Geek looking at data

Resources

Data are only as good as their methodology

Mar 11, 2019, 10:38 AM by Tony Lea, PH.D.
ClickSpend™ takes the best data available and makes them actionable
 

At Environics Analytics, we are committed to making people’s lives better and our clients more successful through transformative data, insights and analytics.

We work collaboratively with our clients and partners to develop solutions that they can trust and fit their needs. To deliver on this promise all of our products adhere to rigorous methodologies. The importance of adopting a sound methodology cannot be overstated.

We employ some of the best methodologists in the world and we continuously refine our models to ensure they are as accurate and reliable as possible and line up with established control totals.

The work that went into developing ClickSpend™ powered by J.C. Williams Group is a prime example of the rigorous process we follow when designing all of our products. Before we build our methodology, we take the time to make sure we are leveraging the most authoritative data sources available. For ClickSpend, we used data from reputable sources including Statistics Canada, J.C. Williams Group, Euromonitor and others.

To make these data actionable, we model higher-level data down to the neighbourhood level. This is a transparent process, which we disclose to our clients and partners. For ClickSpend, our team of experts and methodologists worked with J.C. Williams Group to devise a methodology that provides the best reflection of what consumers spend online.

As is the case with most data products, we are required to make some reasonable assumptions in our models to ensure they tell a story that is consistent with established control totals. Different estimates come from the various assumptions and approaches developed during data collection and modelling process. One large conceptual difference is whether these data are collected from a sample of stores or consumers. Each approach has merit, but they produce different estimates.

Data experts know that various issues and biases need to be addressed to ensure their data are reliable. For instance, were the right stores or people surveyed or were all parts of the businesses captured? Because of these issues, it is not always appropriate to compare the final estimates from different databases with others.

For ClickSpend, we mainly base our estimates on the demand side of retail spending and tie them directly to Statistics Canada’s data from the National Accounts. Retail experts widely acknowledge the differences between the supply and demand views of retail spending in Canada. Depending on your approach, different assumptions and adjustments must be made to ensure the final data are consistent and reliable.

Here is an overview of the methodology and data sources we used to develop ClickSpend:

  • We estimated total retail expenditure from Statscan’s Survey of Household Spending in its microdata form. The process we follow for ClickSpend is consistent with the well-established methodology for our HouseholdSpend database, which offers 480 variables. The advantage of ClickSpend is that it provides a new perspective about the purchase behaviour of consumers online that was previously unavailable to our clients.
  • We used survey data from the J.C. Williams Group “Etail” survey data to estimate what proportion of that total spending was online.
  • We adjusted the estimate amounts spent online to match the total online spending data from the respected Euromonitor.
  • We reconciled the dollars with Euromonitor to ensure the adjusted estimates add up to the appropriate and respected Statistics CanadaNational Accounts totals.

Here at Environics Analytics, we are proud of our thorough and multifaceted approach and the resulting widely used expenditure estimates.