As a consultant, I spend time keeping up with the latest technologies that can assist us in the delivery of analytics-based projects. Many software vendors, both large and small, claim that using their software will automatically deliver great business value. I beg to differ. At Princeton Consultants, we always start with a business question whose answer brings value to the organization before building a model and analyzing data.
I've recently come across several people who have been echoing a similar sentiment, and I'm now hearing it more often from analytics leaders. Here are some references that support the importance of solving the right problem that delivers value:
1. In the article "Broken links: Why analytics investments have yet to pay off", ZS Associates published results from a survey conducted by The Economist Intelligence Unit, where they contacted many firms to determine why analytics projects fail. They conclude in part::
Companies have progressed on the technology side of analytics, but processes should be embedded more closely into the fabric of the business. The biggest challenges in the analytics value chain are those at the front and back end-areas where senior analytics executives interact most with the rest of the business. Solution design and change management are particularly widespread challenges.
2. I attended a recent webinar organized by the team at Anaconda, which included speakers from Anaconda and Forrester. The slides can be found here. In this talk, which addressed how best to scale data science initiatives across the organization, the speakers outlined steps for success. The first is to "Tackle projects with large, clearly defined business value." The slides have more detail about how to approach those steps.
3. Bill Schmarzo, CTO, IoT and Analytics, Hitachi Vantara, gave a great plenary talk at the recent INFORMS Conference on Business Analytics entitled "Big Data MBA – What is the Value of your Data?" He described his "Thinking like a Data Scientist" Flow, which can be found here. His 7 steps are:
- Identify Target Business Initiative
- Identify Business Stakeholders
- Identify Business Entities
- Brainstorm Data Sources
- Capture and Prioritize Analytic Use Cases
- Identify Potential Analytic Scores
- Identify Recommendation
4. Lisa LeVange, the President of the American Statistical Association (ASA), has chosen a theme for the 2018 Joint Statistical Meetings (JSM) as "#LeadWithStatistics", and she writes in her February column in the Amstat News:
One of my presidential initiatives is to establish a leadership institute at the ASA that provides resources and opportunities for members to develop leadership skills as they progress through all career stages.
The idea of this initiative came from work led by former ASA President Bob Rodriguez, who was on a panel at the 2017 JSM entitled "The Leadership Journey for Statisticians." In that panel, Bob said:
You become an emergent leader by recognizing a problem that matters to your organization, moving to the middle of the situation, and influencing others to accomplish a common goal. Statisticians who can do this are increasingly valued in today's data-driven organizations, and so we should all be prepared to become leaders.
5. I found this quote attributed to Gartner analyst Alexander Linden here:
A data scientist must possess the knack of being able to "identify business value from mathematical models." But that vital business value can only materialize if the data scientist also networks with other departments, understands their objectives, is familiar with their data and processes – and can spot the analysis options they provide.
6. From "A Beginner's Guide to the Data Science Pipeline," which describes a data-oriented pipeline for data science projects:
Before we even begin doing anything with "Data Science", we must first take into consideration what problem we're trying to solve. If you have a small problem you want to solve, then at most you'll get a small solution. If you have a BIG problem to solve, then you'll have the possibility of a BIG solution.
7. In a 2015 Forbes article, author Piyanka Jain writes: "Successful analytics start by identifying the question you're trying to answer from the data."
8. In a 2014 Analytics Magazine article, Eric King discusses how organizations are proceeding with a technological focus, rather than a strategic, goal driven approach. One of his key takeaways is:
It is infinitely more effective to select the most viable and valuable modeling project after having surveyed leadership, team members, resources and the environment than to perform great work on a doomed initiative or start sifting for insights without a performance target.
9. Finally, I was included in a panel of analytics "rock stars" asked to give advice to analytics professionals. Here are some relevant quotes from my colleagues:
- Ranga Nuggehalli, Principal Scientist, UPS
Identifying the REAL problem is always the toughest problem in practicing O.R.
- David Hunt, Senior Specialist, Oliver Wyman
My one top piece of advice to analytics professionals is to never lose sight of the business problem you are trying to solve.
- Jonathan Owen, Director of Operations Research, General Motors
Make sure you have agreement on problem framing - including what's in and out of scope - and a good understanding of how the results will be used.
- Pooja Dewan, Chief Data Scientist, BNSF Railway
We must always remember what's most important when solving a problem and never lose sight of our stakeholder's needs.
- Peter Buczkowski, Manager, Workforce Management, Disney
My one top piece of advice to analytics professionals is to dive into your client's business to make sure you are solving the right problem and that your solution can be seamlessly implemented by the area leadership.
If you need help in determining the right questions to ask so that you can gain business value from your data, contact us and we can discuss how we can assist you on your analytics journey.