One of the hot topics in the industry right now is predictive analytics. Clients are asking: “Is this a trend? Is this a paradigm? Should I be concerned about this or doing something?” The answer is: The practice of Predictive Analytics is here to stay and is important for the engineering and construction industry.
Predictive analytics is the next generation of analytics. Data makes up the building blocks for analytics. But instead of using data analytics to make sense of the past — to understand what we’ve done and what trends we can pick out from our past behavior — predictive analytics uses data to predict what is likely to happen.
Predictive analytics is driven by what you don't know. When you use predictive analytics, you let the data tell you the answers that you may not be aware of; that are not obvious to you. You’re looking at data you may not be collecting yourself. You are searching for outliers. You're trying to find patterns and relationships. Companies typically use predictive analytics for three reasons: to manage risk, improve return on investment; and increase yield.
The following example demonstrates how predictive analytics can be used to increase yield. For the U.S. retailer Target, their yield comes from customers that buy consumer goods. A couple years ago, a guy in Minneapolis went into the local Target store and talked to the manager. He had a stack of coupons (and these are for maternity clothes and bibs and diapers), and he said, "You sent these to my16-year old daughter. Are you just encouraging her to get pregnant?” And the manager said "Hey, I don't know, that all comes from corporate. Let me see what I can find out and I'll get back to you." The manager investigated, and he called the parent a week later and said, “I haven't found anything out. I'm just checking with you to give you an update of where things are.” And the father responded, “I need to apologize to you. In fact, my 16 year old daughter is pregnant, and Target knew it before I did.”
Target has a team of data scientists parsing through consumer purchasing behavior data. Pregnant women are good customers and the team searched for patterns and relationships associated with that group of women so they could maximize sales. What they found was that women that buyunscented lotion and zinc and magnesium as prenatal supplements were almostassuredly pregnant. If you have that buying pattern, Target can predict your pregnancy. Based upon your habits, they can predict within a 90% probability when you are due. And of course if you buy a blue bath mat versus a pink one, they know if it is a boy or a girl. They’ve even learned how to prevent the type of reaction the father had. They send you maternity and prenatal coupons mixed in with coupons for wine glasses and lawnmowers to reduce the sense you are being targeted.
Target is pouring over data and finding the themes, patterns and relationships to build predictive models to achieve business outcomes. Using the data and a focused marketing approach, over a short period of time, sales to pregnant women grew from $44 billion to $65 billion. It’s pretty amazing what people are doing.
Across all industries, companies are looking for analytics skills. By some estimates, sixty-three percent are not familiar with “Big Data” or “predictive analytics.” According to Gartner, sixty-eight percent lack a strategy around predictive analytics and business intelligence. Consider your company and those statistics. If you have the people, the strategy and understand the concepts and applications, you are ahead of the pack.
The question for the E&C industry is: “Are we making data-driven decisions in our business?” I truly believe decisions with data and analytics will be the cornerstone of our business going forward. How do we know? Well, we can look at other industries that have been transformed and see parallels. Many other asset builders have transformed industries and generated completely new revenue streams by applying data analytics and using technologies. They have engineering department; they build assets.
The auto industry is an example. Automobile companies are asset builders. General Motors (GM) On-Star division, which is driven by technology and analytics, made $1.5 billion in 2012 with a 35 percent grossprofit margin, nearly five times the 6.2 percent margin from GM's first-quarter adjusted earnings before interest and taxes in North America, according to Automotive News.
Thirty years ago cars never had a level of service. I can remember timing the carburetor for midwestern cold winters, gapping spark plugs and all kinds of other manual maintenance. You don't do that anymore. The automobile companies drove technology into the car. They built new revenue streams first creating service departments in their retail shops. Then, when they sent logs back to their engineering department, the service departments provided valuable asset feedback. The companies saw an opportunity. “If we put monitoring technologies into the cars, they could tell us right away. We can engineer a better asset immediately, and then we can guarantee a level of service for the asset”.
In the past, cars didn't have warranties like they do today. The builders of the assets found a means to influence the levels of service for the performance of their assets for longer periods of time. And not just in the building of the core asset business anymore.
We already collect a lot of data in our industry but we need more. To increase data assets and the ability to build useful models, data collection must to become part of everyone’s job. At MWH, we put the iPad in the field to collect data. We make punch lists and forms easy to use and readily available on different devices. We codifyprocesses in software to use data, to ensure data quality, and to test the applicability of analysis. And we’ve developed the approach and software to the point where it is useful and useable so that our partners and clients can buy it and build their own data assets. With solid data foundations, many in the Engineering & Construction industry will begin to build and leverage predictive analytics.
My hypothesis is (and we're seeing this in other parts of the world): If you’re caught flat-footed in the E&C industry and you're not looking at smart ways to use data analytics and predictive models, companies that are moving quickly will take your market share. Companies will lose if they can't envision new data-driven, revenue streams outside of our traditional set of offerings.
Dan Kieny is vice president and global director of consulting and knowledge at MWH Global, a strategic consulting, technical engineering, environmental and construction services firm with nearly 8,000 employees in 35 countries. This First Read viewpoint was inspired by his comments last month at the ENR FutureTech East event in New York.