Description

Summary thesis

SAP is the world’s largest ERP software vendor. It operates in a number of software verticals focused on large enterprises. The typical SAP customers is a Fortune500 company. The reason we are excited about SAP is because it is a very high quality business, with tremendous barriers to entry, high visibility into its long-term trajectory and currently misunderstood due to an imminent inflection in the business profile. Two value drivers are about to inflect:

• Top line growth is about to accelerate following years of subdued sales growth thanks to the launch of the new S4HANA ERP system

• EBIT margins are about to increase following 5 years of declining / stagnant profitability

About the business

SAP was founded in 1972 by former IBM engineer Hasso Plattner and launched its first financial accounting software in 1973 – that was the R/1 system. In 1982 it launched its second ERP system – R/2. In 1992 the non-publicly listed company launched R/3 which was truly revolutionary at the time. It allowed SAP to gain considerable market share in enterprise applications becoming the global #1, adding CRM / HCM capabilities to large and medium size companies.

In 2003 Hasso Plattner resigned from the Executive Board and a period of relatively slow and confused innovation begins. In 2004-05 a new version of R/3 is released providing a renewed upgrade cycle.

In ‘05-10 SAP became the victim of its own success: its heritage of an on-premise ERP business with very high switching costs and high profitability made the company resistant to any change. Lack of innovation allowed more agile solutions / SaaS pure-players to nibble market share from SAP in non-core applications. The core ERP suite was never at risk though. To compensate for loss in market share in other non-core applications, SAP management just hiked maintenance fee in 2008, further alienating customers. The traditional license + maintenance model drove an ever-expanding margin profile for the company:

Customer stickiness and business innovation

It’s virtually impossible to switch out of the SAP ERP system. We have spoken to numerous CTOs. Each one confirmed that changing ERP provider would entail significant costs, lengthy (multiple years) transition phases and career risks. It’s simply not worth attempting to change an ERP system. Entire factories, distribution centres, logistical operations would come to a halt. SAP used this customer stickiness to upsell adjacent products, some developed organically, others via acquisitions. SAP pushed its own CRM and HCM offerings.

However, this strategy faltered as customers just stuck do ERP (without upgrading) but moved some functions to SaaS competitors in CRM, HRM and other verticals (e.g.  Salesforce, Workday). In February 2010 Hasso Plattner hit the reset button and took control of the company replacing Leo Apotheker. The mandate for the new CEO was clear: innovate or die, SAP’s efforts in innovation are giving their fruits only now - it has been a very long process.

SAP acquired Sybase in July 2010 to accelerate the move to in-memory database. In 2011, the first HANA generation was launched (HANA = HAsso New Architecture). Between 2011 and 2014 SAP spent c. €14bn in M&A to revamp its product suite and innovation. What happened though was that the innovation cycle left 2 years of product hiatus: by 2013, all R/3 customers would have had to leave to ECC, so all forced upgrades were done. ECC has been around for already 10 years and didn’t offer much in terms of innovation:

The product cycle was such that following the introduction of ECC in 2004, license sales continued to grow until 2012-13 when license revenue growth started to stall. As evident in the following chart, license revenue growth stalled between 2013 and 2016. This led to a derating of the stock, from large premium to market to virtually no premium at all. Our bullish thesis is predicated on the reacceleration of such license revenue growth. We note that our view is non-consensual: at its capital markets day in February 2015, the company guided for long term license decline of 5% per annum. We are assuming a return to growth from 2017 onwards:

There are 2 main reasons for this bullish view: HANA and S4HANA.

Enter HANA

HANA is a database platform for next generation analytics and applications. HANA provides an in-memory platform to run applications faster and more efficiently. In layman terms: older (e.g. Oracle, IBM) database platform need a double process to perform analytical applications: all data would need to be written on a database. This database would need to be “copied” on a second database that would be used to perform the analytics. Under this old system, it is impossible to write data in memory and read data for analytics at the same time. All data used for analytics is therefore old / stale. Under HANA, which is “in-memory”, applications can be run in “real time” memory. There are significant advantages:

• Everything is in real time, allowing the user to perform analytics on the latest available data. Tasks that required hours can be now done in seconds

• Memory requirements are lower, therefore reducing Total Ownership Cost

Practical example: it would allow large corporate much better working capital management as the CFO always knows in real time its inventory / payable / receivable data. Essentially, SAP is integrating with HANA Business Intelligence with Core transactional activities.

HANA impact on SAP is massive for the following reasons:

• All future SAP applications (ERP, HCM, SCM etc.) will be developed on HANA database. Customers will no longer be able to use new applications if they stay with Oracle / IBM solutions. SAP is basically “forcing” people to drop their traditional IBM / Oracle database and move to HANA. Those HANA licenses are 100% incremental to its revenue base (just to give order of magnitude: Oracle database sales are c. $15bn p.a. We expect Hana sales to reach c. €4bn by 2020)

• HANA enables applications like S4 HANA to deliver substantially higher functionalities to customers. Speaking to the CTOs of multinationals, they described HANA and S4HANA as “Copernican Revolution”, dramatically changing the way to run a business from an IT perspective. The step change should trigger an accelerated refresh / upgrade cycle. The CTO of a large multi-billion-dollar business also told us they are dropping Oracle to move to HANA

• SAP announced that all traditional non-HANA based applications will be supported until 2025 and then customers will have to move their database to HANA or change ERP provider. Whilst 2025 seems a long time away, these projects can take 5-7 years. We therefore expect most current non-HANA customers (probably 95%+ of customer base) that are on SAP to start the move to HANA before 2020

• HANA seems to expand the TAM for SAP. Out of the c. 4,000 S4HANA signed in the last year, over 30% are new SAP customers. HANA seems to prompt non-SAP customers to migrate to SAP applications to benefit from HANA

All our channel checks suggest it’s much easier to keep SAP as ERP provider and move your database from Oracle to SAP than staying with Oracle database and change ERP system. What SAP has effectively done was to force the entire customer base to upgrade by 2025. Note that HANA was introduced in 2011 but only in 2015 the first fully HANA based was released: S4HANA. We are therefore in the very first inning of the upgrade cycle.

Enter S4HANA

S4HANA is a revolutionary ERP system based on HANA and its importance cannot be underestimated. Simplistically, this is just the latest iteration of the previous ERP application (ERP / Business Suite). However, it is much more than this:

• S4 HANA offers a real step-change from previous versions, allowing data analytics that was never possible before. Whilst Business Suite / ERP was just another updated version of R3, S4 HANA is a complete new release, comparable to what R3 was when it was released in 1992

• There has been widespread criticism on SAP in recent years over its failure to innovate. S4 HANA is the answer to this and just as R3 triggered a very strong upgrade cycle in the 90s (and to a certain extent Business Suite in the mid 00s), S4 HANA could boost license sales over the next 5-7 years

As per bottom chart below (illustrative), each major ERP platform release comes with a larger addressable market in terms of offerings and should drive an increasingly larger uptick in license sales. Considering the Street assumes negative license growth, there is certainly substantial scope for positive upside surprises.

From an empirical perspective, initial conversations with experts confirm the significant change S4HANA will bring in terms of speed, data granularity, functionality etc. It is also clear that all previous ERP / R3 customers will eventually have to move to S4HANA. S4HANA also offers a new presentation layer called Fiori – one of the big criticism of the old system was that it wasn’t user friendly / had a clunky interface. Fiori fixed this issue. Recent quarterly results also suggest S4HANA is gaining acceptance. Not only existing SAP customers are upgrading to S4HANA. There is evidence that S4HANA is attracting customers that weren’t using an SAP system before. At its recent Q3-16 SAP disclosed that 40% of S4HANA customers were net new ones.

Let’s try to quantify the size of S4 Hana license cycle: ECC6.0 (Business Suite) cycle was worth c. €36.5bn in license cycle (double than the previous cycle), it’s fair to assume that the next cycle to 2025 should be worth at least c. €50bn. Even assuming 10% cannibalisation to cloud, that’s €45bn, or €5bn per year on average. Our model assumes c. €5bn on average license sales for the next 5 years. This doesn’t take into consideration the incremental license sales from HANA database which will come at c. 15% price of the application price. In theory, it could add €7.5bn in license sales over the next 9 years or c. €700m per annum. It shouldn’t take much to see license revenues to grown from the €4.8bn achieved in 2015 and 2016.

Business networks

Business Networks represents today c. €2bn in revenues, which is almost entirely Cloud Subscription Revenues. It is made of 3 separate businesses:

• Concur – acquired in 2014 for $8bn is currently growing c. 20%+ YoY. It’s the leader in corporate expense management, travel expenses. It allowed corporates to monitor, process and authorise employee expenses. C. 75% of the business comes from non-SAP ERP customers

• Ariba – acquired in 2012 for $4bn, it’s the largest supplier network in the world. It allows corporate to centralise their supply purchasing via a platform making the whole process seamless

• Fieldglass – acquired in 2014 for $1bn t’s focused on continued labour management, i.e. managing workforce (extended HCM)

These businesses represent the majority of SAP Cloud business (c. 55%) and they are not cannibalistic to SAP core ERP business. Each of these business is expected to grow into the multi-billion-dollar opportunity. Management doesn’t give guidance on Business network. Together with other cloud business representing c. 45% of total cloud, the company believes it can grow total cloud sales from c. €3bn in 2016 to €7.5-8bn in 2020.

These 3 businesses are currently profitable but disclosure is poor. For the first time, ever, at Q1-16 SAP disclosed Business Network contribution margin (+15%).

Bear cases

There are primarily 2 bear cases on SAP: dilutive M&A and margin stagnation. We are comfortable on both.

Dilutive M&A

Over 2012-14, SAP spent almost €14bn in M&A, often overpaying for its targets (see Concur). From a net cash balance of €1.5bn in 2011, SAP went to a situation of Net Debt of €7.5bn in 2014. Many bears worry the acquisition binge will continue.

Whilst we certainly don’t like expensive M&A, management and CEO McDermott has been very clear that the current portfolio requires no large acquisitions. In 2012-14, SAP saw cloud competitors gaining ground and they did not have the technology to respond. M&A was the only solution to bridge the gap. SAP didn’t make any large acquisitions in the last 2 years and we don’t expect any in the immediate future. The business is very cash flow generative and Net Debt fell from €7.5bn in 2014 to an estimated €3.5bn in 2016.

Cannibalisation from cloud and margin dilution

The move to the cloud has clearly depressed margins. There are 3 potential effects to consider:

1.    In a Cloud world where switching costs are lower and competition with new applications is high, will SAP lose market share to the likes of Salesforce.com, Workday? Will they need to lower prices cannibalising margins?

2.    Will the shift to a SaaS model be cannibalistic for SAP, thereby reducing its license growth rate?

3.    Will margin be structurally lower going forward given the mix shift?

Regarding question 1, over 50% of SAP license sales come from applications like ERP and Supply Chain Management where the shift to cloud is minimal and where competition is weak, especially given the high switching costs. CRM is the most sensitive area to SaaS competition (see Salesforce.com) but represents only c. 13% of license sales. Regarding question 2, as mentioned above, we believe cannibalisation will be c. 10% per year, so not dramatic. More core type of applications are not well suited for hosted delivery platform and are therefore relatively immune from SaaS competitors. It’s also important to note that most Cloud revenue today is not cannibalistic. Almost 50% of cloud revenues today derives from Business Networks, which is entirely incremental.

The real big question is one about margins. Over the last 6 years, EBIT margins declined:

SAP doesn’t disclose cloud EBIT margins. It discloses cloud gross margins (targeted at 80% for Business Networks and Public Cloud, 40% for Private Cloud) and a vague contribution margin from Business Networks. The strong growth in cloud revenue is clearly depressing margins as all sales and marketing costs are recognised up front whereas revenue is rebated annually in subscription format. As in every cloud business, the higher the growth, the lower the margins. In a steady state where 80% of bookings are in mature state and 20% are new / renewal, we can expect 80% gross margin. Today, mature bookings represent only 60% of total revenue. It’s clear that in the initial years, cloud revenue is significantly margin dilutive. However, we think we have reached an inflection point and cloud EBIT has just turned positive in 2016 and we expected to reach 20% margin by 2020.

The bear case on SAP margins seems to stem from the lack of understanding of cloud margins. In our view, if we exclude cloud margin, EBIT margins for the group would have continued to climb since they peaked in 2010. It’s only the growth in cloud that clouded the picture (excuse the pun). Consensus expects margin to continue to be flat going forward while we expect strong growth:

Long term targets

At its investor day in February 2015, the company disclosed the following 2020 targets:

We believe there is a good change the company will upgrade these at its forthcoming capital market day in February 2017. We believe 2017 EBIT would end up at the higher end of the guided €6.3-7bn and 2020 EBIT could exceed €10bn against a €8-9bn target. Most the upside will come, in our view, from an upgrade to license guidance. Back then the company guided for 5% annual decline:

We believe SAP could sustain mid-single digit growth for years to come. Our approach to modelling is the following:

We take a top-down view of how much should revenue grow to 2020 based on Gartner forecasts per application. We adjust Gartner’s market outlook to allow for SAP specific factors:

• Business Network business growing 20%+ which is not cannibalistic (+)

• HANA license opportunity (+)

• S4HANA new upgrade cycle (+)

• Market share losses in CRM (-) and to a certain extent in HCM (=/-)

• Cannibalisation due to move to SaaS, especially in HCM (-)

We take SAP’s long term (2020) guidance for Cloud revenue as a given (€8bn in cloud subs). We back solve the implied non-Cloud license revenue as the difference between our estimated top-down derived billings and SAP guided Cloud revenue. We then derive support revenues as a function of license growth and churn assumptions.

Note that SAP does not give billing guidance but gives Total revenue guidance as well as Cloud Revenue. This allows us to compare the company implied billing assumptions and license growth to ours:

• SAP implied guidance of -1% to -5% license growth to 2020 is in our opinion way too conservative

• It’d effectively imply billing growth of c. 2-3% to 2020 against our estimated 8% growth

Triangulating Gartner assumptions and tweaking this data for SAP specifics, we get to total billing growth of c. 8-9%, whilst SAP guidance seems to imply just 2-3% growth

Very simply speaking, from a top-down perspective we get there with the following assumptions for billings:

• Overall market growing c. 6-7% as per Gartner. SAP should grow faster than the market given the new S4HANA product cycle which has just started

• SAP to benefit from new HANA license growth which is entirely incremental. This should add c. 1-2% to revenue

• SAP will benefit from the Business network growing in the high teens, purely incremental

• Some negative growth in billings due to loss in market share in CRM and older on premise applications

Support sales are a crucial part of the equation. We are assuming c. 95% renewal rate and a blended 20% support fee. This leads to long term support sales of c. 5% growth.

As per below, we get to sales close to €30bn in 2020 against company target of €27bn (mid-range) and EBIT above €10bn against company target of €8.5bn (mid-range):

Valuation considerations

Between 2008 and 2013, when the business was growing strongly and margins were expanding, SAP used to trade at c. 40% premium to the market. Today, it trades at little over 18x 2017 earnings, a very small premium to the market. The relative derating occurred due to margin decline explained as flattish license sales. With a reacceleration of the business and with margins about to inflect, we believe SAP will rerate close to where it used to trade. We believe SAP should trade at 20x P/E, which is line with comps:

Applying a 20x multiple to our 2019 EPS estimate of €5.6, we get a total IRR over 2 years of approximate 20% including dividends, which we view as attractive given the high quality of the business and the extreme long term visibility offered by the business. Also, we note that our 2019-20 EPS figures are 15-20% above consensus respectively.

Catalyst

24 Jan 2017: Annual 2016 results and 2017 guidance

9 Feb 2017: Capital Market Day, 2020 targets possible upgraded