A Market Model that makes you go, “hmmm”

If there was a way of identifying which stores were likely to give you more business, you would want to know, wouldn’t you? In this article, I will be giving you access to a functioning market model that does exactly that but, more about this later.

In prior articles, I expressed an opinion that it was important to have a market model in place to establish an expectation for business rather than accepting, “it is, what it is”. The benefit being, that having a reasoned anticipation of retail potential will help focus Sales and Marketing resources in areas or stores where more value can be derived (it’s there to be had, you're just not getting it).

The premise is simple enough but, the reality is, it is not that easy to implement. It has been our experience that some will point to internal sales systems, CRM’s or external sales data as their means of developing such a model but, as indicated in the other articles this data may have unknown artefacts that will distort the underlying potential. It is important to make the distinction that a market model is a predictor of market performance, whereas using primary and secondary data is measuring actual performance. The difference between the two are distortions which may cause an outlet to trade above or below its potential. Above potential is good but, below potential means a growth opportunity.

To help demonstrate this point, this link will give you access to a market model for FMCG Suppliers into the Grocery and Hardware Industry for the state of Victoria (Australia). It is a very simple model that uses population as the key sales driver, the premise being, more people equals more sales. For many this may be sufficient but, we have found that other indicators (Number of Houses, cars, sex, age and income profile) may give a more refined view of what drives sales for your product category. Upon opening you will see a screen like this but, without the ID numbers;

 I’ll explain the layout of the model with reference to the ID numbers used in the above screenshot and how you can change the view of what is presented.

1: This is the list quadrant which ranks from highest to lowest each store by the share of the population. In the quadrant is the store name, the population that is within the radius of the store, the market share the store has of the radius and the resulting share of the population.

2: Indicates the number of stores in the database. The current selection is ranking 1,705 stores while 3: indicates these stores service, 3,836,926 people.

4, 5, and 6 are selection filters allowing you to change from Grocery to Hardware, Urban or Rural stores and a radius measure which will be discussed later in this article. Unchecking the Channel or Urban category filters will show all stores.

The Third and fourth quadrants provide further detail when a store is selected in quadrant 1. Like this;

In this screenshot I have selected Coles, Greenvale by left mouse clicking on the row in quadrant 1, quadrant 2 indicates that I have selected 1 store and that the population in its immediate vicinity is 15,418 people but in a 5Km radius there is still only 1 store and the population captured in this radius is now 43,369 people.

There are four dimensions being used in the algorithm to determine store share of population and they are;

1.      Location

2.      Population

3.      Service radius. And;

4.      Competition within the radius.

Using this algorithm, Coles Greenvale is shown as the largest trading supermarket in Victoria (I can hear the comments of, "that’s not possible", already) because it has a population catchment of 43,369 people and it has no competition and if you look at the Hardware view, Home Timber in Doncaster services a population of 148,037 people and has no competition in the radius either.

However, I wouldn’t advise immediately re-structuring the salesforce based upon this information… as mentioned before a market model is a predictor of performance based on a set of parameters. It is possible that some of the parameters require more investigation.

One of the factors involved in the calculation algorithm is the radius used for calculating the population catchment area and the amount of competition. A fundamental question for defining a market boundary is how far a consumer is likely to drive to consume a service. A long-standing adage in FMCG was a belief that the greatest value a supermarket had was convenience (proximity to the consumer) and that anything beyond a 2.5Km proximity meant the retailer would have to offset the loss in proximity value with other differentiating factors such as range, format or price. The Distance slider allows you to test what would happen to population capture and competition if you were to increase or decrease the capture radius (give it a try, the slider can be adjusted to ranges between .5 to 200Km).

By adjusting the slider to 10Km we can see that the competitive landscape for Grocery and Hardware markets changes and the store order adjusts dramatically. However, although I can understand a pre-disposition for consumers to travel further for hardware services, as there are fewer stores to choose from, I would argue that this is not the case for Grocery, where there are a plethora of stores in the urban area to choose from.

However, although this may be the situation in urban areas, in rural areas the choices become less plentiful. The Urban/Rural filter allows you to separate stores upon proximity to the city. By selecting Rural, you filter the store list to show those stores outside of the city, which you can then increase the distance factor to account for the long driving distances consumers are required to undertake.

However, it has been our experience that increasing the radius beyond 20km does not alter rankings significantly for the rural zone. A case in point is Australia’s largest population zone which is Toowoomba in Queensland with 107,000 people and 16 supermarkets in the area; even if the range were extended to 100km the population increase is only to 110,000 people and there are still only 16 supermarkets.

Within the algorithm, competition within the radius is the next important factor. Not only does the number of competitors, reduce the potential for the population per store but, there is also a matrix that acknowledges certain banners are more likely to compete better than others. Accordingly, the share of population calculated for a store is determined by the population, the number of competitors in the area and the competitive capabilities of those in that area.

Although, the data shown here is selective (Grocery and Hardware Industry, Australia), the model can be adapted to any industry where there is a service location and any country with population details.

If you wish to explore the data further, please PM me or go to the Market Grunt website and submit an enquiry.