In my previous posts in this series about predictive analytics in healthcare – an overview on why the time is right for healthcare organizations to implement an advanced analytics program as a core capability, and the first stage of the startup framework, Formation – we began the journey to show you how predictive analytics can be an inflection point in an organizations' journey to become data-driven.
In this post, we look at the Validation stage. This is where the fun really starts as the first modeling work gets underway. This stage is often characterized by rapid iteration, exploration, and lots of trial and error. As the Startupcommons framework states - this is the "Lean Startup" phase.
The three primary goals of the Validation stage are:
- Achieve product/market fit
- Evaluate organizational commitment
- Execute the transition to the growth stage
Goal 1: Establishing Product/Market Fit
The product/market fit concept is attributed to Andy Rachleff, current CEO of Wealthfront. The concept is used by many start-ups as a simple framework to understand if the product that they are developing addresses a compelling and unmet need (or even better, multiple needs) in a large market. As an intrapreneur, the market analysis is a quantitative evaluation of what the impact of predictive analytics can have on the organization. This is a step that is best done in the formation stage.
The idea of finding a compelling and unmet need is the critical consideration at this stage. In the setting of an internal analytics program, this takes the form of finding the compelling and valuable use cases that if solved, will force the organization to take action. As a new organization, it may make sense to pick a smaller, faster “quick-win” to prove initial value. Identifying your first use cases may be an informal process, but as the program matures you’ll want an established method for mining new use cases from the organization.
Goal 2: Evaluate Organizational Commitment
The second goal is evaluating organizational commitment to the program. This is about changing the organization to embed advanced analytics as a new way of being. Things to consider are the ongoing participation of management champions, follow through on human and financial resource commitments, and gradual inclusion of the program into key internal and external messaging. You can review this white paper from Intel to take a deeper dive into this topic.
A true test for this goal can come when initial modeling attempts prove unsuccessful. Challenges might be met by calls to change course or prematurely abandon the program. Sometimes this can be minimized by setting clear expectations in the formation stage about the iterative process of building predictive models. Carefully control the flow of information out of the project team to manage expectations. While it might be enticing to overemphasize a critical break-through to stir up excitement, project teams can put themselves in a position of needing to feed the organizations addiction for positive soundbites, which again, can distract from the work at hand. At the end of the day, it’s the results that matter, and teams that stay focused get results faster.
Goal 3: Validation to Growth
The third goal, transitioning from validation to growth, is a critical inflection point for advanced analytics programs. If all has gone to plan, a model or models have been developed that show great promise. Management is happy with the results, and the team is feeling encouraged that the investment to date seems to be paying off. This can create a short-term high that can be quickly grounded as a realization of what lies ahead sets in. As a runner, I like to compare it to completing the second lap in a 1500 meter race. Your initial adrenaline is spent and the reality of the work ahead starts to set in. Keeping the pace takes more focus as breathing begins getting strained, and you mentally get ready to grind out the next 600 meters until the finish line comes into view.
This is the time that teams need to roll up their sleeves and begin business process change management in earnest. Re-writing procedures, training teams, developing new applications and integrations into clinical systems, and re-evaluating infrastructure for scalability. With proper planning in the formation stage, these tasks should be well-scoped and resources ready to execute. Organizations that don't do an adequate job of scoping the business process change management required will find that they stall, and possibly fail, as the momentum fades when trying to move from proof of concept to production. Done right, this is when the advanced analytics program matures into an organizational capability, at scale. This is growth.