-times opposing wants and needs. They want to enhance a business, but data intelligence can be complex, making it tricky to implement.
How can we better manage expectations for different stakeholders? How can we incentivize our analytics leaders to make the right decisions? And how can we better communicate with our business stakeholders to see the value of occasionally taking bolder actions?
Finding answers to these questions can also help with recruiting. The recruiting scene for data leaders in IT is currently much like the wild west. Performance-tracking and related incentive structures are in their infancy, and hiring can happen in a haphazard manner, with little caution or reflection. As in the classic western film: The Good, the Bad, and the Ugly, this process includes elements of all three traits.
Clint Eastwood in ”The Good, the Bad and the Ugly”, 1966 | Produzioni Europee Associate, MGM, United Artists
Digital transformation requires defining a few core terms
Before we take a deep dive into what we believe should be the key performance metrics, let’s cover some basic terms.
What is Business Intelligence (BI)? It is a traditional yet vital function responsible for developing and maintaining data warehouses – central repositories of integrated data from one or more disparate sources. BI usually sits within IT and is generally a highly technical area of expertise. BI does not get much visibility with business management, but BI warehouses power all the critical metrics, from financial to operational KPIs, marketing, sales, and more.
What are Artificial Intelligence (AI), Machine Learning (ML), and Data Science? These are the new incumbents, professionals with a rich statistical and mathematical background, discovering new and profitable ways of utilizing data. They are “the smart kids that will make us cash”, so they are incentivized mostly by hard KPIs very often directly connected to ROI. They have plenty of business visibility because of the direct link to these KPIs and ROI.
The Ugly: Priorities
Due to the potential value that AI and ML solutions can bring, upper levels prioritize AI and ML initiatives. They want to see value-driven data project initiatives coming in. On the other hand, middle and lower levels are in charge of operations and need more standard, and straightforward reporting features. Analytics leaders are in the middle trying to figure everything out.
New analytics leadership focuses primarily on running data projects with measurable KPIs. While this is a priority, it takes away from all the other, less quantifiable work needed to have stable data solutions in place.
What type of less measurable work? One example is to operate a well-run technology platform that analysts can use to model. Providing access to data in warehouses or lakes should then be simple. Clearly, having uncomplicated but effective reporting solutions for operational business results is fundamental
Poorly supported tools heavily impact the careers of data scientists and analysts. Therefore, one can imagine why they so often tend to change careers.
Our study shows that 35% of the data scientists have been at their current jobs for less than a year. Additionally, 30% have started their present employment within the last two years.
The Bad: Expectations vs. Reality
Companies often hire data science and AI leaders without a clear understanding of their role, so they are assigned to lead Business Intelligence initiatives. Sadly, these new leaders quickly recognize the situation and attempt to adapt, but their new Enterprise incentive structures don’t allow them to.
BI and AI play different roles, and both are essential. Companies can hire AI leaders to do BI, but incentive structures usually prevent this from happening. BI was once rewarded based on support KPIs, such as stability and reliability, not value KPIs like ROI. As AI becomes increasingly incentivized by ROI, it will strive to reach its KPIs no matter what. Nevertheless, reliability and stability are essential to business stakeholders.
As a result, AI and data science leaders face a dilemma. Should they pursue ROI projects to satisfy upper management? Wouldn’t it be better to fix infrastructure, improve warehouses, and do BI to please their peers and subordinates?
The Good: Evolution to Equilibrium
As technology advances, AI and BI are becoming more convergent. Consequently, the distinction between BI and AI is becoming less apparent. An opportunity now exists to bring the data organization incentives closer and group the support KPIs and the value KPIs into one department. The other option is to keep these two departments separate, which rarely happens.
As companies mature in data analytics, they must invest in suitable activities and arrange their data focused organizations in a way that will successfully support their business needs and long-term business goals.
This effort requires exemplary data leadership. Before recruiting, a company must do their homework. To plan a future data leader’s role, companies must better analyze their needs, manage their candidates’ expectations, and help them direct their focus once they join the company.
It is important to remember that we cannot quantify all aspects of data efforts in these analyses. Even so, companies must focus on fixing the most critical problems in this area. This includes, for instance, so-called support KPIs. It can be difficult to scale and maintain continuity if a company disregards these aspects.