Extracting Value from Insights in “The Last Mile”
Recently, I read the McKinsey article “Breaking away: The secrets to scaling analytics.” The article discussed their findings on the importance of analytics to a company’s success.
Unsurprisingly, they found that the most successful companies spend more money and effort on analytics, specifically embedded analytics. Two thirds of what they deemed “breakaway companies” spent at least a quarter of their IT budget on analytics, with most of those planning on boosting spending in the coming years. But what is most compelling is the challenge of extracting value from insights is what McKinsey calls “the last mile”
The fundamental of analytics is to convert data to meaningful information and improve learnings from past activity. This knowledge is important to help in avoiding mistakes and/or capitalize on opportunities in the future. Just a few years ago, we saw heavy investments in Big Data strategies and technologies to be able to harness this wisdom. What this article, in my opinion, is highlighting, is the inability to deploy this wisdom into the last mile of the data driven strategy i.e. operational decision making. Even in the ones that successfully implemented their analytics strategy, i.e. 8% of the 1000 organizations involved with this study, only two-thirds were at the maturity level to operationalize these insights.
The most important part of this study that I agree with, as evidenced by us at VoltDB, is that the those that deploy their insights into the last mile operational decision-making break away from their peers and competitors. While analytics are useful, they are worthless if not acted upon. You can spend millions creating and implementing a sophisticated analytics strategy, but unless you do something with that strategy, you don’t get any return.
The solution to this problem is embedding analytics into workflows and decision making. This approach is one where analytics are fully connected to any decisions, to the point where analytics power those decisions, in an ever evolving continuous feedback loop. There are a number of uses of this operational approach to decisions, such as personalizing content and suggestions in real-time and stopping fraudulent transactions as they happen, but with the advent of 5G and the true ability to exploit IoT, along with the undeniable power that Machine Learning will deliver, the volume of applications that demand this fundamental requirement is set to explode.
While the value of embedding analytics may seem obvious, most companies are not paying enough attention to this problem. Only a fifth of non-breakaway companies are spending significant portions of their analytics budget to operationalizing their analytics. Comparing that to large majority of breakaway companies spend at least half of their analytics effort towards solving the last mile problems, it’s easy to see how powerful operational decision making is. Their efforts are well placed — companies we work with to embed real-time analytics and decisioning into their products and platforms have seen massive returns on investments.
The value of using analytics to power decisions is significant. One of the reasons why machine learning is receiving a lot of hype is that ML at its core is using analytics to make decisions. Companies that do not progress past the old methodology will be surpassed by their competition, as more companies look to overhaul their analytics and decision strategies.