How to Build a Data Solution From the Ground Up
By Jason Keen
When a new building is constructed, everything must be built from the ground up. The project site must be analyzed, requirements gathered, concepts designed, plans created, schedules finalized, and construction performed to create the final product.
Similarly, creating a company’s data infrastructure and solution must also be built from the ground up.
It is important to visualize the full scope of the company’s wants and needs, regardless of whether they are in the beginning stages of planning or if a large quantity of data has already been accumulated, compiled, and analyzed.
The construction financial manager (CFM) performs a key role in connecting operations, systems, and data into a comprehensive system.
As data is compiled, there are key questions to answer:
- What makes up a quality data solution?
- How does a company go about determining which pieces will make up the final product?
- What are the steps to put an effective data solution in place?
The CFM’s goal is to provide answers to these questions while laying out a simple data architecture.
Every company has different systems and needs. However, the data solution end goal is the same for all companies: Put a system in place that captures data quickly, removes redundancy, provides accurate information, and makes it easy to measure and assess the business.
Start by looking at the systems currently in place, like the accounting system, also known as the enterprise resource planning (ERP) system:
- Does the data have to be entered more than once?
- Is the system technologically up to date?
- Is the system being used fully?
- Is the system being used efficiently?
- Are the right people using the right components of the system?
As the company answers each of these, steps can then be taken to build out the needed requirements. For each “no,” the question then becomes “why” followed by “what can be done” to make the answer “yes.” This methodology can be carried across all systems, whether paper-driven, electronic, or cloud-based.
The next step is to flow diagram all of the pieces of each system/process, look at the gaps, and brainstorm how that system can be used more effectively. At this time, it’s important to also look at which functions are currently lacking in the system. If the system is lacking too many items and/or the technology is outdated, then it may be time to start the process for evaluating and implementing a new one.
Whether the company decides to build-out and leverage the current system or find a replacement, the goal remains the same: Fully utilize the system to automate as much data capture as possible and speed up the data capture process(es) to provide accurate and timely information.
New Systems & Data
Once the current systems have been analyzed and optimized for data capture, usage, and accuracy, the next step is to assess what data is missing. There are several ways to approach this:
- Ask key leaders what information they want readily available.
- Ask industry peers what they measure and how they capture that information.
- Research data and analytics vendors that work in the construction industry. Make sure to select an independent consultant that does not have their own product offering. This approach will allow for an unbiased analysis and opinion.
Then, take the information captured in the analysis and create three rankings: must have, nice to have (wish), and long-term potential. Rank these lists in order of importance, taking note to identify easy wins.
Analyzing Resource Requirements
Along with identifying missing data and information, decide which reporting/analytics product to use based on whether or not the information is presented in a useful manner. There are many tools available, but it is critical to make sure that all of your systems – from capturing to reporting – are built on a compatible architecture.
Here are some questions to consider when looking at both current and future systems:
- What are the current core systems built on (SQL, .net, Oracle, software as a service (SaaS), etc.)?
- Then, go a bit deeper: SQL and .net are Microsoft-based products, whereas Oracle is an Oracle product. Different systems can be integrated, but the external or internal resources needed to support both structures will be more labor resources.
- SaaS is a cloud-based system that does not require maintenance or knowledge of the underlying architecture. SaaS-based systems can integrate with other solutions via an application programming interface (API). Each API comes with documentation of what can and cannot be transferred in and out.
- Does the company have the correct human capital to use the selected capture systems, reporting systems, and system integration? If the answer is “no,” then there will be an additional cost to hire the needed talent, build the existing employees’ abilities via training, or use a third-party to manage and implement the systems.
In the end, it will take a combination of systems and staffing to put the proper architecture and systems in place.
Once the missing data points are identified, the systems are selected, and the talent is in place, it is time to start implementing the changes.
The best place to start is by building out some core reporting and dashboarding to deliver quick wins for users. With reporting in place on existing data, users can conceptualize the bigger picture and value of the systems.
After core dashboards are in place, build a project road map that accounts for all of the changes to current systems, implementation of new systems, and any other necessary items. This road map is a living document that captures timelines, plans, and dependencies of every system and change. This document should be adjusted throughout the implementation process for changed timelines, additions, and/or other systemic decisions. The road map gives everyone in the organization insight into what is ahead, what has been done, and expectation alignment.
Use the road map just like a project plan – as the blueprint for how the company’s systems and processes will change over time, while also highlighting potential resource needs and challenges. For example, there may not be enough staff to support the required pace of change, or certain projects may need an external resource put in place. Think of the road map like a Gantt chart that shows committed resources and the critical path for completion.
So why are most companies pushing for more data, dashboards, and analytics? The key driver is timely and accurate information to drive business adaptation, change, and decision-making. Historically, companies have been stuck in the rearview mirror rather than looking ahead and anticipating change. A real-time capture of data that is quickly available for decision-making could be the difference between a job gain or fade.
Let’s walk through an example. XYZ Constructors has determined that labor cost is a key driver of its business. Company leaders want to be able to see job productivity trends 24 hours after the work is performed, along with a trend analysis of productivity rates. This information will help to anticipate where the job will end up on labor vs. bid rate.
They currently capture time on paper and then input it a week after the work is performed. XYZ Constructors decides to implement a digital time capture solution. The foremen are given tablets and told they must enter their crews’ hours at the end of each business day. The project managers (PMs) must then approve the time by 10 a.m. the following day. This system has the same architecture as the company’s ERP, so once the PMs approve the time, an API is written to automatically move the time into the ERP system. The payroll team then reviews, corrects any errors, and processes the time to the job cost subledger by 3 p.m. each day.
XYZ Constructors also puts a dashboarding tool in place that can source information from the ERP. Along with the time information, the company enters all estimated labor production rates into the ERP.
With this information, the dashboard automatically pulls both the actual production information along with the estimated production and computes key metrics. With these key metrics, PMs can see how their job has performed job-to-date, month-to-date, and week-to-date by 3 p.m. the following day. They can see this information along with estimated and forecasted production.
Since they are now able to see all of this information in one place, it quickly becomes apparent what is going well and what is not. Rather than looking in the rearview mirror and questioning what could have been done better at the end of a project, PMs can react to issues quickly and course-correct problems when they happen.
This is only one example of putting the key pieces in place to move data quickly from capture to reporting. It takes people, resources, processes, and systems to be successful. The other key aspect is for each dashboard to identify all of the needed data points. Using the XYZ Constructors example, the actual data would mean very little without the original estimated production rates to compare against. Also, by layering in the forecasted rates based off of trends, it makes it easy to anticipate where things may end up.
While not every implementation goes as smoothly as the XYZ Constructors example, with proper planning and buy-in, implementation success can improve/increase significantly. Here are some key items to consider, regardless of the magnitude of change.
Obtain Feedback from the Targeted User Group of the Solution
Come with a plan of the dashboard’s elements, then obtain feedback. Users will often provide perspective about items not originally considered since they know the day-to-day process.
Obtain Executive Buy-In & Identify a Project Champion
People may think a certain dashboard is a great idea, but without executive support, there is risk of producing information that does not fit the corporate vision or strategic goals.
Dashboard Considerations & Key Items
There are several key aspects to consider when developing a dashboard:
Consider Where All of the Data Is Going to Live
Will data be disaggregated across multiple systems? Will all of the systems be integrated? Will one system hold most of the data as the core database? Or, will it be a combination of both approaches?
Build in Layers
Rather than creating a dashboard that is granular, build a summary-level dashboard that drills down into the underlying information. Use the top level of a dashboard to emphasize key trends and/or issues; then, design it to allow users to dive in and analyze the information.
The difference between a dashboard and a static report is the ability to drill further down into the information. Make sure logical thought is put into each level of the dashboard:
- Identify the key metrics and patterns that need to stand out for the top level of the dashboard. The top level is usually driven by graphics in the form of charts and trends.
- After identifying the top-level metrics, determine the next level of drilldown to provide context to the outliers. This level of the dashboard may still have some graphics, but it is much more numbers driven.
- Finally, think about the third level as a pivot table of information in a spreadsheet providing all of the transaction details and then allowing users to manipulate the data. The goal of this level is to help the user determine the exact driver of a particular issue.
Of the many considerations when building a data solution, here are the four most important ones:
- Look at the company’s current systems. Determine how to optimize each systems’ utilization or if system replacement is needed.
- Identify missing pieces of data that need to be captured. Determine the best method(s) of capturing the data and where the data will be housed.
- Look at the needed resources to put the data solution(s) in place and layout a road map that provides an implementation timeline.
- Implement dashboards on both the current and new data systems concurrently to provide users with information they need while also creating the ability to build out more data over time.
Building a data solution is an interactive process that takes time. There will always be more solutions to consider and implement. Continue to adapt and progress, and over time the company’s data will become one of the most valued business drivers in the organization.