Page 18 - AeM_Aug_2018
P. 18
BEST PRACTICES & STRATEGIES
Making real-time operational
decisions in the blink of an eye
Simply having predictive models that suggest what might algorithms to identify potential fraud, websites
be done won’t be enough to stay ahead of the customize content in real time, and airlines
competition, Bill Frank states in his article “Let algorithms automatically determine how to re-route passengers
decide – and act – for your company”, part of the when weather delays strike while taking into account
Harvard Business Review “Predictive Analytics in myriad factors and constraints. All of these analytics
Practice”. happen rapidly and without human intervention. Of
course, the analytics processes had to be designed,
“Instead, smart organizations are driving analytics to an developed, tested, and deployed by people. But, once
even deeper level within business processes—to make they are turned on, the algorithms take control and
real-time operational decisions, on a daily basis. These drive actions”, Bill Frank explains, adding that “the
operational analytics are embedded, prescriptive, power and impact of embedded, automated,
automated, and run at scale to directly drive business operational analytics is only starting to be realized, as
decisions. They not only predict what the next best are the challenges that organizations will face as they
action is, but also cause the action to happen without evolve and implement such processes”.
human intervention”, he continues, besides emphasizing
that the same evolution the manufacturing went through “Just as it is still necessary to design, prototype, and
during the industrial revolution is actually happening in test a new product before an assembly line can
analytics. produce the item at scale, so it is still necessary to
design, prototype, and test an analytics process before
If predictive analytics have been a pretty cost-intensive it can be made operational”, he clarifies. Although
customized endeavor in the past, operational analytics many most probably won’t be comfortable at first “with
recognize by now the need to deploy predictive analytics the prospect of turning over daily decisions to a bunch
more broadly, but at a different price point. By creating of algorithms”, it makes sense doing it. The fact is that
an automated process that builds a reasonable model for predictive analytics applied in batch to only high value
hundreds or thousands of products or offers, rather than problems will no longer suffice to stay ahead of the
just the most common ones, predictive analytics can competition, rather it is necessary to evolve to
impact the business more deeply. operational analytics processes that are embedded,
“Operational analytics are already part of our lives today, automated, and prescriptive. ◊
whether we realize it or not. Banks run automated
By MediaBUZZ
18 August 2018 - Predictive Algorithms & Native Ads