How to benchmark your locations and understand the drivers of the model?

With our predictive models you are able map the full potential of every location and every neighborhood. 

Additionally, it can be interesting to compare this predicted performance with their current performance and pinpoint underutilized locations and regions. 

The frameworks configured in the Location Performance module will help you find an answer to these questions, as well as help you understand all the relevant drivers that have an impact on the predicted performance.

1. Benchmarking framework (existing network) 

GOAL: benchmarking the predicted outcome (= benchmark) for each location with their current performance and identifying underutilized locations

Building blocks

• Benchmark: estimation of the predicted potential for each location, considering the current macro and micro environment.
  • For each existing location, we compare the real sales data to this benchmark. By measuring the deviation between the two, we can distinguish different location classifications.
• Drivers of the predictive model*: these are the columns in the framework displaying the macro and micro environment indicators that define the predicted potential of each location. (*below you can find a brief summary of these indicators)

How to use

• For each line/ location in the framework; analyze the real sales and the benchmark column. The deviation column displays the deviation percentage.                                                      
• To understand the benchmark of a certain location, analyze the columns displaying the drivers of the model (i.e. market size, competitive pressure, willingness to travel and all relevant location characteristics).
  • You can easily sort the data for a specific column or driver by clicking on the arrows () located to the right of the column name. Clicking the arrow pointing upwards will sort the data in ascending order, while clicking the arrow pointing downwards will sort the data in descending order. This way you'll see which locations in your network have the highest or lowest value for that driver.
  • The colours within the framework can be understood by analyzing each driver or column. The colour scale is determined by weighing each location against the average of that column for all locations displayed in the framework. 
    • Green indicates that your location performs better than the average on that indicator, while red indicates that your location performs worse than the average.
• If clusters (homogenous groups of locations with similar macro and micro environment indicators) were defined for your network, you can identify for each cluster which locations can still grow. By investigating the possible underlying causes, you can develop specific actions:
  • Actions for underperformers: identify growth opportunities
  • Actions for unprofitable location w/o growth potential: explore real estate opportunities to enhance profitability

2.  Benchmarking framework (new locations)

GOAL:  analyzing the drivers of our predictive model & benchmark it with existing locations, in order to have a better understanding of the predicted outcome. 

Building blocks: 

• Cluster: Each location is categorized into a cluster based on similar characteristics, such as macro and micro environment indicators. 
• Predicted sales: estimation of the predicted sales for each location

Remark! Predicted sales are not the same as the benchmark! 

• Drivers of the predictive model*:  these are the columns in the framework displaying the macro and micro environment indicators that define the predicted potential of each location.(*below you can find a brief summary of these indicators)

How to use 

Very much in line with the previous framework, but the current focus lies on benchmarking the new location(s) (displayed with a in the framework) with existing locations of which you expect similar performances.  We aim to compare locations that share multiple similarities (both in macro and micro environment) in order to get a better understanding of the predicted sales.

• Load the simulation with the new location(s) in the left settings bar.
• Make sure to display only the relevant locations (e.g. to benchmark the new location with) by using the configured location filters.   
• Examine now the relevant macro- and micro environment indicators that have an impact on the predicted outcome and benchmark them with the other locations. Sort each column from high/low or low/high. This way you'll easily see how the new location relates to the others for each specific performance driver(column).

3.  Opportunity framework (not by default configured)

GOAL: uncovering the causes of underperforming locations and define actionable strategies

Building blocks: 

• Similar to previous frameworks, we maintain columns for location classification and cluster, but these columns are now expandable to reveal the location indicators upon which they are based. Within your location filters on the left, you can filter both on location classification & cluster.
• Extra explanatory sales KPIs: these key performance indicators (KPIs) are organized into categories such as activation, loyalty, and conversion. Each of these columns can be clicked open to see the data it is based upon.
• Action descriptions: when you scroll all the way to the right, you will find detailed descriptions of recommended actions to be taken in collaboration with relevant stakeholders. This is based on easy ruling logic, and should be validated by the customer.

How to use

Read the following article: Which actions can you take for underperforming and/or unprofitable locations ?

*The indicators used by the predictive model are different for each customer and depend on the customer's preferences.