Data is an incredibly valuable strategic asset. According to Gartner, by 2021, AI augmentation will generate $2.9 trillion in business value and recover 6.2 billion hours of worker productivity. Data unlocks value because it is itself useful and because it frees up time and resources for other useful activities. Harnessing the power of data is the future of construction. To do so, it is more important that all members of a team be more data aware and data literate than technical data experts.

What do data awareness and data literacy mean? There is both a cultural and technical component:

• Cultural. Construction teams must genuinely seek to verify their decisions and test their assumptions with data. The team should, as a matter of course, reject hunches and “guesstimations.” Data need not be taken as gospel – experience and firsthand knowledge has vital roles to play – but it should always be consulted. A team with this attitude will outperform a team with a crack data scientist but a weaker commitment to leveraging data nearly every time.
• Technical. The team, in aggregate, must have the technical skills to push and pull information from databases, use dashboarding and data visualization software, and code and develop in-house analytic tools, as necessary. For a team without a current data focus, if the goal is to perform sophisticated analytics such as machine learning, more technical personnel should be added. If the focus is instead consuming analytics provided by others, some basic upskilling and data-literacy training is more than sufficient.
• Leaning into data need not be scary and technical. There is plenty of value to be had simply by using basic analytics and metrics in the right ways. Jumping from zero to advanced machine learning is a fool’s errand. It is much better (and more feasible) to spend time on the intermediate steps – data pipelines, dashboarding and simple analytics. Once the groundwork is set, one can bound into advanced machine learning.

Job roles in a data-informed environment

Transitioning to a data-informed process doesn’t mean that more traditional roles will become obsolete. Every task that must be completed on a construction site remains. The difference, however, is that the decisions being made about each task will be more informed. Traditional roles will pivot to become more data-centric. Slightly more of the day will be spent accessing, viewing and analyzing data. Less time will be devoted to paper pushing and dealing with rework and avoidable errors.

Leveraging data is not at all the first step to making construction jobs redundant. In fact, quite the opposite is true. It allows an individual to take advantage of all the information his or her job generates and funnel that learned experience into future decisions. Data makes a given construction job more valuable because it can be done better.

Avoiding common missteps with data integration

Surprisingly, the biggest missteps aren’t technical in nature. There are great technical people to be hired and most organizations have many employees who could easily be trained to handle a larger data-focused responsibility. The technical skills are more than within reach.

The most common mistake is asking data to do too much when the organization does not yet have the appropriate infrastructure. Databases must be set up, integrations must be complete and the data must live in a central location. Without these fundamentals, analytics efforts start at a significant disadvantage. The great risk here is not that the data won’t be useful, but that the organization won’t find the quick win they were looking for and will then lose faith in the value of data as a strategic asset going forward. In these cases, one can quickly lose the battle and the war.

The second biggest misstep is thinking too small. Many construction and project managers recognize the importance of data. They go a significant way to making their teams more data-focused, succeed in this endeavor, generate real value and then stop. Surprisingly often, an organization will halt all data progress once dashboards are up and working well. This kind of business intelligence can sometimes be seen as the end of the data journey. This is not at all true. There is a whole world of machine learning and statistical analysis out there that can offer more accurate, specific and actionable insights (and value) than even the best and most useful dashboards. Automated alerts, real time exception identification, highly accurate risk forecasting – it’s all out there, but it is beyond the capability of dashboards and basic KPIs.

Investment vs. ROI – choosing the right functions for the team

When the goal is increasing data literacy or basic upskilling, the time investment can be as little as a day or so of training for key team members and a few organizational change management check-ins to ensure that adoption of any new approaches remains high. A complete data overhaul, however, requires a more significant outlay. Moving towards a fully integrated machine learning approach takes time. Going from zero to data-centric is typically takes a year. To move from data-centric to machine learning maestros often takes another year.

Organizations that have implemented a data-focused approach typically see cost reductions in the range of 10 to 15% (sometimes 20% or more with a sophisticated machine learning solution). It’s common to see a similar drop in schedule delays and rework. Estimates often become more precise, which kicks off a virtuous cycle of less budget overage (and fewer nasty shocks when bids come in) as well as better project performance. The effect of more accurate estimates alone can result in hundreds of thousands of dollars in savings each year.

The bottom line

The data-centric construction team stops to ask the question “what do the numbers say?” By consulting and reviewing the available data in a serious, sophisticated way (rather than just glancing at a figure in a spreadsheet and moving on), the team makes a decision equipped with all relevant information. The team won’t always follow exactly what the data prescribes -- compelling reasons may exist not to -- but an attempt will have been made to quantify their decisions. In a nutshell, the data-centric team is constantly justifying their decisions in a quantitative way. Biases and bad habits tend not to stick around when this is the norm.