subscription services

Everyday life is filled with examples of subscription services.  They range from newspapers and magazines, where one pays in advance for copies that are delivered over, say, the subsequent year; to monitoring services that guard against burglary or fire; to cellphones, where the network operator offers a handset at a subsidized price in return for the customer signing a long-term contract; to cloud computing, where a customer “rents” storage space or other hardware, or software tools to run his enterprise.

All these kinds of companies have common characteristics.  Apart from the cost of setting up or participating in a delivery system (from coaxial/fiber optic cables to the postal or telephone service), the key variables are:

–the number of customers

–changes in that number as time progresses

–per customer revenue

–per customer operating costs

–customer acquisition costs, and

–the length of time the average customer retains the service.

These are the bare bones.  Of course, there can be other considerations, like a company’s ability to sell add-on services after the initial customer relationship is established, or the fact of general, administrative and (possibly) financing costs.  But let’s put them to the side.

my point

The point I want to make in this post is that these companies sometimes exhibit earnings patterns that equity markets find difficult to understand and value.  In some cases, this has meant that companies are ultimately taken private after their stocks have languished in price in the public markets for an extended period of time.

An example:

Consider a company that provides burglar and fire alarm monitoring to residential customers.  Typically, the firm will offer “free” installation of monitoring equipment in return for a two-year monitoring contract.

Let’s say installation expenses are 300, that the customer pays 20 per month in fees and that the average customer remains with the monitoring company for a long as he owns his house.  Assume that’s 10 years–but it could be a lot longer.  Let’s also assume that the cost of setting up the remote monitoring station is trivial, but that manning it costs 100,000 a year.

the company take on its business

The company probably does a present value calculation to evaluate how much it gains by adding a customer.  Ten years of revenues at 240 per year = 2400.  Subtract installation costs of 300 and the customer’s share of monitoring costs, say, 250.  Then the net value of a new addition is 1850.  Present value is lower, but the possibility of rate increases and operating leverage in expenses mitigates this to some degree.  Yes, I could have done a “real” calculation on a spreadsheet that would be much more sophisticated (though perhaps not much more accurate), but this is the basic idea.

the stock market’s view

Here’s what the income statement for the first five years of such a company’s existence might look like:

 

	year
	1	2	3	4	5
new subs%		50%	50%	20%	10%
total subs	1000	1500	2250	2700	2970
total revs	240000	360000	540000	648000	712800
op costs	-100000	-100000	-100000	-100000	-100000
startup cost	-300000	-150000	-225000	-135000	-81000
net profit	-160000	110000	215000	413000	531800

In year 1, the company is unprofitable, even though on a present value  or “asset” basis it has added 1,850,000 in value.

In year 2, the company becomes profitable on a financial reporting basis, but still has negative net worth.

In year 3, earnings explode, even though the firm is adding less asset value than it did in year 1.

Year 4 is the really interesting one.  Reported earnings continue to rise at an astronomical clip.  Yes, profits are only up 92%, vs 94% in the year earlier.  But is this something to really be concerned about?

Actually, yes.  The concern isn’t about profits but about revenues.  In year 4, subscriber additions show a sharp drop, from 750 in the year prior to 450 in the current period.  There are two reasons the earnings are still so strong, and don’t reflect this falloff:  lower expense for new installations (startup costs) and positive operating leverage from monitoring costs being spread over a larger number of customers.

How does the stock market treat a case like this?  In my experience, the answer is “badly.”  Investors are accustomed to looking at earnings per share or at cash flow per share and this kind of company doesn’t fit either template.  While the company is expanding rapidly, the costs of linking up new customers depresses eps, and cash flow may be negative.  Paradoxically, the profit numbers look their best only when the firm begins to show signs of maturing.  But investors will begin to take fright when they see that revenue growth is slowing.

This situation is a big reason that most monitoring companies have either been taken private or are divisions of larger companies, where the unusual earnings pattern isn’t so evident.

One other observation.

This concerns accounting technique.  In the example above, the installation costs have been expensed in the year incurred.  What would the financials look like if those costs had been capitalized and depreciated over ten years.  Take a look.

 

	year
	1	2	3	4	5
new subs%		50%	50%	20%	10%
total subs	1000	1500	2250	2700	2970
total revs	240000	360000	540000	648000	712800
op costs	-100000	-100000	-100000	-100000	-100000
startup cost	-30000	-45000	-67500	-81000	-89100
net profit	110000	215000	372500	467000	523700

In the first four years, the company now looks a lot more profitable and cash flow looks better.  In other words, the monitoring company looks like a conventional firm that equity investors would have no trouble evaluating.  Expense deferral only starts to catch up with the company in year 5, when the growth rate drops off significantly.

why expense instead of capitalize/depreciate?

For one thing, expensing is the more conservative technique.  For another, in the case of a monitoring company, there’s no capital equipment.  The sensors being installed are all low-cost items that are normally expensed.  Labor cost is probably the biggest factor in the installation.

relevance for cloud computing?

As this industry develops, it will be important, I think, to distinguish between companies that rent hardware (which can be depreciated) and those that rent software (whose costs may be expensed as R&D).  Their income statements may look very different, as the monitoring case illustrates.

There may also be wide company to company differences in accounting technique for basically the same services.  More speculative firms may capitalize all the customer acquisition costs they can–and maybe some that they aren’t supposed to.  Others may have a much more conservative bent.  It’s not clear that brokerage house analysts will appreciate the differences, or flag them in their reports.

In addition, there may be firms whose financials will mimic those of the security monitoring industry.  Absent considerable shareholder education, such firms may have less positive experience for their stocks than the company performance merits.