Data-Driven Digital Transformation
By Russell Villemez, Senior Partner, Dialexa
Current State Analysis is Foundational
A common technique used to plan a transformation is the familiar three-phase project approach: current state analysis, future vision design, followed by the gap plan/roadmap. But some efforts will gloss over the current state analysis because planners either presume to already understand the current state or view the future visioning exercise as more important or more urgent. And their stakeholders often buy into this approach, because, after all, who’s interested in rehashing out how we got here, when it’s more interesting to talk about future opportunities?
However, across dozens of clients, we have found that the current state analysis is foundational to a well-planned transformation. When the current state analysis is skipped, the resulting roadmap carries within it significant blind spots — such as the true nature of the dependencies between business operations and legacy systems. Such un-exposed realities can cause the whole initiative to suffer setbacks and cost overruns as they become more evident in the execution. But when the current state analysis accurately sets the baseline against which all transformative actions take place, we end up with a more realistic, actionable and dynamic roadmap that accurately captures the necessary steps required and the complex dependencies that must be addressed in order to achieve the future vision.
The Role of Data in Current State Analyses
Current State analyses most often start with interviews, which is a good way to collect the opinions of key stakeholders, getting to know company history and context, and preview some of the key issues motivating the transformation. But as interview responses are recorded, collated, correlated, and summarized, we do run the risk of creating an output that merely tells the stakeholders what they already know. Now it’s true that these stakeholders may not have ever received an analysis that aggregates or integrates their own collective insights in way that an unbiased consultant can deliver. But how can we add net new objective facts and insights, and reveal the patterns and root causes that may not be a part of the stakeholders’ pre-existing appreciation for why their outcomes have been what they are?
For example, most clients have a hard time getting their arms wrapped around some fairly important aspects of their current operating state:
· The number of business applications in use across the enterprise
· What are the functional redundancies in the application portfolio, and what are the potentially nuanced differences in why those redundancies exist?
· The scale of missing or inadequate business capabilities required for the transformation
· The true non-discretionary cost of running each application within the portfolio
· The role that each application plays in the overall data ecosystem, and the resulting dependencies created on making modifications to them.
· The degree of dependency on service providers and contractors in keeping the lights on
· The scope of complexity associated with competing technologies and sources of data
On the surface, obtaining insights such as these may seem fairly elementary. But you would be shocked to find out just how many companies can’t even answer the first bullet above (which is foundational to all that follow), or how surprised they are when they finally obtain the answers. Of course, these are just a few of the current state realities that must be dealt with in the execution phase. But whether revealing brand new insights, or simply adding texture to interview outcomes, the specific nature of these topics (and how they relate to each other) are out of reach without data.
Data includes inventories (software, hardware, people, and places), process designs, organization charts, financials (actual and budget), capability models, data models, business-of-IT operational statistics, and all kinds of data that have often never been collected before. Or they may have been collected, but they exist in multiple formats across multiple organizations, and have never been normalized for analysis. Or they may be available in a single format but are not relatable to other, equally important artifacts, or they are missing critical attributes for analysis.
For example, an IT organization may be able to inventory all the applications that support a given business function, say, order management. But that inventory is not mapped to a canonical enterprise business capability model for order management, so it’s hard to determine which system has coverage for which order management capabilities. Or which order management capabilities should be delivered by common enterprise solutions versus unique local solutions. This makes functional gap analysis very hard to do, or at best, subjective.
Additionally, that inventory may not capture the annualized labor and licensing cost of supporting each order management application, without which it will be impossible to assess the relative budget impact of replacing any of these apps during a given year of the roadmap. And it may not capture the technology stack sitting underneath each order management application, which makes it very easy to oversimplify the effort to modernize/digitize each step of the order management process, much less the technology skills required to do it.
Blending Data and Interviews for Real Insights
A current state analysis should not feel like sequential information gathering. It should feel more like an archeological dig. Our search for concrete artifacts in turn points to other artifacts. Even the absence of the needed artifacts (a dig with dead ends) can be a valuable observation on its own, as it may point to poor IT governance. As a result, the interviews should not be the end of the story. Every current state analysis should anticipate the need to blend real data with interview notes, and the need to dig elsewhere and dig deeper, until the full story of what “is” can be told. Finding the people with access to the required data is just as important as identifying the initial interviewees.
For example, in a data-driven approach, you must also gain access to the lower-level IT people that have intimate access to databases, run books and spreadsheets that are the foundation of how your technology stacks are managed. This data can take you from simply understanding that there’s a complex mess to knowing precisely which applications, which business capabilities, which business locations and business units, and which permutation of each are costing the most money to support. Among other things, insights such as these can make the difference in choosing the best “beach head” from which to start the transformation.
Last but not least, visualization of the problem areas becomes possible when you have access to data across the intersections of business, technology, financial and organization. Whether through simple heat maps, integrated dashboards, or advanced multi-dimensional views of your business, the process of understanding the true state of your business becomes much easier to digest while the insights become more objective. No more leaning on just the narratives of stakeholders. The truth is in the data, and the conclusions to be drawn from it are often a surprise to the stakeholders themselves.
A data-driven framework, such as Dialexa’s IT Cube offering, is able to capture, orchestrate, and visualize the data needed to perform ongoing analysis as a part of a digital transformation. Derived from the Feld Group Institute’s Management Framework, it leverages proven models for analyzing business models, business systems, technologies, information flows, economics, organizations, etc., for quantifying and visualizing the complexity of interdependencies within a technology organization.
Beyond the Current State (and Beyond the Plan)
Practically all transformation plans are executed in phases. Consequently, there must be some rationale as to what each phase entails. What is that rationale? At a minimum, we have found that when that rationale takes a “construction-dependency” approach — so that each phase is informed by the phase preceding it, and vice versa — the entire sequence is bound together in a least-cost attainment of the future vision. To that we then incorporate the best business sequence, which could be driven by market-facing objectives, business-unit-specific objectives, cash flow implications, and a myriad of other big priorities in play. And once we have adjusted the initial build-sequence for these other variables, we apply funding, people, and other feasibility constraints to arrive at a fully interlocked execution roadmap.
This is the hardest part of the planning because some really complex dependencies and collisions across competing priorities and organizations need to be worked out for the good of the enterprise. It’s an incremental negotiation of technical, economic and political priorities, all rolled up into a careful balancing act. But armed with data, these “interlocks” can be traversed more quickly and orchestrated with less emotion. Data helps everyone operate off of the same facts of how we got here, where the starting point is, what’s doable, and how fast we can go. As a result, data improves buy-in to a shared plan.
A subjective current state analysis without data leads to vague roadmaps that are directional instead of prescriptive. They allow key interlocks to be swept under the rug, simply delaying the eventual need to resolve cross-organizational issues. It allows all of the root causes and patterns that created the current state to begin with to be glossed over and not attacked directly in the early phases of the roadmap. And as execution gets underway and blind spots are revealed, organizations struggle to drive the program management specifics of what must be done — first, next, or last.
In contrast a current state analysis formed from both interviews and data creates a starting point for your digital transformation that is grounded in terms of understanding the work necessary to reach your future state, and the sequence in which it should be tackled. And this, in turn, is what creates a roadmap that prescribes how to achieve the future state — at every step along the way. Instead of a collection of individual projects that are, at best, related to each other only through a thematic lens, this kind of roadmap describes an integrated sequence, paced across time and across each of the dimensions (business, applications, technology, budgets, and organizations). It creates a roadmap that incrementally builds the future state as opposed to simply pointing thematically in that direction.
Last but not least, data-driven analysis helps you come to grips with “ongoing reality”. It leads to roadmaps that are consulted and refined every step of the way during execution. It produces roadmaps that become the “plan of record” by which progress is continuously measured. In fact, the data-driven visualizations from the current state analysis should be updated every quarter as a part of that measurement. It is even possible for these current state artifacts to become key measures used in day-to-day technology operations. These kinds of current state analyses (and the roadmaps they initiate) do not sit on a shelf. They become the operational lingua franca for execution.
There is a reason why so many companies don’t conduct this kind of deep, concrete current state analysis: It requires focus and a deliberate effort. It requires an inquisitiveness about patterns and root causes, not just symptoms. It requires systems thinking to reveal (and in the execution phase, how to unravel) the multi-dimensional aspects of why things are the way they are. And in each of these, it requires data that can quantify things like scale, complexity, outliers, gaps, etc. in the various dimensions of the current state. This is just hard stuff for most organizations, especially without a framework and set of proven practices for implementing it at speed.
Digital transformation is too important to leave to chance alignment and vague notions of “unifying themes” and silver bullets. Instead, start — and sustain — your transformation with data. And in the process, you will come to rely on that same data to manage ongoing enterprise business alignment, the pace of legacy systems modernization, progressive digitization of the customer experience, and continuous improvement of the overall IT operation itself.
Dialexa builds the world’s preeminent technology solutions. We are the arsenal behind the world’s largest and most successful companies, building revolutionary technology products that solve today’s complex business challenges. Please contact us if we can assist with your digital transformation or data science needs.
