The key thing to watch out for in software investing is the product cycle. By and large, software companies are not very good at managing through multiple product cycles. They tend to be born in a particular paradigm or even create that paradigm, grow fast while the paradigm is in the ascendant, churn out cashflow during the period that paradigm is dominant, and then when the paradigm is on the wane, they decline with it. Very few software companies navigate their way across to the next computing platform; they go through the traditional cycle of grief being, denial, resentment, anger, etc, but generally don’t do anything much positive until they are dragged kicking and screaming into the new era, usually by overpaying for a small acquisition they hope will transform the core business, but given the lack of enthusiasm they have for the new new thing, they generally don’t do all that well.
This is why you don’t see most any of the names that proliferated in the mainframe or minicomputer era around today. They pretty much all got acquired or died on the vine. You know this already. Mainframe software businesses have mostly been rolled up and re-re-re-re-financed by leveraged buyout sponsors, to good effect for the various retirement systems that invest in the LBOs albeit sometimes at the expense of product development for those customers still using IBM zSeries and their kin.
But what of the client-server cohort? These names are now very much under duress, in our view. If you Google “transitioning to the cloud” you will find a large number of companies born in the dedicated single-enterprise datacenter who are now struggling with the public cloud; and who, in addition, were weaned on and became addicted to the adrenaline hit of upfront perpetual licenses when now the world wants to pay on the drip. This is the perfect storm. If you’re the CEO of one of these companies, you need to achieve the following simultaneously:
- Develop a replacement core product that works natively in a multitenanted datacenter environment, delivered over the public cloud, and;
- Change your entire revenue model so that you have fundamentally lower amounts of revenue to recognize each quarter and less cash coming in each quarter too, and;
- Spend enough money on R&D and also on sales and marketing to ensure you can actually get that new product out and start selling it, and;
- Not crater earnings or revenue growth rates, lest your stock collapse, and;
- Hire top people in your sector to do all the above even though you aren’t the hot company any more and your stock is in the toilet so nobody wants your stock based comp, but you can’t pay them in cash because, increasingly, you haven’t got any to spare.
When you see the phrase, “transitioning to the cloud”, it sounds all measured and nice. But it isn’t. Those bullets above - that’s transitioning to the cloud. Also, if you’re doing it now, you’re about a decade late, at best.
The rational board of directors, if they stood back to see just how hard this would be to achieve, would simply put up a for-sale sign on the company early on, call Morgan Stanley and Goldman Sachs to go do the “strategic partnership” thing, and sell the business. But boards are often infused with founder DNA and rational they often are not. So they labor on in the hope that true cloud revenue will grow and suddenly become both GAAP and cash profitable and then the dying license business won’t matter anymore and the stock will move up and they can hire the good people and also use their stock awards to get that bigger house in Atherton now that everyone is leaving for New Zealand. But this is just hope. It is the operational equivalent of believing that #StonksOnlyGoUp.
We’ve charted the problems facing Splunk (SPLK) in a series of notes that you can find here
. Guidewire (GWRE) has the same problems
. Alteryx (AYX) too. If you ask each of them, they’ll quote you Adobe (ADBE), Autodesk (ADSK), Synopsys (SNPS) and others as evidence it can be done. And maybe it still can. But the pain of trying shows up in the stock price and our expectation is that the pain is going to linger on.
Indeed our own view is that for the most part this particular problem can be filed under “Too Difficult”, and we can just avoid the stocks. We would rather be buying the Ghost Stocks Of FinTwit, because there, pretty much all we have to do is to wait for the prices to come back. In the end, even when all the world is pretending to be interested in dividend yields and inflation hedges, everyone loves a growth name really. And the joy of watching a CrowdStrike (CRWD) or a Twilio (TWLO) or a CloudFlare (NET) take flight is truly a wonderful feeling in software investing, particularly if you have been spending your time in the 2021 Tech Wreck doing that very 2020 thing, Buying The Dip (except in 2021 finding yourself grimly wondering whether indeed this be a dip or whether it in fact be the Great Reckoning wherein your elders and betters be telling you that you really should have been buying Chevron, Boeing and AT&T all along, just like they told you every day since 2009).
So. If you are going to play with legacy stocks that don’t know they’re legacy yet - they be all Bruce Willis in The Sixth Sense - plan your strategy accordingly. If you’re trading short term, have an exit plan. Because if you’re trading long, they can let you down when earnings continue to sag; or short, they can scare the life out of you when the stock hits a historic support level on a bright green day in the market, or when earnings decline just a little less than they were supposed to and an army of sellside analysts declares the undead alive once more. For a quarter or two. Just long enough to smoke your buy-stop order, before recommencing their slide into the abyss.
Us? Right now you’ll find us with the depressed growth names over in the corner. We have a feeling we might be fashionable again in H2 this year. So now we be off to tell grandpa where to put his Chevron.
Cestrian Capital Research, Inc - 21 May 2021.
DISCLOSURE: Cestrian Capital Research, Inc staff personal account(s) hold long position(s) in BA, CRWD, NET, T, TWLO.