In my 6 years working in the analytics industry, for most of the Fortune 10 clients, across different geographies, one thing I have realized is that while people may solve business problems using analytics, the journey is lost somewhere along the way. At the risk of sounding cliche, “Enjoy the journey, not the destination”. So, here’s my attempt at creating the problem-solving journey from what I’ve experienced/learned/failed at.
The framework for problem-solving using analytics is a 3-step process. On we go:
1.Break the Business Problem Into an Analytical Problem
Let’s start this with another cliche – ” If I had an hour to solve a problem I’d spend 55 minutes thinking about the problem and 5 minutes thinking about solutions”. This is where a lot of analysts/consultants fail. As soon as they come across a business problem, they straightaway get down to solution-ing, without even a bare attempt at understanding the problem at hand.
To tackle this, I (and my team) follow a Gap-Analysis Framework. In this, the first step is to identify the Current State of the client – where they’re currently at with the problem, followed by identifying the Desired Future State – where they want to be after the solution is provided – the insights, the behaviors driven by the insight, and, finally, the outcome driven by the behavior.
The final, and the most important step is to identify the gap that prevents the client to move from the Current State to the Desired Future State. This becomes your Analytical Problem, and thus the input for the next step.
2.Find the Analytical Solution to the Analytical Problem
Now that you have the business problem converted to an analytical problem, let us look at the data, shall we?
We will start forming hypotheses around the problem, WITHOUT BEING BIASED BY THE DATA. I can’t stress this point enough. The process of forming hypotheses should be independent of what data you have available. The correct method to this is: after forming all possible hypotheses, you should be looking at the available data, and eliminate those hypotheses for which you don’t have data. Once the hypotheses are formed, you start looking at the data, and then the usual analytical solution follows – understand the data, do some EDA, test for hypotheses and do some ML (if the problem requires it). This is the part which most analysts are good at. For example – if the problem revolves around customer churn, this is the step where you’ll go ahead with your classification modeling.
Let me remind you, the output for this step is just an analytical solution – a classification model for your customer churn problem. Most of the time, the people for whom you’re solving the problem would not be technically gifted, so they won’t understand the Confusion Matrix output of a classification model or the output of an AUC ROC curve. They want you to talk in a language they understand. This is where we take the final turn in our journey of problem-solving – the final step.
3.Convert the Analytical Solution to a Business Solution
An analytical solution is for computers; a business solution is for humans. And you’ll be dealing with humans who want to understand what your many weeks’ worth of effort has produced. You may have just created the most efficient and accurate ML model the world has ever seen, but if the final stakeholder is unable to interpret its meaning, then the whole exercise was useless.
This is where you will use all your story-boarding experience to actually tell them a story that would start from the current state of their problem to the steps you have taken for them to reach the desired future state. This is where visualization skills, dashboard creation, insight generation, creation of decks come into the picture. Again, when you create dashboards or reports, keep in mind that you’re telling a story, and not just laying down a beautiful colored chart on a Power BI or a Tableau dashboard. Each chart, each number on a report should be action-oriented, and part of a larger story.
When someone understands your story, they are most likely going to purchase another book from you. Only when you make the journey beautiful and meaningful for your fellow passengers and stakeholders, will they travel with you again.
Summing up with a pictorial representation of the 3-step process
As an Analytics Specialist at Brillio, Subham brings in 6+ years of experience in solving business problems using analytics for clients across multiple verticals and geographies. His area of interest and expertise lies in generating action driven insights and storytelling through data.