Red Hat-A primer in how not to paint a still life
Red Hat (NYSE:RHT) reported the results of its fiscal Q3 Wednesday evening. The balance of the next few paragraphs may be the only non-controversial components of this article. Just for the record, reported revenues were up 18% but still missed the consensus by 0.5%. Earnings were $.61; $.03 above the consensus. The CFO, highly regarded industry veteran Frank Calderoni, announced his resignation to take a position as CEO of another company at the end of January. His departure is a loss for Red Hat and inevitably has led to an investor loss of confidence reflected in the share price battering today.
The share price fall was primarily a reaction to the company's 13% growth in calculated bookings which had grown by 19% and 16% the prior two quarters. Much of the balance of this article is related to an analysis of what happened with the bookings metric. I think that my conclusion is that the miss was far below what the share price reaction reflects. The subject occupied essentially all of the conference call and management was unable to overcome analyst skepticism and to an extent was unable to even make its point given the …chagrin, perhaps that analysts felt about the miss. It might be nice if I could write about the specifics of the business-but in this case I think that readers might want another perspective as to what are the relevant factors to consider in evaluating the share price earthquake. And that has all to do about bookings.
Operating cash flow (CFFO) declined by 8% year on year. The decline in CFFO was a product of both balance sheet items and a smaller increase in deferred revenues. Overall, however, deferred revenues grew 15% year on year.
Revenue expectations were pegged at $614-$622 million for this current quarter with non-GAAP EPS estimated to be about $.61. The revenue guidance has been impacted by $16 million or 2.5% because of the strength of the USD. This is something that almost all tech companies are going to face as a headwind going forward. Net of the impact of currency, management reduced its revenue guidance compared to the prior consensus by $4 million or about 0.7%. Perhaps the use of the word "disappointing" to describe such a finely grained forecast might be a bit overdone?
CFFO was forecast to be $305-$320 million for the quarter, which represents a seasonal high point for that metric. Through 9 months CFFO has only been $465 million, down about 8% from the prior year for the same reasons that depressed CFFO in the third quarter. The Q4 CFFO guide represents growth of 22%-28% from the same period in the prior fiscal year although CFFO will still be marginally down for the year as a whole.
At the end of the day, management said it anticipates that billings growth is likely to return to more "normal" levels in the mid-teens per cent range this current quarter. Bookings, deferred revenue and CFFO are all interrelated. The company doesn't normally forecast bookings growth. To the extent it has done, it seems that its visibility is better and its outlook is more positive than might be adduced from some review of analyst commentary.
The strange love-hate affair analysts seem to have with Red Hat
The problems that Red Hat encountered last quarter are neither particularly unusual or necessarily serious in the software world. When most software companies reported license revenue and that was the metric on which to focus, it was a rare company indeed that could go for an extended period without issues with its forecast. Sales of perpetual licenses have always been lumpy and if that was the way users were consuming software these days, they would be just as lumpy now as in the past.
Bookings in their own way represent a commitment by users that is the same as or greater than perpetual license sales. As a result, it is hard to forecast bookings accurately on a quarterly basis and given that bookings can relate to both the absolute volume of software consumed and the length of the commitment to that software, they are going to be even lumpier and harder to forecast. As the CEO of this company said, "Karl, so as you mentioned, we don't guide billings, primarily because it's a challenging and you can see based on what you just said, there is a lot of volatility. And because of that, it's hard for us to actually predict that (billings) as if we do get more information over time and we start to see some kind of a trend, we'll take that in consideration and use that to try to forecast. But at this point, we feel that probably it wouldn't be appropriate to provide a range from guidance, or even kind of state what would be considered a trend." Put a different way, it is probably better not to try to do something that cannot be done accurately on a consistent basis.
I do not think the above is any deep revelation of an unknown truth. Red Hat shares had an average rating of buy prior to the earnings release on Wednesday evening. There had been 15 buy ratings, 8 strong buys, 7 holds and a sell. 2 analysts have lowered their ratings in the wake of the earnings release and just about every analyst covering the name has lowered his/her price target.
Many analysts do not like uncertainty. And yet a metric like Red Hat's bookings simply has to be uncertain. It makes little or no sense to attempt to draw trends based on the experience of one quarter. I have absolutely no reason to believe that management's forecast of a rebound in bookings next quarter is wrong or that management would ever be able to consistently and accurately forecast bookings. Changing price targets or ratings based on what happened at Red Hat last quarter is close to intellectually indefensible. The fact that the company didn't close $20 million of Federal deals this past quarter is unfortunate. It is also not a very good rationale for thinking that Red Hat shares are worth 15% less now than they were before that shortfall was revealed. The story really hasn't changed. The share price has and that is the opportunity to consider.
A short excursion into the world of software bookings
You say you would rather go to St. Bart's. So too would your writer. Heck, I will even get on a plane and go to Hawaii-well if I can go Business Class. But for now readers, all I can offer is software bookings and a better understanding of Red Hat's quarter as a destination. No warm sun or soft sand or…oh well, no more day dreams.
Most analysts who cover Red Hat, and indeed most analysts who cover subscription software companies, have learned that bookings are the metric on which to focus. Right-well sort of. Bookings are the metric on which to focus. But bookings are not billings. Again, for the record, calculated billings for the quarter including billings that become part of the deferred revenue balance came in at about $680 million. The consensus forecast for that metric was probably about $700 million. So that looks to be a 3% miss.
In the wake of that miss, the shares have declined 13%-14% in yesterday's trading. Well they missed a derived bookings proxy, the shares are expensive, and the market seems…toppy. Is that real or is that an issue with the derived bookings proxy? I think it is almost certainly the latter. I will discuss the subject in some detail below; it apparently was something not readily appreciated by the listeners to last evening's conference call.
Investors have voted with their wallets to flee the name since the earnings were released. I think that is a mistake. I think the shares represent an excellent value at this level with significant upside opportunity. I think that the valuation has compressed to a point where investors are not overly dependent on a hyper growth story to anticipate a return.
On the demand side, the company missed expectations in its Federal Government sector because of the failure of Congress to agree to a continuing resolution before the end of October. Management claims that the shortfall will be made up in subsequent quarters. I have no reason to believe that the missing $20 million won't be booked this quarter or in subsequent quarters. It is hardly a unique event in the annals of Federal procurement of software capability.
Is demand growth decelerating for Red Hat? I do not think it is likely and I think the reaction of analysts and investors is more than a little misguided. If Red Hat were a still life, the picture would be one in which the individual fruits or flowers or shells all looked as they should but were perhaps a bit juxtaposed. If readers can see the picture in focus, then the conclusion they will reach is something far different than is depicted by today's share price action.
What is the difference between bookings and billings?
It really isn't meant to be a trick question for use on some quiz show. As subscription software companies developed in the past decade or so, it became apparent to many observers that just looking at revenues and revenue growth was an inadequate analytic tool for determining their success on a quarterly basis. Software companies that earn their revenue on a subscription basis record the revenue ratably and not up front so it made more sense to look at the value of the contracts they signed in a given period as opposed to the revenue figure that was actually reported. Most software companies that sell their products using subscription contracts provide metrics including the change in long and short term deferred revenue as well as revenue reported in accordance with GAAP. It didn't take long for most analysts to decide to use a calculated billings proxy as the most accurate tool in terms of forecasting the business performance of a company. Most analysts and investors have come to realize that the number that they ought to focus on is bookings and not reported revenues.
There has been a limit as to just how far this realization has extended. For the most part, while subscription software companies record their income ratably, they record their costs as they are incurred, although there has been a trend lately for companies to recognize commissions ratably on the same basis as revenues. While many readers have accepted that bookings are a better metric than reported sales in attempting to evaluate growth and relative worth, most observers have baled about using a bookings number to put together economic earnings. I do not propose to get into a lengthy discussion of the set of arguments in this article but I do think it is a consideration in appropriately analyzing relative valuation.
So, bookings have become acceptable as a valuation metric. What's the problem? The problem is that what Red Hat missed were billings and not bookings. In point of fact, it seems likely that all in, bookings ran at, or ahead of prior expectations while billings did miss. The problem in terms of evaluating the Red Hat is that not all bookings are created equal. In some cases, and indeed most of the time, a booking creates a billing; if the billing is for services in the current period it is revenue, if it is for service beyond the current period but within 12 months, it is short term deferred revenue. If the service is paid for but delivery goes beyond 1 year, the portion beyond one year is counted as long term deferred revenues. And if the service is contracted for, but not paid, the amount of the contract... can disappear into the ether. It can be called unbilled backlog and some companies, such as Salesforce (NYSE:CRM), release the metric quarterly. And this company releases the metric but once a year and this was not that quarter.
As subscription licensing has become more and more popular in the software world, users have been willing to make very long commitments to vendors. These contacts go out far beyond the usual 12-month time span of a deferred revenue. Depending on size, users are willing to commit to 7 year deals and Salesforce has reported several deals in excess of 9 figures in which the upfront revenue contribution has been negligible but in which the commitment is completely contractual. In the case of RHT this past quarter, it received orders for two deals in excess of $20 million.
At one point, it was fairly typical for users to pay up front for multi-year deals but as the size of deals has increased, the tendency has been for users to pay annually as a component of their operating budgets. This is a broad-based trend across the software industry that far transcends Red Hat or any other particular company. As I commented about a few weeks ago, in an article on WDAY, the issue is that fewer users are paying the cash value of their entire commitment up-front and company's report bookings based solely on cash and not on contractual commitments.
In this past quarter, the two large customers who made large commitments to RHT did so based on multi-year non-cancellable agreements, but only paid for the 1 st year in advance. The additional value of those contracts was said to be $27 million. The $27 million is certainly not lost and while it is a booking, it is not a billing. So, while Red Hat missed its billings expectation by about 3% it did not apparently miss its bookings target. To be fair, it did miss about $20 million of bookings/billings in its Federal vertical (almost no deals in Federal are for greater than a single year based on procurement policies); again, management says that these missed orders will close this quarter and given the nature of federal procurement cycles, I would not be surprised to see that happen.
Management did not guide down because it "missed" a proxy billing target. It "guided down" because FX is going to pummel its revenues this quarter as will be the case for many multi-national companies. Management kept EPS guidance intact for current quarter despite FX headwinds. Presumably costs denominated in dollars will show a similar benefit to RHT to offset the negative impact of FX on revenues. Management increased its target for CFFO.
This was not a perfect quarter at Red Hat. Neither was it much of a disaster. Linearity could have been better, no doubt, but it part of the price that companies pay when they hunt down very large transactions. This company chooses to provide data regarding unbooked backlog once a year-in point of fact when it reports fiscal Q4 results which will be in March. The company is seeing more of its business come from its cloud application development and management tools, than from sales or RHEL subscriptions. What the impact of that transition may have been last quarter is not precisely known.
Most companies in this space would be far better off if they provided investors with metrics centered around annual recurring revenue growth rather than around either bookings or billings. It is hard to mess up the meaning of ARR. It is a number that can be readily compared year after year and between companies without explanation. I have no way of calibrating what the ARR growth of this company has been. The annual contract length for this company is 24 months and has been relatively constant for some time now. I have to assume that the two large deals, that closed, but which had a small impact on calculated bookings, would had a significant impact on ARR if that metric were reported. Obviously, $27 million would have had a significant impact on a metric such as unbilled backlog.
Even Paranoids Have Enemies
Who hasn't heard that expression. And the fact is that every time anyone looks at Red Hat the question will arise as to how the company is dealing with the cloud. I initially wrote about Red Hat on this site about 8 months ago, and recommended the shares then because of its strategy of offering tools to users who were migrating to the hybrid cloud. Basically, the company's growth going forward is a function of how successful it will be in selling its service called OpenStack. I thought then that the strategy would be successful. It "grew" bookings by between 32%-33% last quarter and presumably it saw the greatest percentage impact from the large, multi-year deals that weren't recognized.
I do not think that there was any great revelation in this quarter with regards to the competitive progress or lack thereof with regards to OpenStack and OpenShift, Red Hat's Docker based cloud development tool. The company sold two large enterprise deals of that tool last quarter including one worth more than $10 million. The company's overall large deal count rose by 20%, it had a strong renewals quarter, and the argument about the relevance of OpenStack and OpenShift would appear to be less relevant given that 7 of the company's 30 largest deals included the former and 5 included the latter.
Some Comic Relief or lessons in professional stock analysis
The Mikado was a Gilbert and Sullivan operetta of the late 19th century. It occasionally is revived on Broadway and far more often is performed by a high school or college theater club. There is a well-known song in the operetta called "I have a little list." The song was about things that griped the author about life in late Victorian Britain. We all have our own little lists. On mine, is visibly sloppy analysis. Perhaps some readers will fault my definition. But regardless, it is these kinds of reports that make for opportunities. As to the individuals involved-well I have no intention of finding them via tweet or elsewise.
One analyst, who really ought to know better, wrote in a note that "the reasoning was somewhat difficult to follow and adds little to explain the weak fiscal Q4 guidance other than assuming another more back-end loaded quarter than usual." Jim Whitehurst and Frank Calderoni actually do not speak in tongues and I would be embarrassed to try to have to defend those remarks to any reasonably knowledge able institutional investor. If the analyst in question doesn't understand the impact of FX after all this time in the business…well other observers are less challenged. If Q4 guidance, down 0.5% from prior expectations is weak, again perhaps we need to define words more carefully.
Some other commentators had what might be charitably labeled idiosyncratic logic progressions in their thesis. One analyst wrote that "Even with improved billings growth in Q4 from the closing of delayed deals, we believe that questions on billings growth in 2018 will keep the stock range bound." Again, the perspective for all this is breathtaking. The company booked two large deals and the buyers only paid the first year upfront in cash and that means that investors are going to believe that questions regarding billings growth in the next fiscal year will keep the stock range bound. It may be that shares will be range bound-but investors will somehow determine just how fast this company is growing and price the shares accordingly. If the billings proxy rebounds, they will think it is growing faster and reward the shares...and if not-well that is a different consideration. It is hard for me that investors are going to be particularly focused on a metric that the CEO says is one he cannot forecast successfully.
I feel compelled to close this section by quoting briefly from the question of a well known analyst who has been in her position a long time. "I am just, I guess, trying to understand this a little bit. We're talking about people are driving up the long term, and people paying less upfront. But if we really just focus on your short-term billings…you still keep seeing a pretty big deceleration in growth in short-term billings… So, can you help us reconcile all the talk about customers paying you less upfront with what's being indicated by the short-term billings trends, because maybe I just don't understand this?"
I just never knew that making a living my listening to and reporting about conference calls was all that difficult. Not surprisingly, this analyst had a hold rating on the stock. I really find it difficult to credit that this was a hard explanation to understand. Analysts can say, they wish, that they do not believe management. Many times I have implied that. But this is not some complex scientific problem.
There just aren't all that many moving parts. A definition of bookings that is inappropriate to the analytical task at hand, a tendency to larger deals-a trend on-going for several quarters now, a concomitant trend to bank-end loaded quarters and a miss of $20 million on Federal.
I lived in North Carolina for several years and I confess that some of the wonderful people I met in the state, particularly those who were raised in the Inner and Outer Banks are a bit difficult for a Yankee like myself to understand. (Red Hat is headquartered in Raleigh, NC) But honestly, this explanation was about as straightforward and credible as any other explanation one gets when a company misses an operational target. One suspects that in the fullness of time, or perhaps in a much shorter period than that, cooler heads will prevail and analysts will attempt to evaluate the good and the bad on a far more nuanced scale than the above comments represent.
The last time I wrote about Red Hat I was concerned about the valuation of the shares. Not that stopped me from recommending them-the company had just printed a significant upside quarter and all seemed well. I mused about finding the right entry point.
The market has done the work of finding a much better entry point with less stretched valuations. The investment is far more controversial today than it was a few months ago, when the shares made their all-time high. The upside is significantly greater. And as I have argued, perhaps in a less than collegial spirit, very little has really changed. At this morning's share prices, Red Hat a market cap of about $12.7 billion. It has net cash of about $1.2 billion yielding an enterprise value of $11.5 billion.
The company is forecasting total revenue for the current fiscal year of $2.4 billion. Looking at fiscal 2018, which starts in just a bit more than 2 months, consensus estimates for revenue had been $2.75 billion. The consensus will almost certainly decline, if for no other reason than to take account of currency fluctuation. But I will leave that number in my valuation calculations as I think the projected organic revenue growth expectation had been too low to start with.
The company actually increased its full year EPS forecast to take account of the beat in Q3 and it is now $2.27. Next year's forecast for EPS had been $2.60. I imagine it will continue to be at that level.
The company has decreased its CFFO guidance to about $780 million for the year. Some of that is due to FX headwinds; some of it has to do with a slower cadence of deferred revenue growth. I think a reasonable forecast for CFFO next year is about $900 million taking into account some of the trends that were evident both this quarter and throughout the fiscal year. Capex is a small expenditure for this company and is running at about an $75 million annual rate. So, I will estimate free cash flow at $825 million for fiscal year '18, recognizing that it is one of the more difficult metrics to properly gauge.
So, the typical valuation metrics are an EV/S of 4.2X, a P/E of 26.5X and a free cash flow yield of 7.2%. None of these, except perhaps the free cash flow yield, is deep value. On the other hand, they are all far lower than similar valuation metrics at which RHT shares have traded in the recent past. One observer suggested that the valuations had approached the trough seen back in the financial panic.
How fast can this company grow? I have made the point in the past that the company's cloud offerings have a large TAM that is barely penetrated. Like its rival VMware (NYSE:VMW), it has made a significant bet on developing a portfolio to assist users in their transition to the hybrid cloud. That is still there strategy. If readers buy the shares, that are betting that hybrid cloud deployment will continue to grow strongly.
Enterprise Linux is a mature product that continues to find new use cases and has enjoyed something of a second life with the availability of solutions that facilitate the use of Red Hat Linux on data bases that are offered by Microsoft and Amazon. This is a small but fast growing revenue stream.
I think a reasonable case can be and has been made that the company, as currently constituted can enjoy years of growth in the mid-teens and that there is a reasonable opportunity for the company to average 20% growth. This is not a margin story and I have no reason to think it will become one at any point in the near future. Still the company spends 43% of its revenues on sales and marketing on a GAAP basis and 39% of revenues on a non-GAAP basis. There is plenty of runway in terms of margins should the company not see the current magnitude of growth opportunities.
This is near the season when some intrepid souls jump in the surf for a refreshing swim in icy waters. I think investors needn't be anywhere near as intrepid to jump in the Red Hat pool. The water is far more comfortable than yesterday's trading action suggested.
Thanks for reading if you have stuck with the article thus far and my very best wishes for a happy, healthy and safe holiday season!
Disclosure: I am/we are long RHT.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.