30 Mar 2018

Ten years ago, I was paid to write research on investing in software companies. My USP was that I was pretty much a technophobe with little or no interest in software but with something of a passion for finding how to make money by investing in companies. Back in the early 2000s the world of software was full of exclusive jargon which, deliberately or not, served to discourage scepticism. The following is selected more or less at random from the 2006 Annual Report of SAP AG.


Today, the IT sector stands on the verge of widespread adoption of service-oriented architecture (SOA), a development that promises to change the dynamics of the software industry as much as the shift to client-server architecture did 15 years ago. In essence, SOA defines the technical standards that enable the various enterprise software applications used by companies and their business partners to exchange data effectively. Thus, SOA will help reduce the costs of creating and maintaining data exchange interfaces, a factor CIOs consistently cite as one of their top challenges.

When I read stuff like this I used to think a) does this mean anything and b) even if it does, how is anyone going to persuade the board of a major company to throw money at it? That is why I hit upon a personal rule that can be usefully extended from the narrow world of software analysis to life itself – never be afraid to ask the stupid question. Experience teaches that it’s often the hardest one to answer.

Fifteen years ago the normal software business model was to sell a perpetual licence that allowed the customer to use the software plus an annually payable fee that got him maintenance and upgrades. The analyst community was obsessed with the growth of software companies and consequently watched the “new licence sales” number in every quarterly earnings release.

Share prices were highly volatile as a result. A “miss” from the quarterly numbers often resulted in carnage for the shares. I was actually frightened. How I could I make recommendations to buy and sell shares when my view could and probably would be rendered farcical every three months? Even worse, the companies themselves were more anxious than I was and they heaped pressure on their own sales teams.

Goodhart’s Law states that when a measurement becomes a target it starts to misbehave. So it was with quarterly licence sales. The customers began to realise that delaying their orders until the last days of the quarter allowed them to negotiate price discounts. So quarterly revenues became “back-end loaded” meaning that, with two weeks to go, the company management would have no way of telling if the sales pipeline would convert into actual sales. What was a poor dumb analyst to do?  

And then I noticed what the smartest guy in the room was up to.

Larry Ellison was the founder and CEO of Oracle Corporation, a vendor of database and software for enterprise customers. A couple of fashionable software companies, Peoplesoft and Siebel, had seen their shares dumped after growth in new licence sales disappointed. And then, in 2004 and 2006, Oracle turned up with takeover offers. What was going on? Why did Larry Ellison want these apparently fading businesses?

The answer was that the wanted the annual maintenance fees from the existing customer bases. Ellison understood that the value of a software business is found in the customers rather than in the software. Once I twigged this it was great news for me because customers are easier to understand than software. And any business that relies on selling the same thing over and over again – even if it’s software – is inferior to one that sends an annual bill for an installed service.

This has become more obvious since “cloud computing” largely displaced applications that you download and run from your own computer. It is easy to overlook the distinction now because we are so used to everything running interactively on the internet. The cloud computing model, pioneered by (which was founded by an ex-Oracle employee) hosts all of a company’s critical data and charges the company for access to it. This charge is a rolling subscription. Salesforce uses, controls, improves and sells software functionality but it does not sell software. It sells a service (known as SaaS or software-as-a-service).

Thanks to Oracle’s acquisitions I started to value all software companies on the basis of their recurring revenues. It is a very simple calculation: enterprise value divided by recurring (maintenance and subscription) revenues. On that basis Oracle paid 7.2x for Siebel and 6.2x for Peoplesoft. That gave a usable range for looking at other shares and it often proved a reliable predictor of whether companies were vulnerable to takeover, usually after they had fallen out of favour with the stock market.

In 2007, Oracle paid 6.8x for Hyperion and SAP paid 7.6x for Business Objects, both add-on businesses that came with installed client bases.

As the software giants began to release that cloud computing was a threat they naturally began to buy those companies too – viable cloud computing businesses were a limited resource and this fact was reflected in the prices. SAP paid 11.9x for SuccessFactors in 2012 and Oracle paid 12.2x for NetSuite in 2016.

As a rule of thumb I still think that recurring revenue multiples are useful for valuing software and SaaS companies.

The reason I have been thinking about this is the recent profit warning and share price halving of Micro Focus (985p).

Micro Focus is a company built up by a British businessman called Kevin Loosemore. Micro Focus has made many acquisitions over the years, mostly buying mature businesses with long-standing customers. The shares rose from c.£4 to £20 and joined the FTSE 100 as the market bought into the strategy. At which point, in 2016, Mr Loosemore and his colleagues decided to bet the company (sort of) with a transformational deal.

Micro Focus used its strong underlying cash flow to gear up and buy H-P Software. This was one company with revenues of c$1.4bn taking on another with revenues of $3.2bn that were in decline. It was a deal that could only have been contemplated by a management team with a high opinion of its own abilities.

It didn’t take long to go wrong. The merger was completed officially on 1 September 2017 and a warning was given on 19 March 2018. This mainly concerned disappointing licence sales in the acquired business and was blamed on integration “issues”. In a conference call with its lenders the company said that cash flow is still positive, that leverage is falling and that it still continues with its dividend payment policy. As one would expect from an executive like Kevin Loosemore, serious action has been taken including the replacement of the CEO.

This does not sound too bad. There is a decent probability that things will improve.

All that remains is to ask whether Micro Focus shares are now cheap enough to buy. The answer, at 6.2x maintenance revenues is, tentatively, yes. Such is the equity leverage with fairly indebted companies that the shares would rise by 25% to 1220p if that maintenance multiple rose to 7.0x – it will surely do this if the stock market’s confidence is partly restored.

Leave a Reply