The Art and Science of New Site Selection

Finding new sites and executing a growth strategy in retail and hospitality is not simple. Good site selection involves reviewing multiple sites and prioritising which sites are likely to be the best performers. Property teams need to compare lots of sites using multiple factors in a consistent manner.  Do you really want to make property selections based on your “best guess”?

With access to some data sources, it is possible to do this yourself. But ranking and prioritising sites can take substantial analysis time and management review time.  With so many factors to consider and modelling options it can be quite a challenge.

  • Existing sites and relevant operating metrics.
  • Choice and span of location data variables.
  • Sourcing, quality and preparation of data.
  • Correlation among variables and the number of variables to include.
  • Location catchment size and footfall considerations.
  • Choice of model(s).
  • Time periods, seasonal impact and pandemic impact.
  • Validation and testing considerations.
  • Point in time sales forecasts and predicted weekly trading patterns.

A model that can accurately prioritise and predict sales at the next 5-10 locations is very valuable, especially if capital expenditure is limited or you are raising new capital. Moreover, you may have come across an opportunity for a great rental deal and want to predict sales and evaluate profitability in a consistent way.

A data driven approach is an excellent foundation, but the final location decision always involves the human touch.  This qualitative validation includes a wide range of factors such as site lay out, parking restrictions, competitor price lists, strange odours and other nuances. This is where property specialists and management teams really shine – the art of making the final forecast adjustment and investment decision.

Algospark is a specialist in delivering AI solutions for new site selection. Find out more about how we helped Gail’s Bakery expand from 35 to 80+ sites.

We believe that understanding key considerations for data collection, data selection, modelling, validation and visualisation is also an art – the art of successful data science.