Turbocharging Analytics project adoption: It’s not about Math

Everybody is talking about tensor flow, DNN, AI, and ML. Most enterprise want a piece of this action. Some are even doing it! But there is still a gap between expected and realized ROI from analytics projects. Ironically, most analytics projects don’t fail because of inefficient neural networks …
Turbocharging Analytics project adoption: It’s not about Math

Everybody is talking about tensor flow, DNN, AI, and ML. Most enterprise want a piece of this action. Some are even doing it! But there is still a gap between expected and realized ROI from analytics projects. Ironically, most analytics projects don’t fail because of inefficient neural networks or bad machine learning models. They fail because of reasons beyond sophistication and accuracy of algorithms – the Human reason. Here’s an attempt to tackle this problem of the present as we take a leap towards the future.

Data has it that the worldwide revenues for big data and business analytics (BDA) will grow to more than $203 billion in 2020, witnessing a CAGR of 11.7%. In the same breath, a Gartner survey shows that only 30 percent of organizations have invested in big data, of which only eight percent have made it into production.

So, the question to then ask would be, “How do you plug the gap such that analytics projects don’t stop short of delivering to the potential?”

Here’s a 3-point mindset framework recommended by Shashank Dubey, Co-founder and Head of Analytics, Tredence:

  1. Business Case Mindset
  2. Entrepreneur’s Mindset
  3. Adoption Mindset

Business Case Mindset: Let us take a case in point here. Recently, we were approached by a leading e-commerce marketplace. They “wanted” us to help them reduce customer churn. But not all churn needs to be negated; what they really needed was to retain their profitable customers. Evidently, the need was different from the want. This client is not alone. “What the business wants Vs what the business needs” battle is seen in many large enterprises. As data scientists, we are forever ready to pounce at a problem and solve it to completion. We don’t pause to realize that many of these are what we call ‘fickle problems’ that emerge as a result of 5-minute brainstorming sessions.

We need a framework to filter the ‘real’ problems from the universe of ‘fickle problems’. This calls for three perspectives: Proof of Concept, Proof of Value, and Proof of Implementation.

Proof of Concept takes on the lens of a data scientist. We need to examine constraints of data, algorithm and engineering. The question being explored here is whether the problem is “solvable” given the finite resources at our disposal including time.

Proof of Value takes on the lens of a finance controller. The problem here is being evaluated from the perspective of financial ROI, with the question “Is the problem “worth” solving?”

Proof of Implementation takes on the lens of an engineer. The big question being addressed, “is the solution “implementable” on ground?” Most of us use one or two of the above lenses but rarely all 3. Ironically, we use these to determine how to execute the project and not to determine if we should execute the project in the first place! This is the first phase of the project which we call conceptualization. This is the ‘thinking’ phase.

The next phase of the project is the ‘doing’ phase. This is where the potential discovered in the thinking phase needs to become real. Here you need an entrepreneur’s mindset.

Analytics executives must double up as product entrepreneurs.

Entrepreneur’s mindset: Imagine your project is to build car for a client who has never seen one before. This client is going to drive the car you build. She is understandably anxious and needs to be involved in the build process. There are two ways you can execute (assuming you know how to build cars!): you can start from scratch or you can start with a prototype. In the first approach, you may start with showcasing a chassis and then in stages add tyres, steering mechanism, engine, etc. This is how cars get built on a shop floor. In the second approach, you start with showcasing a prototype. This prototype looks like a real car; just that none of its components work! You gradually keep on replacing the dummy components with the real ones, evolving the overall design on the go. This is how successful products (and companies) get built. We need this product entrepreneur’s mindset to execute analytics projects.

Frontline managers must be the loudest voice in analytics project conversations.

Entrepreneurs also distinguish themselves by prioritizing outcomes over tasks. They seldom land in ‘operation successful, but patient dead’ situations. Here, the outcome mindset warrants the need for effective stakeholder synergies. But before we get there we need to ask ourselves: Who is the ultimate stakeholder? In most enterprises, there are many proximate stakeholders the analytics leaders, business executives, the IT group etc. However, the ultimate stakeholder – the frontline manager – is often discounted. Ideally, your frontline managers must be the loudest voice in key conversations. In reality, they don’t even have a seat on the table in most cases! Effective synergy across the Business Analytics, IT & Leadership with frontline managers at core is the cornerstone of outcome mindset.

The next step after the ‘doing’ phase is the ‘adoption’ phase.

Adoption mindset: This last lap is critical. A shiny toy which does the job doesn’t guarantee frontline adoption. Simplicity, scale, and integration will need to triangulate here.

Simplicity must be approached from the perspective of who the end user is. Your analytics solution should make the end user decision making simpler and faster, not just more accurate. The solution must also be scalable enough to seep across frontline managers who can leverage the solution when and where they want. And finally, the solution to the analytics project must be one that can seamlessly integrate into legacy systems, with minimal resistance and orientation. Nobody wants one more App!

In today’s world, data analytics is no more a luxury. It is basic hygiene. But Analytics as such is not just the work of numbers. The human neurons will need to connect. The three mindsets – the business case, entrepreneur’s, and adoption – will need to concur and converse to convert analytics projects into profitable ROI.

Has your organization faced challenges in meeting its Analytics objectives? Share your experience with us.

And if you are an enterprise that is facing challenges in taking your analytics investment to fruition, reach out to us on info@tredence.com.