May 4, 2021
When a brand’s real estate team begins using analytics for site selection and sales forecasting, often one of the first questions asked revolves around the data used to build their real estate models. With the wide range of data sources available, knowing what data to use and how to provide it can have a significant impact on the quality of the real estate models and the timing of the solution delivery. Our experts have outlined tips to make data integration as seamless as possible.

Make a data plan
Even if you’re not quite ready for real estate forecasting models, it’s never too early to start thinking about how to obtain data and what to collect. Consider what you would like your mapping system to do and what data you’d like to see. Use this as the basis to develop a plan for data collection.

Prepare a plan to collect the data you’d like to see in your mapping and modeling system.

Collect the most valuable data points

What is the most important data to collect for your brand? Having accurate address information for each of your units is critical, as well as the lat/long coordinates if available. Sales, open dates, owner and franchise info, and unit or building type can provide the basis for developing real estate models. If you are looking for a more in-depth analysis of your sites to use in your sales forecasting, noting any attributes that vary considerably between units helps determine the potential impact on sales. 

Operational “Inside the Box” attributes specific to each site—ownership and management metrics, brand awareness measurements, ad channels—can help fine-tune predictive models and account for data that an analytics partner cannot provide. For example, if you have a downtown business district location that is closed on Saturdays and/or Sundays because surrounding businesses are also closed, this is something to take into account to improve accuracy when creating your models.

Identify and collect the most important data points relevant to your brand.

Organize your data
Using a template and checklist to organize your data will allow you to easily view and double-check information to avoid common errors like invalid or inconsistent field values.

CBRE Forum Analytics provides clients a data checklist and location templates to make it easy for everyone to stay on the same page with what has been collected and what is still needed for the project.

Implement a template and checklist to organize your data.

Check for Accuracy
Ultimately, the data you provide is used as the starting point for all subsequent analytics work, and the accuracy of that data directly affects the quality of that work. It pays to take ownership of your data collection process and ensure what you are providing is as complete and accurate as possible.
This can mean consolidating all your data into one file with a unique ID for each record. Check for duplicate IDs, addresses, and review your data to see if anything looks strange or out of the ordinary.

Analytics and predictive modeling can only be as good or accurate as the data that is used. There’s a reason statisticians say “Garbage In, Garbage Out” when discussing the validity of predictive models.

Clean, accurate, and up-to-date data is the foundation for analytical work.

Streamline Your Data Transfer
Probably the biggest misconception when providing data to create forecasting models is that sending information as it becomes available will speed up the project process. This couldn’t be further from the truth. In reality, model work cannot begin until all data is received and organized. Each time new data is sent, it needs to be combined with previously sent data, which can take time and leaves greater room for error. Not only does this not speed up the process, but it can lead to significant delays in the project.

Collecting and transferring all your data in one consolidated file is the most efficient way to begin your analytics project. It will save time and greatly reduce the possibility of errors.

Clearly define your data
When you transfer your data, it’s important that it is clearly defined for the team providing your analytics. Sharing definitions and metadata for each file will save time and help address any questions. Knowing exactly what each data field represents will also help the statisticians better understand when they are deciding which factors to include in predictive models.

Include definitions for the data you provide for forecasting models.

Keep data updated
You do your brand a disservice by using outdated data. Updating data as often as possible will help keep the analytics in your platform and reports as up-to-date as possible. 

Optimally, data updates should be made on a monthly schedule if not more frequently. Location data should be evaluated each time you run a report where they are included in the output, and update the full location file where needed as often as possible where the platform allows. While changes to model attributes won’t change forecasting models on the fly, keeping data updated throughout the year will reduce data collection time during the model refresh and ensure the update is as accurate as possible.

Make a plan to update your data on a regular cycle.

Contact us today to learn more or if you have questions about your data strategy.

Put our intelligence to work.