The best business model in the world

But if Carlsberg did make business models...

As I wrote yesterday, IMHO running a restaurant is probably the worst business model in the world. Which begs the question, what is the best?

Enterprise software.

Oracle, SAP and the boring bits of Microsoft rarely set the pulses racing. While Apple make's sexy phones and Intel whiz-bang chips, these guys basically make accounting software for big companies. But for an investor this is a thing of beauty.

Like the cigarette business, except it doesn't kill its customers

Here's an ideal enterprise software business model in a nutshell. You sell a licence to use your software for a million bucks (cash up front), you then go in and install it and charge another million bucks for the service. Them every year you charge the customer a $200,000 "maintenance fee" in exchange for regulatory updates and bug fixes. On that you make an 80-90% gross margin, and the rest of the opex base (split roughly evenly between R&D, sales & marketing and admin cost) lets you mint an operating margin of 20-30%.

And here's why its so great:

• Asset light:  Upfront capex requirements are close to zero. Unlike a manufacturer which needs to build a new factory before they can even contemplate doing business, software companies require little more than an office, some computers and a phone line. 
• Cash generative: Software companies are scarily good at converting accounting profits directly to cash in the bank, because there is very little cash tied up in working capital. In fact if you collect maintenance fees upfront, you can even have negative working capital - your customers are paying you to sell them stuff!
• Scalable: Software has virtually zero marginal cost of production. Want to create twice as many copies of your software? Simply generate twice as many licence keys and point your customers to the download site.
• No exposure to input costs: If you don't manufacture physical goods you have no suppliers and no exposure to fluctuating input pricing.
• Flexible cost base: Normally the downside of high gross margin (i.e. low cost of production) businesses is a high fixed cost base which means small a small sales decline can wreck your margins. Software companies have this covered too. Sales and marketing staff (who's costs amount to about a quarter of revenues) are paid on a minimal salary+bonus basis. This means when times are hard you can almost instantly by slashing discretionary bonuses. Sure the sales guys aren't pleased but for them its better than a P45, and you can have a big cost buffer to protect your margins.
• Customers never leave: Enterprise software is like the Hotel California - You can never leave. Once a large company has gotten stuck into a complex accounting system like SAP or Oracle the sheer institutional inertia and hassle to involved in ripping it out, junking the historical investment, and retraining everyone makes it virtually impossible to switch suppliers. Which makes the price negotiations very relaxing - for the software vendor.
• Recurring maintenance revenues: Remember that $200k maintenance fee I mentioned earlier? That's money for old rope. In reality the company gets very little for the cash apart from the right to use something they've already paid for. And remember even if a software company isn't growing its licence revenues, if they are selling any licences at all the installed base of recurring high-margin maintenance revenues will carry on growing.
• Eco-system: All good bullies have their sidekicks. In this case an eco-system of professional services companies like Accenture and IBM who's job is to sell even more services around your enterprise software installation. This means a) they make money, b) they do a lot of the sales & marketing legwork for the software company and c) customers become even more entangled in a spaghetti of legacy IT systems which makes it even harder to leave.

Warren Buffet loves the cigarette business because "It costs a penny to make. Sell it for a dollar. It's addictive. And there's fantastic brand loyalty."

In short, software has all of that, plus the added advantage that it doesn't kill its customers.

And if you don't believe me, here is SAP's revenue and profit growth over the last decade or so. OK there's been a couple of acquisitions (Business Objects and Sybase), but the overall trend is clear. Note also how resilient profits are even when licence revenues are falling - that's due to that huge blue-shaded block of cash-rich maintenance revenues which grows at a steady pace, rain or shine.

As I said, probably the best business model in the world.

PS - So where's the catch?

I'm aware this post is a bit of a love letter to the enterprise software business. But where's the catch? Well nobody's perfect, so here's a few:

• Lack of visibility: Unlike manufacturing industries there is little supply chain or end-market data to analyse, so forecasting software licences on a period-by-period basis can be quite difficult (although forecasting recurring maintenance is very easy). Not an issue for a true long-term investor but this means software companies can be volatile on a quarter-by-quarter basis.
• Competitive market: Precisely because the software model is so good (and generates such high ROIC), the industry tends to attract competition and new entrants. For the industry behemoths like Oracle or SAP this isn't a huge issue, but for smaller companies it means competition can be brutal.
• Piracy: Because software is easy to make its easy to pirate. Again for complex large-scale enterprise software this is not a big issue given the complexity means you can't just boot up and run pirate code, but this can be a challenge for simpler packages.
• Maturity risk: Software companies tend to de-rate as they get larger (base effect - harder to sustain growth) and maintenance grows to a larger proportion of revenues. Actually from the point of a fundamental investor this is completely nuts - if companies are deriving a larger proportion of their revenues from high margin recurring maintenance streams their stocks should up-rate, not de-rate. However tech investors tend to be obsessed by growth-at-all-costs (even at the expense of profits) so mark down software companies which have substituted licence growth for maintenance growth. Go fig.
• Platform risk: You could be the best damn company at making widgets which run on desktop PCs, but if all your customers switch to iPhones then you have platform risk. This rarely happens but when customers shift the device on which they consume their IT, AND you haven't kept up with them you run the risk of having your legs cut out from under you. (a good example of this is WordPerfect which had the dominant word processing software running on MS-DOS but completely missed the shift to Windows. In short they were late, and Microsoft Word won).