5. Recommend Next Steps. This is “the” hardest part of it all, and it requires teamwork.
- Ask ‘So what?’ If you identified a problem, is there a way to fix it? Or is it even a problem? Let’s not over-prescribe medicine, as that is how analysts lose credibility.
- Form a Hypothesis. “Why” is it happening? Look beyond obvious places. If it is hard to make connections between the dots, ask the stakeholders. Again, is there anything that analysts should have known? End-users, like bad patients who often forget to provide doctors all relevant information, make bad assumptions, too.
- Always Make Multiple Suggestions, as decision-makers may reject the first idea. Be budget-conscious, as every division deals with limited resources. Do not take an easy way out or make habitual suggestions. Not every situation calls for complex statistical modeling as the next step.
- Involve All Stakeholders in the Solutioning Stage. Analysts often try to be the hero, but they shouldn’t forget that this prescriptive analytics is a team sport.
This is just a high-level summary of the five-step process to provide insights out of large reports or even raw data. Prescriptive analytics is a lot like playing jazz music where no note is pre-written, and it is something that should be taught through countless case reviews via a master-apprentice relationship. But I hope this guideline helps kickstart insights-driven analytical practices. Even if one doesn’t get to go all the way through this list, each step will be a bragging point for the analyst and a step toward better insights.
The analytical landscape is changing rapidly, and analysts – and marketing organizations to some extent — who are unable to provide insights will be replaced by machines in the near future. To remain in this analytics game, analysts must be able to raise complex questions and answer them in forms of stories and recommendations. Or, to put it differently and at the risk of sounding like a cliché, to provide insights through prescriptive analytics.
