"Data is the new gold" is the famous quote of the American billionaire Mark Cuban. In the past decades, many trading companies but also banks and insurance companies jumped on the Big Data bandwagon. The construction industry seems to be missing out, even though it would be ideal for the use of Big Data in many respects.

The operation of a construction company generates an enormous amount of data, which is often collected in folders or in different systems, but is hardly ever used.

Without claiming to be exhaustive, these can be divided into the following categories:

• Customer data; e.g., public vs. private, nationality or domicile, corporate vs. small business vs. private
• Geodata; location of the construction site, can be assigned to a municipality or a country, for example, for further evaluations
• Weather data; number of bad weather days
• Personnel deployment data; e.g. site manager, site clerk, foreman, technician, site manager's supervisor, site clerk's supervisor
• Data on equipment used; e.g. make, model, running time
• Subcontractor data; broken down by trades
• Contract data; e.g. size, can be assigned to a size class for further analysis, contract type (general contractor vs. individual trades), claims/extra cost claims

Through structured recording and regression on target variables such as the construction site result, numerous interesting findings can be gained from this or a lot of gut feeling can be confirmed.

Again only exemplarily some statements which could result from this as well as possible explanation approaches:

• Certain site managers or certain construction managers perform better with certain types of customers, because they speak the language of the customer, e.g., due to experiences in their past
• Certain equipment performs better in certain regions, e.g., because it is better suited to the soil conditions
• Certain teams are more proficient with certain types of equipment because they are better trained on them or have more experience
• Certain site managers or certain construction managers perform better on certain types of jobs (e.g., general contractor jobs) because they set their management priorities accordingly.
• Certain types of orders or customers lead to better results in certain regions, e.g. because the people involved are known.

Of course, correlations always have to be questioned critically, as they may be based on third party correlations or simply be spurious correlations. Nevertheless, the potential is enormous, especially for somewhat larger construction companies with a broad range of offers.

From a purely mathematical point of view, the evaluations are not overly complicated. Often, such evaluations simply fail due to the lack of structured data collection in a central system.

In order to exploit the potential of such evaluations, it is advisable to proceed in the following steps:

The first step is to define as conclusively as possible all the exciting data that is collected somewhere in the company and that is potentially relevant to the success of the construction site. When in doubt, it is better to think too broadly than too narrowly.

The second step is to define the system in which the evaluation is to take place. For this, BI systems that are as flexible as possible are the first choice.

The third step is to define the input systems and the data transfer processes into the evaluation/BI system.

Finally, it is time for the IT implementation, which is often less complicated than first thought if it is defined properly in the previous steps.

Considering the high revenues and low margins of the construction business, the opportunity/risk profile of such a Big Data project seems attractive.