The last mile problem in the telecom and cable industry is well documented. As the buzz around analytics, AI and machine learning reaches its peak, I am often surprised to see how the ‘last mile’ involved in the analytical processes is the most often ignored frontier within majority of the enterprises. This ‘last mile’ is the most critical path of the analytics journey when a company’s rich investments on harvesting information, analytical techniques and generating insights gets translated into tangible business outcomes based on the strategic, financial or an operational decisions made by someone in their workforce or an algorithm. The decision could be focused on targeting customers, pricing, recommending products/services, M&A, cost cutting or market entry etc. Companies spend so much of their time, energy, and dollars investing in deploying technologies, integrating data and generating insights that they forget the obvious—to connect the human chain involved in the ‘last mile’ and make it their core mission of their analytics strategies. As a result, most of the harvested information and insights run out of steam before the finish line, failing to achieve large scale adoption within the enterprise to drive business value.
I was assisting one of our large retail client define their enterprise analytics strategy and near term goals. One of the key areas was to address analytics around their store and labor performance. The client had 500+ retail store locations and their challenge was to minimize the information overload, digitize manual business processes to enable the territory managers to measure retail store and labor performance. As we started the initiative, I found myself at the client’s headquarters in a conference room with key selected stakeholders from Business and IT, discussing hypotheses and ideas for defining the problem statement. After couple of days of intense brainstorming and failing to nail the problem statement for our Minimum Viable Product (MVP), I decided to go and see the work in the field in order to observe and unearth insights for our problem statement and MVP. The client agreed and I embedded myself with a senior territory manager on a store visit. Next morning, I started my store visit with Tony who had 25+ years of tenure with the company. Tony told me that he usually starts his day at 7 am, driving 60 miles one-way from his home to make store visits in his assigned territory. He usually spends his time in his visits, listening to employees and customers to improve store operations and customer experience.
"As Data, AI, Machine Learning becomes more pervasive in our daily lives, enterprises can easily get distracted by technology innovations"
Our first stop was one of the oldest and largest stores that was in dire need of sales and traffic lift. Tony and I walked across various sections of the store where he asked the employees about foot traffic, sales targets and logistical challenges that need to be addressed to achieve their weekly targets. Tony then captured their feedback as notes in his paperback. Our next stop was one of the vendor stores located strategically near the entrance to have a conversation about daily sales and performance improvements. The vendor store manager greeted Tony warmly and swished out an iPad. I then saw Tony open his large black tote bag and pull out a white paper binder. This was my second powerful visual observation about the digital divide at play. For the next 30 minutes, I watched as Tony painstakingly turned over pages of the paper binder and keyed in numbers in his calculator as he was trying to query and answer logical questions from the vendor store manager on historical events, forecasts, and data points. After 20 minutes, Tony was visibly frustrated and ended the meeting by telling the vendor manager that he would be back in a couple of days with information to have a productive discussion. As we walked back to his vehicle, he told me in a disappointing tone—“How I am supposed to find information to do my job and be productive with these paper sheets”.
This was the one of the most poignant human moment where Tony’s words exposed the last mile of the analytics problem we were trying to solve. Over the next couple of months, the powerful visual of three large paper binders became our rallying point in the client headquarters for our problem statement to digitize the whole process of how territory managers would consume information and key business metrics, interact with data and be more productive in their daily jobs. Tony became our persona, insight and change agent who would work with the data scientists, application and user interface (UX) designers to shape the MVP but also educate his entire territory manager community to the adopt the new digital analytics solution.
As Data, AI, Machine Learning becomes more pervasive in our daily lives, enterprises can easily get distracted by technology innovations thereby not focusing on the human element involved in their analytics journey. Companies should drive an outcome driven agenda by identifying the last mile of their analytical problems, connecting and linking the human chain to their overall analytics strategies (from descriptive to predictive and prescriptive). Embedding the below 6 steps in their analytics execution lifecycle will enable enterprises to collectively combine the strengths of their workforce with data, algorithms and technologies and be successful in creating value for all of their stakeholders.
1) Understand and map the last mile of all of your analytic problems, connect the human element involved in the last mile and link them to your problem statements and business outcomes
2) Identify your persona (preferably a real-life stakeholder) and make them your insight and change agent for your analytics problem statement and Minimum Viable Product (MVP) definition
3) Go and observe the work on the field with your persona—You will gain valuable inputs and insights for your problem statements and MVPs that you won’t often find sitting in a conference room
4) As you develop the MVP, include your data scientists, application and user interface designers to work with the insight and change agent to iteratively seek feedback and improve
5) Leverage your insight and change agent to train and educate their peer community to increase adoption and buy-in of your MVP
6) Define success metrics and measure impact of your MVP, celebrate success, replicate success in other business areas