EconomyRice

For a complete index of posts, click here.
s
Why do Starbucks and  Pret-A-Manger (or Coffee Bean) often have their shops next to each other? Or the same phenomenon with McDonald’s and KFC? Why do you see competing banks clustered along the same road?  Surely, you think it would make more sense for two similiar and competing firms to locate themselves as far away from each other as possible. Yet, all these firms with a clear profit motive often behave in a manner that contradicts seemingly intuitive logic.

We need to figure this out. So let’s play a game.

Prelude: Theory – Hotelling’s model

The first thing I’d like to do is introduce you to the Hotelling model, a famous bit of game theory that deals with this exact situation. There is an excellent YouTube video that illustrates the model here; I highly encourage you to watch it before reading on!

In any case, I’ll briefly outline the model here. If you’re familiar with the model already, feel free to skip ahead.

Say Hilary Clinton and Donald Trump were both competing to sell ice-cream on a beach that was 100 meters long. We’re going to make a few assumptions about this beach to simplify the situation.

• There are children uniformly distributed along the beach.
• Every child is equally willing and able to pay for an ice-cream, and will buy an ice-cream if there is someone selling them.
• Trump and Clinton are selling the exact same ice-cream. You cannot distinguish between their products. (In economics lingo: Homogeneous goods)
• They are both selling at the same price. This makes sense, because if they are selling the exact same good, then the one charging a higher price would probably sell nothing.
• They can both move their ice-cream trucks at no cost, i.e. location is perfectly flexible.
• Given the above, a child will simply buy an ice-cream from the nearest seller.

In an ideal world, Trump and Clinton would peacefully co-exist, and situate themselves at the 25 meter and 75 meter marks. This way, two things would happen:

1. They would both get exactly 50% of the market.
   The circles I’ve drawn on the diagram are to show their business areas.
   Any child between 0 and 50 would go to Clinton since she would be the closest. Similiarly, any child between 50 and 100 would go to Trump.
2. No child would ever have to walk more than 25 meters to buy an ice-cream.

However, Hotelling’s model (and John Nash’s work on game theory) tells us that this outcome cannot last. Either Clinton or Trump will soon realize that they would capture more of the market if they moved their trucks. For instance, if Trump moved his truck to the 25 meter mark as well, he would now have 75% of the market.

Both Trump and Clinton will now be worried about foul play on their rival’s part. So what do they do?

They both move to the 50 meter mark. They still both get 50% of the market, but now, neither one can increase their market share by moving.
For instance, if Trump moved his truck to the 49 meter mark, Hilary would capture 51%. So Trump wouldn’t move. Similiar logic applies to Clinton. We can note a few things:

1. Compared to the previous situation, their market share hasn’t changed. They’ve merely prevented their rival from profiting from dirty tactics.
2. The children, however, are now worse off.
   Previously, no child had to walk further than 25 meters to get an ice-cream.
   Now, the furthest child is 50 meters from the promised ice-cream. Poor child.
3. On a more formal note, this situation, where neither person can improve their position by moving, unless their rival also changes position, is known as a Nash equilibrium.

If you found my explanation confusing, it may help to watch the video I linked you to! Watching an animated version of this model may be clearer.

In any case, the key takeaway from all of this is that Hotelling’s model predicts that two firms competing under these circumstances will end up locked in equilibrium, side-by-side at the halfway point of the market.

The intuitive thing is to do is to proclaim that the first situation (one at 25, one at 75) is what will materialize. This is certainly the best outcome for consumers. However, it leaves firms vulnerable to attack, hence the model’s outcome of Situation 2 (both at 50).

Now that we’ve got a theoretical explanation in our minds, let’s tackle the real world.

---

The Real World: Coffeehouses

The above sight is stunningly common. I love looking out for it wherever I am. However, Hotelling’s model, as discussed above, is insufficient to explain it. A few ways in which real life deviates from the game:

1. The model assumes that the sellers are perfectly free to move around. This obviously isn’t the case. There is a large cost to relocating. Space for relocation may also not be available.
   (space)
2. In the model, both sellers are competing simultaneously.
   In real life, one outlet probably established itself in that location first, before the other came along later.
   (space)
3. Starbucks and Pret aren’t selling exactly the same thing.
   This allows for differentiation in :
   a) type of good (for instance, Pret has a wider range of pastries)
   b) quality of good (I think Starbucks’ hot chocolate is better)
   c) price
   In other words, factors other than location will influence the way consumers choose between the two.
   (space)
4. (This is very important)
   Not all locations are equally good.
   In the model, if you were the only seller present (a monopoly), then you could place your truck anywhere and still capture 100% of the market.
   In real life, some locations are far superior to others. For instance, if Starbucks placed its store right outside a giant business hub, it is likely that it would draw a huge number of corporate consumers looking for their daily coffee fix. However, if it placed its store even 3 streets away, there’s a good chance that a portion of the busy working adults who would have bought coffee from them when it was convenient would now see it as too much of a hassle.
   (space)
5.  You don’t precisely know who your consumers are.
   In the model, as long as there was an ice-cream seller there, a child would buy one.
   In reality, some consumers may decide that past a certain distance, it just isn’t worth it. You also aren’t sure exactly who can afford it, and who can’t, or if some people have very strong preferences for a particular brand.
   (space)
6. Your area of competition isn’t cleanly defined.
   The model is simple. It’s a nice, finite, straight line.
   In reality, the geography of competition is far more complex, and will change over time. A location that seems desolate and undesirable today may boom in 2 years if someone decides to turn it into a residential area, for instance.

Despite these differences, there is, however, a compelling reason to believe that Hotelling’s model does explain coffeehouse competition to an extent.

Imagine that you were the manager of Pret-A-Manger, and you had to choose a location for your store. As you look for the optimum spot, you assess your competition (in this case, Starbucks). I earlier explained that not all locations were equally good. The logical implication of this is that if Starbucks is already in operation, then it has chosen the best locations for itself.

Hotelling’s model offers you an initial solution:
You should locate yourself right next to Starbucks.
The model, on a rudimentary level, tells you that if you locate yourself next to your competition, then you end up capturing the same market share as your competitor.
You could risk choosing a different location in an attempt to defeat Starbucks, but you then run the risk of having your tactics and analysis go horribly wrong. So you go for the safe option of locating yourself next to Starbucks; this is a tried-and-tested location.

That’s as far as the model can take us, because the only factor in play is location.
I want to go further.
There are many, many reasons besides the model’s single variable which suggest that the strategic move for Pret-A-Manger is to locate itself next to Starbucks.

1. Starbucks was established in 1971, has nearly 25,000 stores worldwide, and rakes in about $20 billion in annual revenue. Pret was established in 1983 (Starbucks has a 12 year headstart), has about 400 stores worldwide, and makes about $500 million in revenue yearly.
   
   If it so happens that you choose a location where Starbucks isn’t yet operating that proves to be a prime spot, Starbucks can easily afford to open an outlet there as well. Even if that specific outlet had to operate at a loss, Starbucks can afford to do that. Indeed, one of Starbucks’ strategies is to flood the market, sometimes having 3 outlets within 500 meters of each other.
   This means that even if Pret did manage to win the initial location battle, the advantage is unlikely to last for long. It is better to stick with the safe option of locating next to Starbucks. Much lower risk of a corporate war, given that it really is difficult to steer clear of Starbucks in the long run.One thing to note: Pret is now sufficiently big that Starbucks probably wouldn’t find it worth it to wage a price war against them. However, if it was a completely new and purely local brand, it is possible that Starbucks might try to drive them out of business, for instance by giving out free coffee, if they located right next to Starbucks. Anti-trust legislation might protect against this, but then again, it might not.
   (space)
   
2. Starbucks has already established its brand worldwide. By locating next to Starbucks, Pret gets to piggyback on the Starbucks brand by getting consumers to associate Pret with coffee as well. Essentially, Starbucks has already done your advertising for you. You get to introduce your brand to consumers who aren’t familiar with you.
   (space)
3. You get to steal customers during peak hours.
   Many Starbucks outlets are located in prime spots.
   Starbucks doesn’t sell coffee at a constant frequency throughout the day. There are naturally busy hours (such as in the morning before work) when Starbucks likely does a roaring trade. At these times, lines get long, and there is bound to be some waiting time. By locating next to Starbucks, Pret gets to capture some of Starbucks’ customers whose main concern is to get a quick cup of coffee, and are willing to switch to Pret if it means a shorter queue.
   (space)
4. There is a snowball effect of having two coffee shops in the same location.
   That area now gets seen as the ‘place to go’ for coffee, given that the two firms together can serve a greater total number of customers. This could mean that Starbucks and Pret combined draw in more customers than either could on their own.
   This, by the way, is the reason why you tend to see clusters of similiar firms. It makes sense to establish a location as the ‘place to go’ to buy Item X.
   (space)
5. You allow consumers to compare, and invite them to try it out.
   If Pret, a potentially unfamiliar brand, had located itself away from Starbucks, consumers who were already comfortable with Starbucks are quite likely to be unwilling to head to a different location for an untested brand. Locating a Pret next to Starbucks makes it convenient for consumers to try it out, and compare the two. If you’re lucky, they may even decide that they like Pret more!
   space
   In the UK, where I currently reside, I’ve noticed that prices in Starbucks and Pret aren’t wildly different. Both sell an Americano for around £2, depending on how you customize it.
   In Malaysia, where I previously lived, you’d often get smaller brands like Each-A-Cup or Chatime located near huge brands like Starbucks or Coffee Bean. The price difference is huge (RM10 at least at Starbucks versus RM5 at the smaller store). You get to capture consumers who may prefer a Starbucks, but come to you anyway for a cheaper price, but only because you’ve made it convenient. If you weren’t situated right there, they wouldn’t ever come to you.
   space
   As it turns out, one of the funniest sights I ever saw in Malaysia was a local coffee seller on a motorbike selling coffee by the road for RM1.50, right outside Starbucks. That is one smart guy.

Further Application: Banking

I found point number 5 (allowing consumers to compare) to be of particular interest, so I opened my eyes a little wider and had a look around. I found this:

This is a digital cutout of St Andrew’s Street in Cambridge.
Barclays, HSBC and Santander are literally right across / next to each other.
In addition, if you have a look at the map (click here), you’ll see TSB, NatWest and Lloyds on the same street as well.

All the logic we applied to coffeehouses can be applied here, especially the fact that they’ve collectively turned St Andrew’s Street into the area to go for banking (this makes sense, since this is the heart of town in Cambridge).

However, the earlier point I made about locating yourself near your competitors to allow consumers to compare becomes particularly important.

Coffee is a small purchase (despite its huge importance in the daily functioning of the human race). The average consumer is unlikely to be immensely concerned about getting the best possible deal every time they buy a cuppa.

With banking, it’s a different story. When opening a bank account, you want to get the best possible interest rate on your savings, the best perks, the lowest risk, and so on. It’s also a decision of a rather large magnitude. Furthermore, some banks tailor themselves specifically to students (for instance, allowing larger overdrafts), while others appeal to those seeking the best mortgage deal (for instance, offering lower interest rates).

If all of them weren’t located in the same area, a consumer would be unlikely to be confident of making a decision on the day. They’d want to check other banks out, do their research and make sure they’re not being hoodwinked (the general suspicion of bankers doesn’t help either). By locating themselves in close proximity to each other, they’ve brought the comparison to the consumer’s door. “Feel free to compare”, they are saying. We’re all right here. There is no excuse for you to hold off making your decision today. This increases business for everyone collectively, because complete information is transparently presented to the consumer.

Wrapping Up

The conclusions derived from Hotelling’s model are often stunningly echoed in reality. However, the reasons for it are many, and run far deeper than a simple battle of location. I’ve discussed banking and coffeehouses; an exploration of a different industry would likely yield some further unique conclusions, while perhaps sharing the same fundamental logic.

So what does all this mean for the average firm?

Keep your allies close, and your competitors closest.