While thinking about my talk at the On Demand Conference today in New York, I stumbled upon a tweet which quotes a Bloomberg piece on Baidu’s Robin Lee (the CEO):

To Li, the math is simple. While search advertising is big, the services and retail market is much bigger. The initiative will “definitely” eclipse search revenue over time, he said. In cities like Beijing, the vision is already taking shape. Click an application on your smartphone and someone will stop by your home to wash and walk your dog. A few more taps and someone will deliver groceries or medicine.

This of course sounds a lot like the on demand services that entrepreneurs are building here in the US.

The basic logic here appears to be as follows: you could either operate a search engine and then hand off a customer to a service provider collecting either nothing (organic) or an advertising fee (paid search). Or you could integrate backwards vertically and provide the entire service (or sell the good) directly. Obviously your revenues will be much bigger. Your profits will also be bigger as an absolute number, but you may *not* be more profitable.

When it comes to profitability it is important to keep in mind that this is a ratio measure of profits over capital, effectively the return on capital (lots more detail to this, such as capital structure, but good enough as a first approximation). Now whether or not you can increase profitability by vertically integrating depends on whether you are a more “profitable” owner than the alternative owners (with the base case being separate businesses). 

And the answer to that question hinges on many factors. One that weighs heavily though is the competitiveness of that market. The more competitive it is, the less likely that vertical integration will increase your profitability. Here is an interesting example: book stores and Amazon. Why did it make sense for Amazon to actually become a book retailer? Because book retail is not very competitive. And the reason it isn’t is that book stores generally don’t set their own prices which are instead determined by publishers. And each publisher has a monopoly (by virtue of copyright) of the titles it publishes. So it turns out that vertically integrating let Amazon access some of those additional economics and much of the subsequent legal battles with publishers were over just that!

Now think about something like laundry. In a city like New York you don’t have to walk more than a few blocks to find a dry cleaner. The pricing for laundry and dry cleaning is quite competitive as a result. Hence access to “rents” (excess profits) to increase overall profitability is not really a valid reason for vertical integration. The same would seem to be true for many other on demand services.

That doesn’t mean that there aren’t other valid reasons largely having to do with contractual incompleteness and the need of added coordination in an on demand system (see the Paul Joskow paper for a great summary of the economic literature on vertical integration). It is super important though to keep in mind that you are actually *buying* that increased coordination. It comes at a cost and if anything as a result reduces overall profitability and may also negatively impact motivation as I described in my post on the trade-offs between firms, markets and networks.

So as you start, build or invest in an on-demand business that is vertically integrated, make sure you think long and hard about the existing and future economics of the business. The benefits from increased coordination have to outweigh their cost, which is largely a question of how inelastic the demand is for “on demand” – if you cannot add a significant price premium over the standalone base service then you could find yourself in trouble.

While thinking about my talk at the On Demand Conference today in New York, I stumbled upon a tweet which quotes a Bloomberg piece on Baidu’s Robin Lee (the CEO):

To Li, the math is simple. While search advertising is big, the services and retail market is much bigger. The initiative will “definitely” eclipse search revenue over time, he said. In cities like Beijing, the vision is already taking shape. Click an application on your smartphone and someone will stop by your home to wash and walk your dog. A few more taps and someone will deliver groceries or medicine.

This of course sounds a lot like the on demand services that entrepreneurs are building here in the US.

The basic logic here appears to be as follows: you could either operate a search engine and then hand off a customer to a service provider collecting either nothing (organic) or an advertising fee (paid search). Or you could integrate backwards vertically and provide the entire service (or sell the good) directly. Obviously your revenues will be much bigger. Your profits will also be bigger as an absolute number, but you may *not* be more profitable.

When it comes to profitability it is important to keep in mind that this is a ratio measure of profits over capital, effectively the return on capital (lots more detail to this, such as capital structure, but good enough as a first approximation). Now whether or not you can increase profitability by vertically integrating depends on whether you are a more “profitable” owner than the alternative owners (with the base case being separate businesses). 

And the answer to that question hinges on many factors. One that weighs heavily though is the competitiveness of that market. The more competitive it is, the less likely that vertical integration will increase your profitability. Here is an interesting example: book stores and Amazon. Why did it make sense for Amazon to actually become a book retailer? Because book retail is not very competitive. And the reason it isn’t is that book stores generally don’t set their own prices which are instead determined by publishers. And each publisher has a monopoly (by virtue of copyright) of the titles it publishes. So it turns out that vertically integrating let Amazon access some of those additional economics and much of the subsequent legal battles with publishers were over just that!

Now think about something like laundry. In a city like New York you don’t have to walk more than a few blocks to find a dry cleaner. The pricing for laundry and dry cleaning is quite competitive as a result. Hence access to “rents” (excess profits) to increase overall profitability is not really a valid reason for vertical integration. The same would seem to be true for many other on demand services.

That doesn’t mean that there aren’t other valid reasons largely having to do with contractual incompleteness and the need of added coordination in an on demand system (see the Paul Joskow paper for a great summary of the economic literature on vertical integration). It is super important though to keep in mind that you are actually *buying* that increased coordination. It comes at a cost and if anything as a result reduces overall profitability and may also negatively impact motivation as I described in my post on the trade-offs between firms, markets and networks.

So as you start, build or invest in an on-demand business that is vertically integrated, make sure you think long and hard about the existing and future economics of the business. The benefits from increased coordination have to outweigh their cost, which is largely a question of how inelastic the demand is for “on demand” – if you cannot add a significant price premium over the standalone base service then you could find yourself in trouble.