Call Me an Agent of Change, Not a Human Resource
Organizational transformation is a never-ending process that proceeds on multiple frontiers. Together with Dmitry Smirnov and Anna Bäckström we take a look at certain issues that show up during organizational transformation in pursuit of becoming more data-driven, and being able to take their data in use systematically. Instead of focusing on organizational structures that are subject to change, we want to address a different question - a question we feel is of even bigger importance, namely: what is the role of individual contributors, the people in the organization, on the overall success of the transformation?
Data-driven, what does it mean to you?
Data-driven organization is a badge that is pretty much mandatory for an organization to wear. There are hundreds of articles written on what makes an organization data-driven. Gartner publishes regular reports, consultancies write white papers, and managers drive flagship projects that are aimed at transforming the organizations into some kind of data-driven pegasus (yet another imaginary animal, since the unicorn metaphor is already occupied). So what is it?
It’s a continuum, and it’s not really so that an organization is either data-driven or not. One actionable perspective is that there are tiers of “readiness” of an organization (see https://www.wwt.com/article/data-maturity-curve for a good example). Organizations often start with realizing that data is scattered all over Excel sheets in various units, and while there is some decision-making (and sometimes even some data-driven products and models used for business) - there is very limited understanding and planning around the data resources throughout the organisation. An advanced data-driven organization would often have a dedicated data team for technical solutions, multiple analysts and data scientists, deep understanding and capability of working with data among the business users, and also products driven by data (e.g. recommendation systems, optimization models, automation).
The purpose of transforming into a data-driven (or more data-driven) organization is primarily to gain a competitive edge by a holistic understanding of the performance of various parts of the business, take advantage of the data to design or enhance the products and services of the company, and inform the strategic planning - driven by deep understanding, and not just intuition.
What is human-centered organizational transformation?
Let’s set the stage: what is a successful transformation towards a data-driven organisation? A fairly precise criterion is the value brought to the customers, it’s the purpose of the business after all. At a higher level, transformation means behavioural change, that is, the people in the organization start to act differently, and usually also perceive things differently than in the pre-transformation age. Successful transformation is more than a new set of rules and policies, it’s the change in the way individuals in the organization see themselves, their role, and the company overall.
For example, let’s return to the data and take the decision-making process. Before data, one would rely on experience and intuition, a gut feeling. After data, it’s difficult to trust individual judgment without any data to back it up. Perception is changed and we now expect data as the fundamental block of making an argument. Acquiring data, and relying on it for judgment becomes basic behaviour, compared to the pre-transformation times.
What characterises a transformation that is going well? Our experience is that successful aspects of transformation efforts will focus on individual engagement, hands-on experimentation, and embracing chaos. This can be contrasted with a focus on formal organizational processes, policies and rules, top-down regulation and exhaustive planning. To be fair, the optimum is most likely somewhere in the middle, that is, a formal framework of some kind will be needed, but what we want to highlight is the word focus here, because obviously all aspects have to be considered, formal as well, but what the organization will keep as its key target will define the way transformation goes.
What transformation do you need before you can have AI?
Depending on which tier of “data-drivenness” you are at, the possibilities and areas of focus vary. A typical thing to hear: “It doesn’t make sense to look into any AI approaches if you don’t have your fundamentals (e.g. reporting, enterprise data warehouse, etc) figured out”. Using data for anything beyond reporting is viewed as something that requires significant upfront investment, before any tangible value can be extracted.
This perspective is steering the organization into the direction of first preparing the fundamentals, the platform, and only then addressing the high-leverage possibilities. In practice, that leads to an extensive project focused on breaking down the silos (bringing all the data together) and building reporting and dashboard capabilities. Only once all the data is sitting in a unified data platform and all the excel sheets are replicated in dashboards and reports, there is perceived readiness to focus on the application of available data for core business.
Having these fundamentals surely would help, but this is definitely not a blocker for taking advantage of ML/AI solutions. In the next section, let’s look at what we at Fourkind think about this.
Data-driven, what does it mean for us at Fourkind?
We at Fourkind are firm believers in the value of experimentation and rigorous validation of any approach taken to solve a problem, as early as possible. Think about the “lean startup“ approach - putting experimentation and validation into the core of business development requires the organization to be strategically focused on informing its decisions by data.
What often happens in practice is that while (extensive in terms of time) building the aforementioned fundamentals, the organization loses focus on the true value generated by the business, and creates a plan to build something that doesn’t yet have a clear, tangible value. The time it takes to build these fundamentals is only extended by the fact that we are dealing with a moving target.
Add to that the risk of participants losing mutual trust, and the lack of engagement due to unclear or rather remote perspectives of seeing the real value of the project. The engagement of actual individuals who are to follow through this transformation is surprisingly often completely left out of the equation, especially when a structured top-down approach to organizational transformation is taken. Whether it's a top-down management-driven, or IT-driven transformation project, companies focus on tools like data catalogues, or policies and rules, but not much thought is given to the actual employees that are expected to execute the vision.
Our view is that it is crucial to start from concrete business problems, and while solving them, use the now validated approaches to develop a scalable vision of how this will grow to become the fundamental structure in the company. These concrete business cases enable quick validation of the overall approach and adoption of the practices that work. The hands-on nature of such projects allows the individuals to keep track of the WHY of their work, and keep themselves engaged by seeing the quick wins (or quick feedback) from the actual business metrics affected by the fruit of their work.
To take it one step further, here is a provocative statement: building a data platform is not a prerequisite for being a data-driven organization, and instead of this significant investment, a smaller focused effort driven by business case would have sufficed and paved the way for further more structured steps.
We are the agents of the change, not “resources”
Managing a transformation is a lot more than just providing a plan and finding the resources to execute this plan. To reach a successful ending, it requires finding a way to engage the people in the process of change. Let’s make it clear: people are the agents of the change, not the objects to which the change is applied. When we talk about transforming the organization, we need to remember that it doesn't exist in a vacuum, where a manager operates the “human resources”. The organization can have a sound strategy, however, it relies on people implementing it, and people need to understand and care about this kind of transformation. Without anchoring it in real value, a real feeling of achieving something, enabling something, the actual people who will be doing the trench work will not be committed.
Everyone is creative, not just creative professionals. Managers, engineers, support staff, just to name a few - can apply their creativity and find a solution to a problem which lies outside of the box, or outside of the common process of solving it. But only when people really understand their role in the bigger picture, and feel the agency - they can move the company forward, actively look for new solutions and stay engaged in the process.
How does it work in practice?
The perspective that we shared now is already something that can be used to steer your strategic work in the direction of a more human-focused, business-goal-driven transformation. But perhaps we can also provide a few suggestions and advice that builds upon our practical experience of supporting a change in the organization.
- Focus on making the individual contribution visible, understandable and embedded in the bigger picture of the transformation. Establish an “ownership” mentality, where the people expect to be able to ask questions directly, and are accustomed to keep their street clean, not just their own “home”. This can be contrasted to only caring about your piece of work, and forgetting about it as soon as it left your desk.
- Instead of top-down regulation, observe the existing patterns. Find out why they are the way they are (c.f. process mining methodology). What do successful individuals and units do to solve business problems? Try learning the policies and processes from observing what works, rather than theorising.
- Choose concrete and measurable business cases to validate the policies and practices that you plan to implement in the transformed organization.
- Embrace the “tribal knowledge” (a very accurate term we saw described by Alteryx: https://www.alteryx.com/whitepaper/optimizing-data-governance-in-the-age-of-self-service-data-analytics), as a byproduct of broken silos - as you will establish bridges between parts of the organization in an effort to bring the data together to enable the next business case, you will find that there are knowledge silos - people who just have such a deep contextual understanding that it can’t be crystallized in a single Confluence or data catalogue article. Being the only one to provide an answer to a critical question is a great source of power in the organization. Instead of breaking the boundaries, learn to build bridges, that is, design the organizational change that will make it easy to discover the data resources organization has, and put an effort into creating collaborations that allow “cross-pollination” - exchange of ideas between people and units that don’t necessarily have to continuously interact with each other. A good example here is a data council or an advisory board, a structure often suggested in data governance frameworks.
- Related to the previous point, embrace the complexity and put the quirky business rules on display. How many times have you wondered, after yet another meeting, why the hell is a certain process done the way it is? You are not the only smart person, and often, after some digging, you will discover a quirk, maybe related to legislation, or maybe to some exotic optimization, that only one or two people in the organization know and understand. Make such quirks known and appreciated, because mutual understanding that business is complex and there is a reason why expert people are needed, will make people appreciate each other's efforts more.
- We are proponents of being lean. But fast experimentation always bears a risk of mistakes, and focusing on the business values can lead to neglect of security and best practices. There is a tradeoff, and instead of excessive regulation, it might be better to focus on the “guardrails”. That means making sure that in the process of innovation, solutions are built with a thought of how to make sure that mistakes are fixable, and consequences are manageable.
This publication was prepared together by Dmitry Smirnov (Machine learning consultant at Fourkind) and Anna Bäckström (Principal consultant at Fourkind).
