Analogue Modelling

By Sarah Stowe | 29 Oct 2015 View comments


The analogue model is a site selection and network assessment tool which works by comparing each site in the network to all the other sites based on certain drivers of sales. This model is based on the assertion that analogous sites (analogous in terms of the selected drivers of sales) should be performing at similar levels.


Depending on the size of your network and your knowledge of the drivers of your business, we will ask you to quantify some of these drivers for every site in your network. Spectrum can also nominate some drivers of your sales based on its prior experience and by use of appropriate statistical techniques.

Once this is done, SpectrumÕs specially designed computer programs are able to match every site in your network to every other site and cluster similar sites together. Instead of clustering similar sites together, it is also possible to calculate a Ôsimilarity indexÕ. The clusters or indices can then be used to compare the actual performance of each of the sites to that of the cluster on an average. This approach is also useful for predicting the performance of future sites.


  • Estimate the performance of proposed outlets at any location
  • Perform scenario planning
  • Alleviates Ògut feelÓ from decision making
  • Compare actual vs. potential performance of existing outlets
  • Prioritise ÒNew BuildÓ areas


Lets say you have a network of 10 sites. You nominate 5 drivers of sales all ranked on a scale of 1 to 10. The drivers could be the traffic levels passing by each store, levels of favourable demographics around your sites or anything else you think is relevant.
For illustration, lets say your site A is ranked Ô1Õ on all the five drivers and is doing $1000 worth of sales every week, your site B is ranked Ô2Õ on all the five drivers and is doing $2000 worth of sales and so on.

Once we have developed an analogue model especially for you, we are able to calculate a similarity index. This index can be used to compare an existing site to its analogues in the existing network and also predict the sales for this site.

The table below shows that a new site which is ranked a Ô3Õ on driver _, Ô3Õ on driver _, Ô4Õ on driver _, Ô4Õ on driver _ and also a Ô4Õ on driver _ is most similar to the existing sites C and D. Also, site J is far better that this new site and site A is far worse.

In the above table, shades of green are indicative of similarity to the new site, and shades of red are indicative of dissimilarity. Alternatively, a similarity index approaching zero is indicative of relatively high levels of similarity, and a similarity index going away from zero is indicative of higher levels of dissimilarity.

Please see our Downloads section to view full article.