Portfolio Update: Highlighting Execution, Regulatory and Valuation Risks
SEE LAST PAGE OF THIS REPORT Paul Sagawa / Tejas Raut Dessai
FOR IMPORTANT DISCLOSURES 203.901.1633/.901.1634
psagawa@ / email@example.com
June 11, 2019
Portfolio Update: Highlighting Execution, Regulatory and Valuation Risks
In the first quarter after we added PVTL to our model portfolio, it dropped 51.7% turning modest portfolio outperformance to 546bp of downside vs. our primary benchmark, the tech components of the S&P and a 57bp downside vs. the broader index. PVTL management failed to invest in building appropriate sales distribution and did not adequately advise investors of its problems. This failure highlights the execution risk inherent in smaller, high growth TMT stocks, particularly in complex emerging areas like hybrid/multi-cloud platform software, where the sales process is technically sophisticated, enterprise customers have little expertise and the business stakes are high. While we remain convinced of PVTL’s long-term potential, we did not adequately consider the immediate threat. Meanwhile, regulation risk has become a major concern, with reports that the DoJ and FTC are considering investigations of GOOGL, AAPL, AMZN and FB. While the 2020 election could change things, we believe that it would be difficult under US law to win a broad case and expect any forthcoming suits to focus on narrow behaviors and for proposed remedies to be less than catastrophic. Finally, investors remain skittish on valuation risk, with SaaS application names the biggest concern. We have argued that traditional valuation metrics are of little help in evaluating tech stocks (Model Portfolio Update: What is Value in Tech?) and continue to believe that expectations for the top SaaS names (and the rest of our portfolio) are overly cautious. We are removing PVTL in favor of NTNX, which suffers from its own distribution issues amidst a clumsy business model transition, but which appears further along toward resolution and is a more likely acquisition candidate.
- Execution Risks are high in nascent, complex tech. Portfolio constituent PVTL fell 51.7% since our last update, as it failed to hire and train necessary sales reps while underestimating the time needed to close sales. NTNX, which plays in an adjacent product space, suffered a similar but less serious slip up. It is not just small companies – NVDA recently stumbled on a channel inventory glut. We note that these missteps were more marketing than technology, perhaps related to a tendency by engineering focused leadership to discount the importance of sales/distribution and the relative scarcity of high caliber sales talent. This is particularly true for in emerging product markets like hybrid cloud integration, where solutions are complex and fluid, customers inexpert with the technologies involved, and expectations for early leaders are high. While we believe that the end market is extremely attractive and that short-term problems are unlikely fatal, resolution for execution missteps often stretches beyond a quarter or two. Investors need be patient.
- Regulatory Risk unlikely to upend tech leaders. The WSJ reports the DoJ and FTC have begun prelim investigations of GOOGL, AMZN, AAPL and FB. While a Democratic sweep in 2020 could usher new anti-trust legislations that might be more threatening, the current laws focus on consumer harm – a high bar when many products are free or clearly advantageous to users. The process will take years to play out in markets that change quickly – market concentrations that might seem a problem today may be moot by the time probes turn into trial. Moreover, broad suits with those potentially catastrophic penalties are particularly difficult to win in court and the government is likely to focus on narrower complaints and less burdensome remedies. None of this yet requires adjusting investment narratives for the top digital platforms.
- Valuation risk is mostly a non-issue. Sell side estimates for tech stocks typically cluster within a much narrower range than the distribution of outcomes suggested by history. In real life, tech companies often beat even the highest estimates a year or two out (but also often fail to hit the lowest), making metrics based on forward estimates dicey at best. For fast growing companies, analysts often apply a template of decelerating growth, assuming downward inflection points that can be far off the mark. These overly pessimistic projections yield highly misleading metrics on forward sales and earnings. We believe that many high growth TMT segments – e.g. SaaS software, digital ads, e-commerce, cloud datacenter components, etc. – are very likely to exceed the consensus trajectory that has been established, and that more appropriate estimates would yield far less lurid multiples.
- SaaS consolidation could be upon us. In the past week, GOOGL has acquired privately held SaaS analytics company Looker, while CRM went even bigger with a $16B deal for DATA. This raises the ante for analytics – IBM, ORCL, ADBE, WDAY, and even analytics powerhouse MSFT may look to beef up in response. Other SaaS analytics names AYX, SPLK, TBCO, and MSTR are all up strongly on the news. We could also see M&A interest rise for other best-of-breed midcap software names, like ZEN, TEAM, NTNX, PVTL, PFPT, and many others.
- PVTL killed our model portfolio performance. We added PVTL in March, just in time to absorb a 51.7% decline which siphoned 389 bp from the performance of the 15-stock portfolio. Absent that hit, the remaining 14 names were up 668 bp on the quarter, ahead of the S&P500 by 332 bp and just behind the tech components of that index by 157 bp. Beyond PVTL, XLNX was our other major loser, off 8.8%. We expect XLNX to rally in 2H19 on renewed datacenter capex growth. Our top performers were AYX (+ 40.3%), MSFT (+18.1%), ZEN (+13.9%), ACN (+13.6%), and NOW (+12.4%). Despite our modest underperformance for the quarter, we are still up 2223 bp vs. the tech benchmark over the past 12 months.
- Replacing PVTL with NTNX. Given the serious miss and guide down from PVTL we have little confidence in management’s ability to expeditiously resolve the distribution issues that are stunting its growth. Moreover, DELL’s nearly 70% ownership of the company contributes to volatility and limits the potential for an acquisition. It is an easy decision to remove it from the portfolio despite our continued belief in the long-term potential for its products. We are much more confident of NTNX’s chances of re-establishing its growth narrative as it completes its transition to a subscription model, noting that its SaaS growth was an outstanding 110% YoY in its most recent quarter. We also see substantial potential for a take-out at premium. We are adding it, seeing the current weakness as a significant buying opportunity. We also considered TWTR and PINS, based on their unique and defensible positions in a strong digital advertising market, and NVDA, based on our expectation of 2H19 reacceleration in cloud datacenter spending.
Our Model Portfolio Performance
For the roughly 3 months since our last update on March 04, 2018, our 15-stock model portfolio rose 279 bp, lagging the tech components of the S&P500 by 546 bp and the broader index by 57 bp, our worst quarter in more than three years. The reaction to CRM’s deal for DATA reversed a fierce rally that had softened the downside from the extraordinary sell-off of Pivotal Software, which we had added in March and which declined by 51.7% in the quarter. Absent Pivotal, the portfolio only modestly underperformed, down 157 bp vs. the tech components. Pivotal fell more than 50% in one day after cutting its 2019 full year guidance as a failure to expand its direct sales force to address customer interest led to lost deals and sharply decelerating revenue growth (Exhibit 1). This collapse is a vivid reminder of the importance of management execution for companies navigating a rapidly changing TMT landscape (Exhibit 2, 3).
Exh 1: SSR TMT 15 Stock Model Portfolio Performance Since Last Update
Founder-led tech companies often have a hard engineering focus, biasing management against the importance of a strong sales organization – i.e. if the product is good enough, it will sell itself. The value of software products is very rarely THAT self-evident to would-be customers, particularly when the solution is highly complex, the strategic importance to the customer is very high, and the potential buyers lack the
Exh 2: TMT portfolio cumulative performance relative to benchmark – 5 Yr.
Exh 3: TMT portfolio cumulative performance relative to benchmark last update
sophistication to easily evaluate it (Exhibit 4). For Pivotal Software, this bias manifested in a failure to prioritize sales hiring, leaving the company shorthanded in closing viable opportunities. The customer needs a lot of patient explanation and handholding to get the sale across the line, which Pivotal is ill equipped to provide. As such, sales were lost, customers were alienated, and revenues failed to meet expectations, prompting the serious adjustment in annual guidance and a furious sell-off. Fixing the problem will not be quick. Hybrid cloud software is a rapidly emerging arena with a scarcity of sales talent with the technical chops that Pivotal demands. Hiring these highly sought-after representatives away from the myriad other rapidly growing enterprise software players will be very difficult and training them to be effective with Pivotal’s products will take time. For investors, it may mean several quarters before the growth narrative can be reestablished.
Nutanix, which competes in an adjacent hybrid cloud software segment, reports similar struggles to keep up with hiring although it does not appear to be as far behind as Pivotal (Exhibit 5). It is also suffering in the aftermath of its failure to adequately assess and communicate the financial impact of a major change in business strategy. Previously, Nutanix had sold its hybrid cloud integration platform software as a license bundled with a dedicated hardware server.
Exh 4: Highlight of differences between execution constraints for cloud businesses
Exh 5: Execution risk is weighing down mid-range platform as a service and hybrid plays as distribution becomes equally critical
Last year, management decided to phase out the hardware portion of the bundle, which was sold with no markup and limited customer flexibility, moving to a subscription model. This transition had the immediate effect of lowering revenue both for the lost zero-margin server sales and for the way subscription software spread revenue recognition over a longer period. Ultimately, this will be good for the company and for investors, as the recurring revenue of software subscriptions will be more stable, and we believe, more profitable. Still, management has done a poor job of explaining its business model shift and enumerating its effects on financial results to investors. The good news? Subscription sales revenue is up 110% YoY. The bad news? The share price has been cut in half from its February highs. Still, given the strength in bringing in new business and the continued buzz around the product we are intrigued (Exhibit 6).
Nvidia is another growth company licking its self-inflicted wounds. In the first half of 2018, at the tail end of a bout of crypto-mining mania, ramped production of its PC graphics cards to fill the shelves at retailers, who had been marking up nearly 100% in the face of furious consumer demand. Since the cards had been in such short supply, many stores doubled their orders once the new capacity came online. Unfortunately, the price of crypto currencies tanked, and consumer demand dried up, leaving dealer inventories stuffed to the rafters with extra graphics cards. Nvidia management finally figured it out and slammed on the brakes after 3Q pre-announcement, but the damage was done. Channel inventory gluts are a sticky problem, as not only does the company have to curtail shipments in the current quarter to bring down dealer stock, but it also faces a year’s worth of tough comps artificially boosted by the over-shipments. Two quarters later, Nvidia stock still trades at just over half its September 2018 highs. We expect cloud datacenter CAPEX, down a bit vs. its own difficult compares from a year ago, will roar back in 2H19, pumping new life back in to Nvidia sales growth (Exhibit 7).
While all companies must guard against poor execution, the risk is particularly acute for growth tech names, as time spent correcting management error gives rivals opportunity to gain ground in a race for critical mass, diverts management attention and can affect employee retention and hiring. While it is probably impossible for investors to avoid execution risk entirely, it is fair to say that it is more likely in fields like Hybrid/Multi-cloud platform software, where the solutions are complex and strategically important for customers who lack their own expertise, technical standards are still evolving, and multiple vendors with disparate approaches are vying for attention.
Exh 6: Summary of Execution Risks facing SSR TMT picks
Exh 7: Consensus estimates of cloud capex spending are fairly conservative
In late May, the Wall Street Journal reported that the Department of Justice was considering opening a formal probe into Alphabet over alleged anti-trust violations, and that the Federal Trade Commission was preparing for similar investigations into Amazon, Apple and Facebook. This sparked a sell-off in these stocks that bled over into the whole tech sector, until the market’s attention was drawn back into the China-US trade soap opera. While politicians on both sides of the aisle have been demanding new scrutiny of these tech mega-caps in light of accusations of undue censorship/fostering hate speech, we believe that it is very unlikely that these reported regulatory initiatives will have much bite for the companies in question. Our reasoning:
Timing – Government moves slowly. Microsoft’s November 2001 consent decree, in which it agreed to separate its browser from the Window’s operating system, was the culmination of a decade of government investigation, which began as a DoJ probe in 1992. A 1999 trial led to a guilty verdict and an order to break up the company, but the judgment was thrown out on appeal, and the court ordered to consider the case on narrower grounds and determine a much less drastic remedy. Arguably, by the time the consent decree was negotiated, its significance was moot, as the nature of competition in software had already changed to Microsoft’s detriment and the mobile era was on the horizon. The biggest downside was probably the distraction that the case had made for Microsoft management. We believe that investigations of the current generation of Mega-caps would be unlikely to proceed any more quickly than that.
Exh 8: An overview of key differences between US and EU anti-trust laws
Proving Consumer Harm – Unlike the EU, where a finding that a dominant company used its market power to squelch competitors is sufficient to yield an anti-trust verdict, US law requires that the prosecution prove damage directly to consumers or that strategic actions are taken for no other reason but to squelch competition (Exhibit 8). This is a very high bar. At first blush, digital platforms, which have largely fostered lower prices and often provide substantial services to consumers for free as a vehicle for advertising would be difficult targets. Advocates for legal action assert that under these standards, behavior that erodes consumer privacy, presumed to be a component of service quality, could be sufficient rationale to bring charges. Likewise, acquisitions that forestall competition for a dominant platform could be deemed anti-competitive. Considering this perspective, we believe the businesses most likely to receive scrutiny would by Google’s bundling its franchise applications (e.g. search, maps, YouTube, Play) with Android licenses, Apple’s rigid control of app distribution for iOS devices, and Facebook’s privacy standards and record of buying potential competitors. Still, these interpretations are somewhat afield of historical anti-trust jurisprudence, and the likelihood of a winning broad judgment against any of the four platform leaders seems small and none of them would appear to justify divestitures or other highly disruptive remedies.
Scope – The DoJ and FTC launching investigations does not mean that the probes will necessarily lead to formal charges. The agencies might find against anti-trust violations or determine that prosecution for anti-trust is unlikely to be successful in court, and thus, end the investigation without charges. It may also be that the government decides to prosecute on narrow grounds where it believes it can make a successful case, but where drastic remedies, such as forced divestitures would be off of the table. For example, rather than seeking to prove Google’s dominance in digital advertising and force separation of Search and YouTube, the DoJ could take issue with the presentation of search results and demand better page placement for non-Google site options. While nettlesome, this would hardly be a debilitating outcome for Alphabet. Similarly, breaking Facebook, Instagram and WhatsApp apart or asking that Amazon divest AWS are unlikely.
Unintended Consequences – In the unlikely chance that the government chooses to pursue a broad restructuring agenda, we expect attention to shift to the likely impact of drastic measures. Alphabet, Amazon, Apple and Facebook are rivals, and weakening one or two would create a void for the others, along with other powerful players, like Microsoft, to move in. Requiring Apple to allow alternative app stores and to support iMessaging and other proprietary apps on other platforms would be a gift to Google and Amazon. Regulating Amazon’s marketplace would raise Google’s profile in e-commerce. Splitting Alphabet into parts boosts Facebook, Microsoft, Apple and Amazon, without, necessarily doing much for the consumer while hamstringing attempts to bring more powerful user experiences to market. It’s like a dog chasing a car – what would it do if it caught it? We expect this reasoning to gain attention should the investigations gain more momentum.
Politics – Right now, with “fake news” a major controversy, politicians are talking tough about big tech companies. A progressive President with a compliant congress could change US anti-trust law to soften the requirement of consumer harm, which would raise the likelihood of broader action against the tech mega-caps. Still, we believe the current news flow overestimates the public support for anti-trust action that would threaten to reduce the utility of the major digital franchises – e.g. less selection and higher prices from Amazon separated from its market place, weaker privacy and malware from a disconnected Apple App Store, or severed connections between the various Google services. Even with a change to US anti-trust law, we do not believe that there would be a viable way to “break up” Alphabet, Amazon, Apple and/or Facebook without a cascade of unintended consequences, without significant consumer dissatisfaction, and without major political blowback. As such, we see the most likely outcomes as either minor actions to address very specific concerns with practical remedies, or simply, status quo. Ultimately, progress is its own regulator. The tech sector undergoes generational paradigm changes that tend to upend dominant leaders about every 25 years, and this history suggests that Alphabet, Amazon, Apple and Facebook will face unforeseen challengers that could prove far more consequential than will anti-trust probes (Exhibit 9).
Exh 9: SSR summary of tech regulatory risks and outcomes
We wrote about the perils of valuing tech companies last quarter (Model Portfolio Update: What is Value in Tech?) but given the volatility in the market amidst an unusual and often confusing news flow, we thought it worth reiterating. Traditional valuation metrics, such as P/E or P/S, are historically poor predictors of performance for tech stocks. In almost any other sector, stocks that are “cheap” compared to their history or to their peers can generally be counted on to “regress to the mean” and catch up to a “normal” valuation. This sort of strategy is disastrous in tech, where cheap stocks are very often cheap for a reason, while seemingly expensive stocks tend to outperform (Exhibit 10, 11).
This is compounded by the tendencies of sell-side analysts to cluster their estimates within a narrow band (why stick your neck out?) and to presume that high flying tech names will decelerate back to a “normal”
Exh 10: Historical Quarterly Surprises by Model Portfolio Constituents last 5 years
growth rate over the course of a five-year forecast. The best tech companies deliver growth trajectories well above the assumption of deceleration – i.e. Amazon and Facebook have been “expensive” throughout their tenures as public stocks, but their relentless growth have forced analysts to continually revise their forecasts upward. Against the actual numbers that have been posted subsequently, Amazon and Facebook should have been considered “cheap”.
Valuation concerns pop up all over the TMT landscape, but have been a particular worry for investors in enterprise SaaS software. Stocks like Salesforce, Workday, Service Now, ZenDesk, Alteryx, Splunk, Atlassian, and others have sustained growth trajectories against the projection of deceleration, some of them actually delivering further acceleration at the top line. Given that SaaS bundles many costs (e.g. datacenter infrastructure, networking, user support, etc.) into the subscription price, and that they compete not just against packaged applications but custom built software as well, we believe that many are underestimating the total addressable market, presuming that the end of the runway is much nearer than it actually is.
The Long-Anticipated SaaS Market Consolidation May Be Upon Us!
Last week, Alphabet announced that it would acquire privately held cloud analytics company Looker for $2.6B. On Sunday, Salesforce announced its $15.7B deal for publicly traded Tableau Software, another cloud-based business analytics tool (Exhibit 12). Given significant scale advantages for cloud software and the increasing difficulty for fast growing smaller SaaS names to hire sales talent, we see a lot of synergy in
Exh 11: Average Surprises by Constituents for Most Recent Fiscal Quarter
Exh 12: Snapshot of SaaS M&A activity in the last 3 years
these combinations and suspect that we will see a lot more M&A activity in the near future. We have written about this in the past (SaaS: Handicapping the Unicorn Races), and the potential that the deal market will heat up adds to the reasons for investors to emphasize SaaS. We currently hold five SaaS names in our 15-stock model portfolio – six if you count Microsoft – and can think of a dozen more that would fit nearly as nicely.
Replacing PVTL with NTNX
This swap of hybrid cloud platform software players reflects our strong belief that hybrid/multi-cloud will be the dominant enterprise IT architecture for the next decade or more. While the direction is clear, the market remains fragmented and inexpert enterprise IT departments are leery of locking themselves into proprietary solutions from big cloud platforms or traditional IT players. The combination of urgency and uncertainty makes stocks in the sector volatile but also suggests significant upside rewards for investors. While both Pivotal and Nutanix have made missteps, we see the sell offs as opportunities. Of the two, we see Pivotal’s problems as more serious and expect recovery to last most of 2019. In contrast, we believe Nutanix will show substantial re-acceleration in 2H19 as the transition from bundled hardware sales to the subscription model proceeds and the comps get easier.
Exh 13: NTNX growth slowed due to an expensive transition to recurring model
Exh 14: Summary of financial and valuation metrics for NTNX
Moreover, we believe Nutanix is the more likely acquisition of the two. It is the technology leader in the hybrid/multi-cloud integration software at the core of the new architecture. Unlike Pivotal, which is majority owned and controlled by Dell, Nutanix is also independent of the major would-be hybrid players – e.g. IBM/RHAT, DELL/VMW, CSCO, AMZN, GOOGL and MSFT – and would be a valuable addition to any of them in this high-stakes competition. While it is tempting to wait for the market to mature a bit, particularly given PVTL’s painful blow-up, we believe NTNX’s current share price is an unusually attractive entry point for the stock (Exhibit 13, 14).
We considered adding TWTR or PINS, as we believe both are well positioned to disproportionately benefit from the continued shift to digital advertising but believe that NTNX has more upside from its current valuation. We are biding our time for NVDA, still recovering from its channel inventory glut and from a momentary softness in datacenter capex. We expect a strong rebound before year end.
Exh 15: Summary of updates to the model portfolio going forward
Exh 16: SSR TMT 15 Stock Model Portfolio – RECONSTITUTED