Tackling the Big Data Challenge

Any investments in data and analytics must prove valuable to both the field and the office.

Data is the currency of the 21st century.” If you believe this perspective, then your engineering or construction organization may be the mint. With all of your organization’s internal and external interactions, transactions and activities, you may very well be swimming in a sea of data as valuable as Fort Knox.

In the prior two articles, we focused largely on the “why” and “what” of the changing environment. Here we will focus more on the “how.” We would like you to walk away with a better understanding of how you will build your data and analytics mint.

You Are a Construction Professional — Not IT. Where Do You Start?
Whether you are a seasoned veteran or a millennial (or maybe Gen Z) looking to bring better data and analytics to your organization, we need to agree on a few fundamental principles before we extend this discussion.

For starters, you belong to an engineering or construction company. You are not a part of an IT or accounting company. While improved data and analytics could be the steroids that give you a competitive advantage, your organization’s full-time mission is not data and analytics. Therefore, your investments of time, money and resources must be calculated and weighed against other investments in your business. And you only have so many chances to get it right before you lose momentum. This does not change the imperative to learn and adopt better data analytics, but there are various challenges that dictate how you spend resources to operate your business.

One stumbling block lies in the provenance of the data itself. As an engineering or construction company, you most likely have a mixture of premise-based systems (applications that run on-site) as well as SAAS (software as a service). Moawia Abdelkarim from DPR Construction puts his finger on the problem, noting: “Just getting to the data is a huge problem, not to mention different naming conventions for the same thing as it appears in these disparate systems.” He recommends significant planning and organizing of the company’s data. This can be done in-house, but getting outside support for the early days of data structuring is likely worth the cost. Debbie Pyykkonen is an independent data analytics professional who works in construction-related industries as well as retail and medical sectors. She recounts a number of projects where she has been called in after the fact to “clean up” the underlying data structure. “It’s good work for me, but I’d prefer to be involved with the client earlier in the process rather than having to do all this rework.” Interestingly, Pyykkonen is exploring predictive data methods to actually help with the cleanup of the data itself. “Why not use the same ‘fuzzy logic’ tools to ‘scrub’ the data?” she asks.

Secondly, the whole discussion around data and analytics is more about creating culture and behavioral changes than it is about creating new systems. Therefore, when implementing systems, you must focus on the right level and on the right indicators that are going to drive behavioral change. Focusing purely on “lagging” indicators or on the “rearview mirror” perspective does not change results as much as focusing on “leading” or predictive measures. Likewise, the inability to narrow the focus of your data and analytics strategy can and will dilute the behaviors that you are trying to drive. Adbelkarim observes that the construction industry, even when promoting dashboards and visualization technology, is still tied to these lagging indicators: “At DPR we are striving to develop metrics and data that help prevent problems. We’d like data alarms to be going off to protect the company before the problem occurs.” It is very challenging to get management and staff to agree on upstream measurements and hold themselves accountable without “poking holes” in the chosen measurements; this must change.

Finally, the “creators” and “consumers” of data are typically different in your organization. In other words, those who generate data (typically field or operational staff) are a different set of individuals than those that must actually use it and make decisions from it (typically office-based staff and often the decision-makers or heavy influencers). This is especially important in an industry that is challenged in integrating field-office cultures and ensuring that both work at optimal productivity. Therefore, any investments in data and analytics must prove valuable to both the field and the office. FMI discussed this dynamic with Marc Krichman of Lantern Data Systems, who states that “given workforce dynamics and generational expectations, we must continue arming the individuals that are closest to the project and who can make their own decisions without direction coming ‘top-down’ — after all, they are the ones doing the real work and they have the most impact on quality and customer satisfaction.” In other words, the “creators” in the field are also now becoming the “consumers” — creating a higher imperative to involve the field in the selection and organization of the data, building intrinsic support for data collection and application.

Even when predictive analytics can help drive material savings and create better volume purchasing, for example, the fact is not every company has a culture that allows that savings to be easily propagated. De-centralized decision-making epitomized by a great project manager executing work in the field, for instance, may not support the data results, thus eroding its accuracy and effectiveness even further. To meet this challenge, it’s important to consider a virtuous circle, of which supply-chain management organizations are a great example, having mastered quantifiable reductions in their cycles now for decades. IT data “consumers” might create demand planning or inventory reorder models, but field-level “creators” must be given time to understand and see clear evidence of success before becoming supporters of new data analytics.

Self-Assessment: Where Is Your Company on the Analytics Spectrum ?
If you want to improve your data and analytics game, you are probably trying to solve a challenge (or set of challenges) at one of three levels. As a company, you can measure your analytical efficacy as a progression from level one to level three, as follows:

• Level one — One-off problem-solving: Perhaps you are a project manager, accountant or IT professional who needs a better way to capture project progress, budgets or quality. You are motivated by your personal or project needs, and while your requirements are immediate and local, you can understand the implications to overall organizational efficiency when you see that 10 other project teams are working independently to solve the same problem.
• Level two — Business unit or functional problem-solving: You may be the head of the accounting or project management department, and you are looking for better ways for your team to capture costs or progress. Again, you can understand the long-term implications if you are the head of project management and the accounting group is taking another approach to solve the same problem (or vice versa, if you are in accounting and project management is taking its own approach).
• Level three — Enterprise-level problem-solving: This combines the
  top-down and bottom-up approach to finding solutions. Before we go on, you may have just had an allergic reaction to the word “enterprise,” which may evoke bad memories of an “Enterprise Resource Planning” (ERP) system that was an expensive multiyear, painful exercise. Here is the good news: While enterprise-level solutions may have once meant “ERP,” the world is now flat! You now have the ability to harness data and analytics, bringing in various data points to create intelligence, without the need for those multiyear painstaking efforts that often end up in a divorce and missing limbs. If college students are disrupting entire industries with businesses that they started from their dorm rooms, then surely your firm has access to the right internal and external resources.

We are beginning to see clients who are no longer investing in significant ERP and functional upgrades (such as the newest PM system) and are instead focusing their efforts on analytics to improve decision support for their managers. How much new under the sun is there in ERP, PM systems, imaging and workflow systems? Is it really worth all the money and risk to upgrade, or would it be better to harness that data for consumption by your business leaders?

Recalibrate the GPS
It is interesting to think where the industry is now compared to where it was 10 years ago. We often hear about drones, 3-D technology and virtual collaboration that help us work “in” the business and build projects. There are also some fascinating trends in how we are working “on” the business — both in the boardroom and in the project trailer.

Exhibit 1 shows a road map of the progress we have made in the past several years and some foresight into what may come in the construction industry.

So What Are Others Doing?
There are three obvious areas that construction and engineering companies have started tackling: safety, materials management and pricing algorithms. DPR’s Abdelkarim supports the idea that safety data is a good place to start because it’s been tracked at most companies for many years, the data is usually consistent and “clean,” and it’s important to the success of the company. But “ironically, since we already have a number of programs that have significantly improved outcomes over the years, safety might not be the best place to derive ROI from analytics.” Both Abdelkarim and Pyykkonen echo the idea that materials management has many years of success in distribution, and other companies and are intrigued by the idea of looking at pricing activities with more big data techniques. “However, we’re not the corner gas station, changing our prices every day, driven by supply/demand curves,” says Abdelkarim.

Many construction companies are still thinking of analytics in the old “reporting” mode. We get requests from clients to “rewrite their old reports” in a newer report-writing technology. This accomplishes little more than replicating past information flows and mostly misses the mark of newer visualization technology and totally misses the mark of creating more meaningful and math-driven analytical data. Probably the biggest failure of this approach is that it avoids making meaningful analysis of business drivers — what is important data? What really makes money? What really loses money?

Sometimes the technology even gets in the way. Rolling out a “content management” or “portal” technology such as Sharepoint or even new systems that manage processes such as fleet or tool management are great for the business but do not necessarily create good discussion around analytics. There may be decent but basic reporting about utilization, for example, but no predictive data produced about upcoming utilization. The most valuable economic decisions are about the future, not the past.

Lastly, and especially during this relatively profitable part of the economic cycle, we still see companies missing data in the aggregate that can enhance profitability. For example, excess materials at the end of a job are still often poorly handled. There are lots of job-level assumptions made about handling and transportation costs of these materials. Warehouses and yards get cluttered up with duplicate, unused and difficult-to-dispose-of materials. With better visibility and analysis of these materials — indeed better demand and consumption predictive tools used in the first place — smart management of these materials can produce profit for the contractor. We know a contractor that spends $2 million/year on safety glasses alone with zero analytics in place to manage and control that demand. It’s predictable, it’s manageable. And managing these costs creates value and profits for the contractor.

Test Your Team, Scale Up to Success and Deploy Into Your Organization
At the end of the day, the fact that construction and engineering companies have succeeded for so long without big data is not a winning rationalization. Even if companies are not convinced that safety data, supply chain analytics and pricing strategies — our recommended focus areas — are worth investing in, at least remember that not doing so puts your company squarely outside the norm. Admittedly, this “everyone else is doing it” reasoning is a poor substitute for strategic thinking, but at a very basic level construction companies must compete for talent from other industries and must respond to customers who are demanding data savvy service providers. As the economy shifts to successful service and technology companies such as Tesla, Apple and Amazon (who are building large facilities all over the world), data-focused construction and engineering firms will be more appealing to work with as vendors and business partners. Find ways to get there.

Test your team. Is it transitioning toward big data or do you need to lead it there? Here are a few simple questions to answer about your team:

Five signs your team is not on the right track with data analytics:

• It still conducts most of its data analysis from printed reports.
• All of your business Key Performance Indicators are backward-looking in time.
• Your key data sets take hours to “run” on the “system”.
• You still have to ask an IT person for business data more than 50% of the time.
• Your analytics tools are the same ones used for more than five years.

Five signs your team is on the right track with data analytics:

• There is a project underway to clean up and consolidate the data.
• Younger staff are bringing in new tools to try to produce better analytics.
• IT is collaborating with field staff to develop new analytics.
• Data is quickly available for management meetings.
• Your organization is getting outside help with these new tools.

Once you get a sense for where your team is, try a scaled-up approach. We suggest using the following four levels, as demonstrated in Exhibit 2:

• Tier One: Operational Efficiency and Effectiveness — These are critical initiatives that enhance project performance. They typically involve collaboration, better tracking of changes and costs, quality and other performance factors. While these are important and are becoming “table stakes” to operate in your business, realize that they are the innovations that are the most easily replicated by your competitors and, as such, they offer the least long-term competitive advantage.
• Tier Two: Organizational and Individual Learning — Initiatives in
  Tier Two involve improvements to how the organization and individuals thrive together, give and receive feedback, and change behaviors or processes accordingly. Think of Uber’s real-time feedback as it might apply to the construction industry. It can also involve workforce analytics to better define areas of improvement or deficiencies in skill sets. This tier is increasingly important in a talent-starved industry. Data here evolves more quickly and creates ongoing value. It begins to become more predictive in nature.
• Tier Three: Market Insight and Customer Interactions — The ways in which you acquire intelligence about the market and your target customers are changing. Data and analytics can improve your ability to track customer behavior, acquire feedback and make strategic decisions accordingly. This is largely implemented through Customer Relationship Management (CRM) software but is growing wider to include better monitoring around market trends, changing supply/demand dynamics and competitor positioning. Do not forget that the most important
  data is related to your customers — not simply introspective and retrospective. And the more predictive and future-oriented you want it to be, the more challenging it is to develop. Be prepared to enlist help with the math and statistics involved with this initiative.
• Tier Four: Profit Model — By far, this involves the most progressive and longer-term strategies around data and analytics. The more progressive items in profit model innovations that involve activities such as: data and analytics include integrating predictive risk assessments with self-insurance programs in order to maximize risk management profit centers, capturing historical costs and productions in order to have best-in-class preconstruction predictability, or even finding ways of harnessing data and trends that may have high value that you can sell to your clients (or provide to your customers as a business development relationship tactic). Your managers will need to be convinced that these new profitability drivers are for real before they will provide enthusiastic use and support.

The point of understanding the four levels of innovation and how data and analytics can drive progress is to push you to think critically about which areas of your business need the most attention, which areas are in need of innovation, and just how quickly you can scale data to drive profitability.

At the end of the day, you will probably need to re-orchestrate some of the organizational structure around data analytics. This requires management commitment, enlightened collaboration among business-side data mavens, input from IT data managers and the creativity of more academically and statically oriented mathematicians. None of this happens without tools and methods to mine, clean and manipulate data in accurate, deep, understandable ways.

The good news is that the raw elements for big data-focused business strategies can flow from pre-existing data sets and tools already in use by many construction and engineering companies. We hope that our article series has helped you frame some ideas and concepts for your entry into the inevitable and increasingly ubiquitous world of big data. Your organization may not yet be headed in a data-driven direction — perhaps distracted by other technologies, business priorities or insecurity about what “big data” is supposed to mean — but some basic steps can be taken. Look at your organization: are your people trying to use data in new ways? Are you using more than averages, totals and comparisons? Perhaps there are some simple places to start, where you have good data. Clean data. Lots of data. Many software packages are now building in more typical predictive analytics for ordering, managing material demand and other high-volume transactional areas. Construction companies have not traditionally thought of themselves as managing these types of transactions and this type of data. But it’s there in every construction company and ready for exploration and harnessing for value and profit. Q

Mike Dow is vice president of consulting with Tilson Technology. He can be reached at 720.370.5381 or via email at mdow@tilsontech.com. Jeremy Brown is a consultant with FMI Corporation. He can be reached at 303.398.7205 or via email at jbrown@fminet.com.

The post Tackling the Big Data Challenge appeared first on FMI Quarterly.