23 Jan New 2017 Insights About Some Old Ideas
Recently, we came across this excellent and timely article by Momchil Kyurkchiev CEO and a co-founder of Leanplum, about trends in managing the many new handheld internet of things (IOT) devices introduced at CES in Las Vegas earlier this month.
As these devices get increasingly sophisticated and connected to the world of big data, their influence and impact grows. We thought we’d take a crack at lending some of our ideas and experiences to the discussion.
The three trends identified for 2017 are Predictive Analytics, Chatbots and Intelligent Marketing Automation.
Although these topics are old terminology to most market research professionals, each trend has 2017 adaptations that perhaps extend beyond those mentioned in Momchil Kyurkchiev’s article.
The traditional practice of Predictive Analytics has now become real-time data science where we don’t have time to make studied strategies out of summarized statistical data. We’re now bombarded with so much rapid-fire big data that we must handle each progressive click individually and route it appropriately on the fly, in real-time. And if we don’t precisely anticipate, respond, direct and store, we may never have the opportunity to own that particular pathway that could channel us to a successful sale or lead. The adapted 1712 proverb from Joseph Addison’s play Cato was never more appropriate than right now in 2017: “One who hesitates is lost.”
This introduces the second trend, the Chatbot which is the automated version of the old call center decision tree logic using a question-and-answer mode to navigate the customer’s individual experience. What is important here is the new and unique opportunity to collect click-data and statistically compare actual results as they happen against the expected choices to learn and adjust the offering as the overall campaign progresses. With the volume of big data also comes the opportunity to design an A/B test by temporarily rerouting some of the e-traffic on an alternate path and then statistically analyzing the results of each path. This allows us to optimize sales by selecting the approach that garners the most response.
Intelligent Marketing Automation
Lastly, as is often required to support the first two trends, we have the continuing need for Intelligent Marketing Automation. While much of the underlying decision logic is still the same, in order to be responsive to the billions of real-time clicks, automation software must be replicated throughout the “cloud”. This software, therefore, must be rewritten into new, distributable languages like Hadoop, MapReduce, Python and Scala. To illustrate, a direct marketing client came to us about six years ago with a modeling challenge of scheduling their 250 call center operators for maximum peak volume during the pre-Christmas ordering period, which represented 90 percent of their annual business. We helped them by auto-routing the traffic, but the peak load hang-ups still exceeded the actual orders. Today and into the future, many call centers reside on separate servers throughout the cloud. Doing so enables processing the orders coming from handhelds anywhere. Peak volumes are becoming nearly a non-issue and there is real motivation and examples offering price incentives for online ordering.
It’s truly an exciting yet challenging time to be in the business of data gathering and analysis.