Back

How to weight sites in a local scenario

By default every site counts as one, so a corner shop and a superstore carry the same weight in a local market. Weighted share lets you say otherwise - by what a site sells, how big it is, what format it is, or how far it sits from the centre of the market.

On the Scenarios page, under Configure market concentration, switch from Share of sites to Weighted share. Four kinds of weight appear, each with a Yes / No switch. Turn on the ones you want and set each up with its pencil; they multiply together, so you can combine them.

Weight shares by, with the four weight types and a red pencil on one that still needs setting up

A pencil shown in red means that weight is switched on but not set up yet, and Analyse local markets stays greyed out until you finish it or switch it back to No.

Fascia

Give each fascia in the subset its own weight - useful when one operator's stores are systematically larger or stronger than another's. Every fascia starts at 1; change the ones that should differ.

Weight by fascia, with a weight against each fascia in the subset

Tick several fascias and Bulk assign weights sets them all to the same figure in one go, which saves work when you're treating a whole tier of the market alike.

Bulk assign weights, setting one weight across the selected fascias
Distance

Weight a site by how far it is from the centre of the market, on the basis that a shop at the edge of a catchment competes less for those customers than one at its heart. Linear drops the weight evenly from 1 at the centre to 0 at the catchment edge.

Weight by distance on Linear, with the curve and the calculation behind it

Decay curves instead of a straight line, and you set the rate. Above 1 the weight falls away quickly near the centre; below 1 it holds up and drops near the edge. The panel draws the curve and shows the calculation, so you can see what a rate does before you commit to it.

Weight by distance on Decay, with the decay rate and what it does to the curve
Numeric value

Weight sites by a number already in your data - revenue, floorspace, checkouts, whatever you imported. Choose the column and each site is weighted by its own value.

Weight by numeric value, choosing the column to weight sites by

Choose one - here a revenue column - and Sloc shows what's in it: the smallest and largest values, the mean and the median. Worth a look before you commit. A column running from 1 to 108 weights the biggest site a hundred times the smallest, which is a strong claim; if the spread looks wrong, or the mean and median are far apart, the column probably needs cleaning before it carries the analysis.

A column chosen, with the range of values it holds
Categoric value

Weight sites by a label rather than a number. Choose a column - store size band, format, urbanicity - and give each value in it a weight.

Weight by categoric value, with a weight against each category in the column
Pro tips

  • Only the ratios matter. Weighting a superstore 3 and a convenience store 1 gives the same shares as 30 and 10, so pick numbers that are easy to explain.

  • Run the scenario on share of sites first, then again weighted. If the markets of concern barely move, the weighting isn't doing the work and the simpler measure is easier to defend.

  • A numeric weight needs the column filled in for every site. Sites with nothing there carry no weight at all, which quietly removes them from the share.

  • Weights are part of the scenario, so save it once you're happy - see how to save, open and share scenarios.