Business Impact of Predictive Analytics
This week Bucharest played host to Internet & Mobile World, where, as well as speaking about Social Media and Big Data, I joined a panel on the business impact of Predictive Analytics. The session was chaired by journalist and owner of lastiri.ro, Lucian Mindruta, who did a fine job of keeping myself, Edward Weatherall (VisualDNA), Francesco D’Orazio (Face/Pulsar) and Steve Keil (MammothDb) in order and handling the many questions from the audience. What did we learn? The discussion ran for over an hour, so these are the three highlights that stood out for me:
How do we predict customer needs? Is there a future for focus groups, or will research by social media replace them? Many organisations make poor use of focus groups. They aren’t there to design your next product or offering for you, they are to help gain customer insight. Predictive analytics, driven by big data and on-line versions of traditional data gathering, like surveys and behavioural studies, provide a very cost effective alternative. While they don’t deliver exactly the same insights as a focus group, when they are combined with interaction via social media, they come very close. Close enough, I believe, that they will definitely threaten the future of expensive focus group research.
Do we trust the data or do we trust our gut? A very controversial question, but one that isn’t really the dichotomy that it seems. How do we balance between data driven and intuition lead decision making? By involving both. Both have benefits and limitations. We need to understand that our interpretation of data is subjective; even though it might not feel like it, all decisions are to some extent gut driven. That means we need to ‘program’ our sense of intuition well. We ‘prime the machines’ in the same way that we train predictive analytics, by using the right inputs. Feed your intuition by spending time with end customers, and gathering data about what is and is not working.
Innovate or Increment? The balance of kaizen-like continuous improvement versus radical innovation is one of the hardest things in business. When do you focus on small incremental changes, where analytics are critical for tracking optimisation, and when do you do things radically differently, where analytics are used to target change? I’d argue that you need to plan “innovation experiments” into your programs – try something radically different every so often, and see what learning and opportunity it throws up. Then feed that into how you optimise your business.
Lastly, Lucian made a key point about the dangers of over optimising business processes: be careful. As I ‘ve said may times, a system that is 100% optimal and efficient ends up as 0% resilient, unless resiliency is built into the criteria for efficiency. If we focus exclusively on the analytics, we can fail to spot dangerous change that is outside of our focus. Predictive Analytics are an additional tool for business decision making, not a replacement for good leadership and management.