I don't have a problem with sizing the market for a general product, like the US coffee market or the US mattress market, but I'm having trouble estimating the # of specific stores, for example, this question.
I've tried breaking it down, but I keep getting stumped. I'll write out what I've thought of so far.
Since coffee is the primary product (at least in the US) I thought of doing:
(# of daily coffee drinkers in US) / (# of customers/store/day)
Obviously, I would break each of the branches down much, much more, but if I were to make a tree, those would be the first branches. However, there are a few issues with this. The demand for coffee in general, isn't what drives the creation of a Sbux store, right? Or is it? I'm not sure. Second, while coffee might be the primary product, they have so many other types of drinks that this wouldn't account for.
To make it more specific, I thought of doing:
(# of people in US who go to a Starbucks each day) / (# of customers/store/day)
Once again, I would break these branches down way more, but these would be the first branches. This doesn't seem right either. It seems like the first component would call for too many assumptions. If I were to break that component down further, I would probably look at the # of coffee drinkers → (US population) (% who drink coffee), I'd also make sure to focus on the relevant population, then those who drink coffee regularly. I'd also consider the # who drink coffee at home v. purchase elsewhere, then I'd consider % who go to Starbucks as opposed to other large suppliers. Overall, however, this is just not a good line of reasoning. I think it calls for far too many assumptions.
I'm really hoping someone can walk me through a full problem breakdown because, as you can see, I've clearly tried to think this out on my own. I've only seen vague answers so far, and they have not been helpful. I have an interview very soon, so I would really appreciate some assistance.
Could you please explain the rationale behind the formula: # starbucks = Number of Coffees consumed per day / Coffee Shop capacity per day * Starbucks market share.
The sticking point for me is the coffee shop capacity element. My understanding is that this assumes that coffee shops all operate at the same capacity (no matter if it’s a chain or a small independent coffee) and that they all continuously operate at full capacity. In my view, this leads to low accuracy of the estimation, but I’d love to understand better the thought process.
Thanks!
1. My assumption was that starbucks average store is no different than the market average store. Of course a different assumption could be made.
But I would rather suggest changing conceptually the ORDER in which I did the estimation. So now, I would instead estimate first starbucks value share (total consumption * starbucks market share) and then divide by the average starbucks store capacity. This would be the least complex way to do the estimation.
2. Alternatively you could use two segments to estimate market capacity (smaller independent stores * % of that type of store + large multi-unit chains * % of that type of store).
But then you could not multiply this by Starbucks market share but instead you have to use by Starbucks % of total stores (i.e. 20% of revenue does not translate into 20% of stores). While this is a legitimate way of doing the estimation, it is more complex and provides no advantage vs. the other approach