Thursday, October 12, 2017
Speaker: Rodney Nelson, Senior Equity Analyst, Morningstar Research Services LLC
1. Moats in software.
• Identifying moats in software can be tricky. ROIC vs WACC: developing software is not capital intensive business → inflated levels of ROIC → normalize capital expenditures (capitalize R&D, etc).
• The primary moat sources are:
o Switching costs – cost of implementing, using the software and its end-user training; for application vendors derived from data migration, deep integration, customer loyalty. Ex. Salesforce.com: 75% of top customers use 4+ clouds in 2017 vs 13% in 2013.
o Intangible assets – inherit knowledge in a very specific industry, niche area of expertise, deep customer relations.
o Network Effects – data-based network effect (e.g. Salesforce, Ebay, Amazon, Facebook), ubiquity in user base (e.g. Microsoft Office).
o Cost advantage.
o Efficient scale.
2. Breaking down the Cloud Value Chain.
• The Cloud Value Chain
o On premises - the customer is responsible for literally everything from hardware to code and application; need enough skilled labor, capital etc. to build and maintain the system.
o Infrastructure-as-a-Service (IaaS) - provides networking storage and virtualized compute power; vendor manages and refreshes hardware and infrastructure software.
o Platform-as-a-Service (PaaS) - vendor provides environment for building/testing/deploying applications; customers worry about writing code and managing the application.
o Software-as-a-Service (SaaS) - vendor manages the entire stack; customer worry about actual use of the application by end users.
• Why migrate?
o Expensive to maintain the infrastructure when running an on-premises software – input costs, constant maintenance payment, security, updating hardware and software, etc.
o ~30% savings between On-Premises and SaaS costs.
o Inefficiency is the biggest cloud development driver – the average hyper cloud vendor spends about $1bn/year on system updating/maintaining/security, creating a much more efficient environment including security against a would-be attackers.
o On premises data-center capacity utilization rates are on average 10%-20%, while in a hyper scale public cloud environment a company consumes and pays only for storage that it actually uses/needs.
• Key trends in SaaS
o Still in early stages of the migration to SaaS – looking from the enterprise perspective penetration rate is only 25%.
o Customer relationship management CRM (e.g. Salesforce) and enterprise resource planning ERP (e.g. Workday) are the product suits seeing the fastest migration.
o Profitability: SaaS provider recognizes all the expenses connected with booking a contract right away, but recognizes revenue from the customer gradually along the service term (even though collects cash), meaning the renewals of contracts provides for higher profitability, free cash margin could be considered a leading indicator for future profitability.
o Business model transitions (from on-premises software to subscription payment) can work, but results are mixed. Adobe is a “gold standard”.
o Pure-Play SaaS firms boast strong and improving competitive positions (Salesforce, Workday, ServiceNow).
• Public Cloud (IaaS and PaaS)
o Public cloud moats are built on cost advantages and intangible assets.
o Public cloud is the most capital intensive market - in order to serve the needs of large multinational customers large vendors need to have global scale multinational set of datacenters meeting reliability, redundancy, scalability requirements, regulatory concerns etc. Amazon & Microsoft are the largest public cloud vendors.
o Vendors attract huge swaths of customers, allowing them to capitalize on scale efficiencies on top of a relatively fixed set of input costs.
o Intangible assets - inherit knowledge of an enterprise software and development of unique premium services.
o Capital intensity and intangible assets will limit competition globally. Several vendors have given up already, e.g. HP, VMWare, Cisco.
o Price war market, but scale and premium services are what really drive margins.
o Public cloud market represents a massive opportunity – est. $200+bn market by 2021.
o We see four major vendors that may consume that opportunity: Amazon and Microsoft (Azure) – lion share, and also Alphabet’s GCP and AliCloud.
3. Market valuation and investable opportunities.
• The tech sector has heated up in 2017.
• Cash flow is the most important indicator in software valuations/multiples.
• Best ideas: Microsoft and Salesforce.com
1. What is the difference between Platform as a Service and Infrastructure as a Service businesses?
• Platform provides the defined set of tools built for a specific type of development, for example, Twilio made a very specific set of tools to a customer to build one type of application (communication based).
• Infrastructure provides more flexibility and freedom - a customer has freedom to deploy any software assets or choose from a set of tools.
2. What are your short ideas?
• Tableau Software is incredibly overvalued (trades at 7.0x P/Sales), and it is now in a compromised market position offering relatively unique but replicable product, very similar to a product offered by Microsoft and included free in 365 Office. Being cash reach it can be a good takeover target but not at current multiples. Similar company Click was privatized last year at 3.8x P/Sales.
• We are also conservative on Oracle due to its lack of investments in a public cloud business.
3. What industries are currently most utilizing public space services and what industries are growing the most?
• It is becoming more ubiquitous among industries to use cloud services.
• Government, financial services and technologies are the ones that lead. They are also growing the fastest as they are
consuming the most products.
• In terms of penetration, more of the upside is on the front of consumer and industrials. (Ex. Salesforce has recently acquired Demandware to extend from purely CRM into the retail space with e-commerce technology).
4. What is your opinion of Amazon?
• I follow the AWS component of Amazon business. Looking at the model, AWS is a unique cash generating profitable asset that allows Amazon to continue operating its e-commerce business.
• In terms of valuation, we think Amazon is marginally undervalued compared to its current price. AWS accounts for about 2/3 of Amazon value and is also the most underappreciated component of the valuation.
5. The wireless communications companies has been talking about rolling out 5G and need to reduce the latency; they talk about the wireless communication becoming a new platform, and they really have their eyes set on the autonomous cars. In light of that, who of the public cloud providers are best positioned to profit from the autonomous cars, who is likely to dominate the software that runs the autonomous cars, and are we going to see a massive disruption in general automobile industry function?
• The big challenge for 5G technologically, especially in a view of powering autonomous vehicles, is the need for small cell size everywhere to provide the level of connectivity to send data back and forth between the car and the datacenter.
• From the cloud perspective, the companies that invested the most in the development of a cloud could potentially provide that services just because they have the largest installed capacities, e.g. Amazon and Microsoft (in this part of the world).
• In terms of auto industry, we do not see a major auto manufacturer that has not been investing into development of an autonomous vehicle.
• Our view is that rapid development of autonomous vehicles is still not in the nearest future largely due to underdevelopment of wireless infrastructure, especially in the rural areas, and the enormous associated CAPEX to develop it (which we think would largely be deployed by wireless communications companies/tower owners rather than cloud vendors).
6. Do you anticipate cloud vendors market to display monopolistic or oligarch market features and any regulations following?
• Now it is hard to make a delineation between a free market and something that should be treated as utility. But cloud market has clearly been evolving as an oligopolistic market.
• It could be tough to regulate.
7. In your forecast you have Google cloud service being less successful than Microsoft and Amazon. Given that Google has no capital constraint why do you think so?
• Probably Google’s internal strategy – they only want to develop “next wave” technologies, don’t want to be backward looking. But within companies there are a lot of backward looking technologies that are still relevant.
• Google has been very aggressive on cloud development. They have recently hired Diane Green, a co-founder of VMware, a company that invented a technology which made cloud possible in terms of server virtualization. So the progress has been made on that front, but Google is still way behind the progress that has been made by its main rivals, Amazon and Microsoft.
• Given that they are not capital constraint Google definitely has resources to build a global scale cloud, and has plans to be represented in some 45 regions.
8. How cloud development has impacted hardware producers?
• Hardware producers are facing a very long run of downward pricing pressure.
• Net new deployments of servers in on premises datacenters within the US has been plummeting meaning that hardware producers’ swath of potential consumers is shrinking dramatically.
• We have been seen public comments from hardware producers like HP where they seriously reconsider whether they should be a server business.