The future of business is nonlinear, which means organizational structure must also evolve. Three subscribers asked organizational structure questions this month. Each question is different and focused on a micro problem, but the threads are common at a macro level.
When people ask questions in classes and consulting engagements, they get a long-winded response. I’m doing much the same in this article, so I should explain myself. Data professionals share a common experience and have common questions because the root causes are the same.
I learned early on that addressing symptoms was like playing whack-a-mole. It’s expensive and takes too long. We should optimize our actions for impact. I implement a transformation framework for clients with this image.
A micro-focused question hides a desired outcome. Questions imply intent, but people asking questions rarely have the context to articulate their intent. That makes it challenging to frame a question to prompt a response that addresses their intent.
Teaching strategy has helped me understand that I must be able to:
Take a question with a narrow focus or even a problem statement without a question.
Determine the intent or desired outcome.
Deliver a response that explains the root cause and the steps to address it.
Operators want a simple, actionable answer so they can get to work. Businesses have the same challenge. Symptoms are apparent and straightforward to address, but new symptoms will emerge. It will take more iterations to solve the problem by addressing one symptom at a time, so determining the root cause is critical. Root cause analysis adds time to the iteration but reduces the number of iterations.
The Problems Emerge In The Data Organization First
Data teams are increasingly cross-functional. We support most of the business but own nothing beyond the data team. If we go too far into someone else’s fiefdom, we’re making waves or stepping on their toes. We're not producing enough business impact if we don’t push into external teams and advocate for initiatives.
We need technical resources that are often owned by external teams. If we go too far into another technical team or organization’s domain, we’re stepping on toes and making waves again. If we don’t, the data team won’t be moving fast enough and won’t deliver to production.
Data and AI strategies may be separated, and ownership may live in different organizations. Product strategy and management can also be a separate function with its own organization outside the data team or strategy organizations. Data science and data engineering are sometimes separated.
We inherit legacy organizational structures and data ownership schemes but lack the authority to fix the problems. Organizational structure and cultural debt are two of our most significant barriers to delivering value back to the business.
This is the next image I show clients. Legacy strategists hate this one, and it’s where their pushback begins.