Hotelling’s Law (1929) is an observation in economics that in a competitive market, firms have an incentive to position their products to be as similar or as close to each other as possible. Given the effort firms so often appear to put into marketing and other methods in order to distinguish themselves from their competition - the principle of minimum differentiation can be a somewhat counterintuitive result. Nevertheless, it is one that can be found in many real world contexts, helping to explain why car dealerships often cluster together and why competitors often end up selling almost identical products (think coffee chains in the UK for example).
If you are unfamiliar with the concept, a quick explainer from Ted-Ed can be found here.
As the video lays out, for competing firms, the Nash Equilibrium solution (though it would not be called that until two decades after Hotelling first wrote about it) is for each firm to position itself as close as possible to the other. By positioning themselves as close to their competitor as possible, each firm is maximising their market share, given where their competitor has placed themselves. Each firm will react to the other by moving slightly to one side of it and if there are two firms in a market then both firms will eventually cluster where the median consumer is.
Although in the example in the Ted-Ed video, the positioning occurs along a geographical distribution, the principle can be applied in a similar way to a preference distribution, where instead of clustering at the geographic centre of a market, the firms cluster at the median preference position in the market. For the purpose of this blog, let’s assume that preferences for distributed normally along a single dimension for a given good. So in the case of Fanta and other similar ‘fizzy orange’ drinks, instead of people distributed along a beach, they have preferences distributed from ‘weak’ to ‘strong’, with firms clustering at the preference of the median consumer to ensure they maximise their market share.
Note here that median voter theorem is basically the same idea as Hotelling’s Law only applied to electoral politics.
It is not unknown for large, international companies like Coca-Cola to tailor their products to fit the preferences and tastes of local markets. Consumers in different markets will have different preference distributions and so in order to maximise their revenue across each market, firms will have to adjust their products. A classic case of this is how the McDonald’s menu varies across countries. In our example of fizzy orange drinks, people in different countries will have different strength preferences depending on factors such as culture, diet etc. Some countries will have a preference for stronger fizzy orange drinks and others will have a preference for weaker ones. The position of the median consumer will vary in each case, meaning that a cross-country product like Fanta will have to adjust in order to capture that median consumer and maximise its share of the fizzy orange drink market. The substantially different strength and colour of Fanta across the globe is evidence that this sort of local market clustering occurs.
However, a more interesting example of this phenomenon is when rather than positioning their products directly from the underlying distribution of consumer preferences, firms position their products indirectly using the distribution of existing products. Remember, Hotelling’s Law states that a firm will aim to position their product to maximise their market share GIVEN the action of other firms. If firms have clustered at the median point, then the optimal response of any additional firm is to also cluster at the median point - but there are a number of reasons why this clustering at the median may not occur and if firms have positioned themselves elsewhere then the optimal response for a firm may not be to position their own product at the median.
In the video example above, it was relatively cost free for the ice-cream vendors to reposition their products, they simply had to move their cart. In such a low cost scenario, we are likely to see firms gradually gravitate towards the median consumer. However, in the real world, repositioning is rarely this costless. Altering a product involves significant investment in research and development, the retooling of machines, the retraining of labour and carries the risk of damaging a brand reputation that has been built up over many years (Coca-Cola learnt this lesson the hard way in the 1980s).
In addition to these repositioning costs, firms may also lack the information on the nature of the consumer preference distribution that is required in order to position themselves as close to the median consumer as possible. In such cases, a firm’s best guide regarding where to position their product (particularly if they are a new entrant) may be the position of other existing competitors. If an existing product has already placed itself far from the median, and due to repositioning costs there is little prospect of it changing its position, then the optimal strategy for the new entrant is to simply position itself slightly closer to the median than the existing product, ensuring it captures all of the market on both sides of the median, up to the extreme position of the pre-existing product.
Comparing the relative strength of Fanta in mainland Great Britain and Northern Ireland provides an interesting case study of this behaviour, where a firm adjusts its product in response to existing local competition, rather than in response to differences in taste. (Note: obviously I have no evidence as to what Coca-Cola’s thinking behind their product strategy actually was - what follows is purely for illustrative purposes).
Historically (historically here means pre-2018), Fanta Orange in Northern Ireland was significantly stronger in taste than its counterpart on mainland GB and was a much darker shade of orange. Naturally for such a niche subject, there is little hard evidence on this that I can find beyond personal anecdote, but here is someone raising a similar point on a forum a few years ago.
Given the geographic, cultural and dietary proximity of mainland GB and Northern Ireland, it is unlikely that this difference in strength could be explained by significant differences in consumer preferences for fizzy orange drinks between the two markets. However, there is a significant difference in the existing competition faced by Fanta in the respective NI and GB fizzy orange drink markets that may explain why the product differs in taste so much between the two.
Fanta’s initial history in World War 2 is relatively well known at this stage, but it was only after its relaunch in 1955 that it started to become available in the UK. By the time it was entering the UK market in the late 1950s and early 1960s, it already faced pre-existing competition in both the GB and NI markets and the nature of this competition helps to explain why Fanta developed so differently either side of the Irish Sea. In the GB market, Fanta’s main competition (with the possible exception of Scotland - but it is debatable as to whether Irn-Bru is actually orange flavoured) was and remains Tango, a relatively weak and lightly coloured drink. If the GB consumer population was distributed from ‘weak’ to ‘strong’ preferences for orange drinks on the normal distribution above, Tango would be towards the weaker end on the left hand side. Therefore, when Fanta launched in GB in the late 1950s, it only had to position itself to be slightly stronger Tango in order to get the custom of everyone to the right of it on the taste distribution (which included the median consumer - as shown on the diagram below). Since it only had to be slightly stronger than Tango, Fanta in GB became a relatively mild drink with a light orange colouring.
In Northern Ireland, Fanta faced very different competition. Although Tango is sold in NI, it is fairly rare with the main competition to Fanta being a drink from the Republic of Ireland called Club Orange. Launched in 1930, this is a much stronger drink, with 13g of sugar per 100ml (compared to 4g for Tango), a darker colouring and even actual orange bits in it (as you would get with pure orange juice). Whereas Tango is positioned towards the left of the normal distribution above, Club Orange would be much further to the right, beyond the median consumer.
Therefore, when Fanta entered the market in NI, simply using its GB product would not be an optimal strategy given the position of its main competitor. If it used the milder GB product, it would get the consumers with a weaker preference but would risk losing the consumers with a stronger preference (potentially including the median consumer) to Club. In light of this, Fanta altered their product to suit the NI market, making it significantly stronger in colour and flavour so as to position to be just slightly weaker than Club. This strategy meant that Fanta would get everyone to the left of club on the distribution (including the median consumer), meaning only those with the very strongest tastes were left with Club.
Fanta therefore provides a useful illustration for how new entrants can use the position of existing firms to act as a guide when designing new products. One final note on this example is that it held up until the introduction of a Sugar Tax by the UK Government in 2018, when Fanta and many other drinks changed their recipe - reducing their sugar content in order to avoid the levy. Since the tax applied to the entirety of the UK, the recipe change also occurred to Fanta in Northern Ireland, which is now much more similar to its counterpart in GB than it had been up until that point.