Today we have many, many sources of big data to help support our business decisions. Adopting such analytics tools usually involves a journey towards a more structured decision process. One common misunderstanding that we see is the idea that predictive models and rich analytics eliminate the need for intuitive, thoughtful decision making. Another is that everything must be perfect in your systems from a data perspective in order to start. Neither is the the case!
Analytics tools don’t replace intuition. Instead they refine intuition.
The most valuable systems embed business rules and strategies into the analytics framework to make credible recommendations, to highlight opportunities, and to automate the innumerable smaller decisions necessary to keep strategy and execution aligned. This frees experts to focus on the “what-if” explorations of big picture moves that will make a difference in the market.
The journey through such a transformation can involve many steps, and it is possible to start unlocking business value relatively early in the process. This happens best when you plan for a phased evolution, where each step builds on the data and systems you have at that time. As the tagline says, don’t let the present keep you from the future!
Gartner frames the analytics-driven process in four stages, with prescriptive analytics as the most valuable of the four. Focusing on “how can we make it happen?” question too early, though, is putting the cart before the horse. Getting to prescriptive stage should be the end goal because it is there that the data is working the hardest for you. Just remember that you can’t get to the “how” before the “what”.
At Revionics, we help retailers make better decisions on pricing, promotions, and placement. Our systems drive insights from big data, providing the predictive analytics that help model likely outcomes, and the prescriptive analytics that help retailers map the best ways to achieve a given outcome. We’ve driven scores of successful decision process transformations with retailers, across sectors and at different stages in process development. Here are some key starting points and questions that I have learned from participating in many of these:
Understand (and communicate) the consequences of inaction
- How does current performance compared to goals?
- How does current performance compare to world-class execution and to key competitors?
- Are the trends positive or negative? Why?
- What is likely to happen if we do nothing?
If you always do what you’ve always done, you’ll always get what you’ve always got. — Henry Ford
Establish both feasibility and opportunity from beginning
Working with a partner or solution provider? How does the team map a process for:
- the business benefits in objective, measurable terms, along with the risks, phased roadmap and effort required to achieve?
- honesty and clarity about likely initial results along with current limitations and reasoned plans for improving with better data and directed innovation?
- transparent results from blinded “back-casting” or “hind-casting” using your historical data?
- a results measurement framework that validates success and communicates in terms all agree are meaningful to business?
Know what defines success in near-term and long-term
No one knows your business better than you and your colleagues. It is critical to spend time up-front thinking about what you are trying to achieve and how you will measure progress!
- What changes in the organization’s primary leading indicators (units, profit, share, etc.) would indicate success over a 3-5 year period?
- What initial progress over 3-6 months would establish credibility on achieving those goals?
- What signs or indicators would indicate the need for a course correction or even a timeout?
Keep imperfect data from perpetuating inaction
Don’t wait until all the data is clean and complete before you start! Instead, focus on what you can achieve and let the benefits help pay for better data!
- What parts of the needed data must be right to begin implementation? How about for GO-LIVE?
- What parts of the data stream can be fixed, cleaned and/or supplemented with minimal impact to the organization’s existing processes?
- What new or refined data would drive the biggest impact on success for first phase of project?
There is a myth in these types of decision support or decision automation programs that data must be perfect in order to begin. The reality is different – some data must be right, and you need to know which data that is. No one has perfect data, though, and worrying about this can create excuses for inaction.
If you or your organization operates at a scale that requires thousands of decisions or more each month (as most multi-site or multi-channel retailers do), a transformation creating a structured, analytics-driven decision framework can be the difference between defining the market and being defined by the competition. Start small and let the benefits pay for more innovation!
Perfection is the enemy of progress!
In my next post on this series, I’ll cover some strategies for ensuring a successful decision transformation. In part three, I’ll discuss the specific benefits of and cautions for full decision automation.
For additional articles on hot retail topics, I invite you to visit the rest of the Revionics blogs. Have a great day!