Put your business on the map …….. Know where your customers are coming from!
In today’s competitive markets, businesses need to know how they can better serve their customers and to do that, they need to know more about them. The best indicator of future customer behaviour continues to be an analysis of past and current behaviour. A database of transactions (sales) as kept by most businesses describes this behaviour – it just needs to be analysed. The traditional approach to such analyses may focus on relative popularity of products or seasonal. This is, of course, important information but it is only the tip of the iceberg. Studying where customers are coming from can tell you much more about them; couple that with spend, product preference, seasonal changes and so on, and your data is turned into knowledge – you have real business intelligence. How it's done …….. The table below shows an anonymised sample of data extracted from a transaction database. Each customer is identified by a unique reference number (Ref.) and by the postcode where they live. This is a vital yet easy piece of information to collect during a transaction (e.g. for customer records, delivery, invoicing). Lets start by having a look at the spend of each customer. Using software called a geographical information system (GIS), the spend of each customer can be shown on a map at the correct location of its postcode. This map immediately tells you where to find your best spending customers (larger circles) or where considerable numbers of customers are concentrated into certain locations. But we might want to go further and explain this pattern. If there are clusters in the data, then some mechanism is causing the clustering and driving sales. If you know the mechanism you might be able to manipulate it, such as through marketing. In the map above there appears to be clusters, but this is subjective and we might argue about where the clusters are until the cows come home. So we take an objective approach, one which is unique to Terra Cognita – its called Geo-Prozone analysis. Geo-Prozone Analysis Geo-Prozones are geographical proximity zones where phenomena such as customers, burglaries, illnesses, water bursts and so on occur sufficiently nearby to one another according to some underlying controlling factors. The same pattern of customers as shown above is now given below as a point pattern with a boundary called a convex hull. This is the starting point of the analysis. The algorithm searches the map at different scales in relation to the total area for clusters and results in the map you see below. This shows three densities of clusters plus individual customers who are not close enough to any other customer to be part of a cluster. This is an objective view of the groupings apparent in the first sales map above. Using this as a guide, it is possible to establish same ‘sales areas’ in which some clusters and individuals are grouped together. Each sales area has a total spend and these can be ranked in importance. The map below shows this for our example. There are 15 sales areas that can be grouped into three lots of five in order of total spend. The 80 – 20 rule How often are businesses exhorted to follow the ‘80 – 20 rule’! Eighty percent of your results comes from twenty percent of your time. Eighty percent of your sales comes from twenty percent of your customers. Concentrate on the twenty percent! Sounds good – lets test it. Going back for a moment to the customer spreadsheet above, we can rank the customers in order of spend and graph this against accumulated percentage spend. This is the traditional gains chart.
This shows that twenty percent of the customers bring in forty-five percent of the sales. So what’s gone wrong? Nothing really, just the ‘80 – 20 rule’ doesn’t work all the time, particularly in businesses where customers are spending roughly the same amount. BUT …… all is not lost. Instead of individual customers, lets use the ‘sales areas’ in the last map above and create a geographical gains chart.

	This shows that the fifteen ‘sales areas’ mapped above account for eighty percent of the sales, and indeed the first five ‘sales areas’ account for fifty percent of the sales! Its obvious where you should be concentrating your efforts.
Finding more sales The work already carried out above has already told you a lot about the geography of your business entirely from the data already stockpiled within your business. Your existing customers in the most productive ‘sales areas’ could, for example, be sent a mailshot to give them news of new products, cross-sell or simply build a better relationship with them. But you may wish to dig still deeper into what is driving sales so as to inform your search for new customers. Here, external data such as census and/or lifestyle will probably be needed unless your business has also been collecting questionnaire data from customers. In the example being worked through here, we can test for a number of factors. Sales per postcode sector might be driven purely by the number of households or more subtly by some measure of affluence. The measure of relative affluence used here is an index specifically constructed on the basis of employment type and income groups. Other variables can be used to create other indices. The variables used here were merged into an index of affluence using our unique TC-transform developed for better spatial data handling. The correlations given in the table below show a poor correlation between sales and the number of households, a mediocre correlation between number of customers and households and a very good correlation with the index of affluence. The relationship between sales and affluence can be further explored in the regression model given below. This shows that whilst many of the customers are from middle affluent areas, new customers from higher affluent areas are likely to bring in at least £1,500 per sale on the basis of past performance. Like other data, the index of affluence can also be mapped with current sales to show current customers in relation to the better areas where one would look for new ones. The analysis shown here has been carried out on customer spend. It can be repeated for individual products, groups of products and specific customer characteristics such as gender or age groups. Business Geographics The example worked through above is but one. Business geographics can be used to assist in business decisions on customer relationship management, marketing, defining sales territories, location of outlets, deliveries and so on. And its not just retail business where these map analyses get used. Any organisation that needs to consider where (and lets face it, most things are somewhere) – banks, insurers, building societies, schools, hospitals, doctors, dentists, estate agents, cinemas, logistics, garages, even pizza delivery! Terra Cognita specialises in analysing the where of business – extracting business intelligence. Make your data work for you, don’t just stockpile it.
Put your business on the map …….. Know where your customers are coming from!
| Home |