Did SAP Overpay For Concur?

Since SAP announced to acquire Concur and eventually closed the acquisition for $8.3B many people have reached out to me asking whether SAP overpaid for Concur. I avoid writing about SAP on this blog even though I work for SAP because this is my personal blog. In this case, I decided to write this post because this is the largest enterprise SaaS acquisition ever and this question unpacks the entire business model of SaaS enterprise software companies.

If you’re looking for a simple “yes” or “no” to this question you should stop reading this post now. If not, read on.

People reaching out to me asking whether SAP overpaid for Concur in itself is a misleading question because different people tend to compare Concur with different companies and have a specific point of view on whether the 20% premium that SAP paid to acquire Concur is justified or not.

Just to illustrate financial diversity amongst SaaS companies, here are some numbers:

This is based on a combination of actual and projected numbers and I have further rounded them off. The objective is not to compare the numbers with precision but to highlight the financial diversity of these companies based on their performance and perceived potential.

Market cap is what the market thinks the company is worth. The market doesn’t necessarily have access to a ton of private information that the potential acquirer would have access to when they decide what premium to pay. While the market cap does reflect the growth potential it is reflected in a standalone pre-acquisition situation and not post-acquisition.

The purchase price, including the premium, is a function of three things: revenue, margins, and growth (current, planned, and potential). However, not all three things carry the same weight.

Revenue

For SaaS companies, annual recurring revenue (ARR) is perhaps the most important metric. It is not necessarily same as recognized revenue what you see on a P&L statement and ARR alone doesn’t tell you the whole story either. You need to dig deeper into deferred revenue (on the balance sheet and not on P&L), customer acquisition cost (CAC), churn, and lifetime value of a customer (LTV) that companies are not obligated to publicly report but there are workarounds to estimate these numbers based on other numbers.

Margin

If you’re a fast growing SaaS company the street will tolerate negative margins since you’re aggressively investing in for more future growth. Margin is less interesting to evaluate a fast growing SaaS company, for acquisition purposes or otherwise, because almost all the revenue is typically invested into future growth and for such SaaS companies the market rewards revenue and growth more than the margins.

Margin by itself may not be an important number, but the cost of sales certainly is an important metric to ensure there is no overall margin dilution post acquisition. Mix of margins could be a concern if you are mixing product lines that have different margins e.g. value versus volume.

Growth

Current and planned growth: This is what the stock market has already rewarded pre-acquisition and the acquirer assumes responsibility to meet and exceed the planned or projected growth numbers. In some cases there is a risk of planned growth being negatively impacted due to talent leaving the company, product cannibalization, customers moving to competitors (churn) etc.

Growth potential: This is where it gets most interesting. How much a company could grow post-acquisition is a much more difficult and speculative question as opposed to how much it is currently growing and planned to grow pre-acquisition (about 29% in case of Concur) as this number completely changes when the company gets acquired and assumes different sales force, customer base, and geographic markets. This is by far the biggest subjective and speculative number that an acquirer puts in to evaluate a company. 

 

To unpack the “speculation” this is what would/should happen:

LTV 

This number should go up since there are opportunities to cross-sell into the overall joint customer base. LTV does reduce if customers churn, but typically preventing churn is the first priority of an acquiring company and having broader portfolio helps strengthen existing customer relationship. Also, churn is based on the core function that the software serves and also on the stickiness of the software. The most likely scenario for such acquisitions is a negative churn when you count up-selling and expansion revenue (not necessarily all ARR).

CAC

This should ideally go down as larger salesforce gets access to existing customer base to sell more products and solutions into. The marketing expenses are also shared across the joint portfolio driving CAC down. This is one of the biggest advantages of a mature company acquiring a fast growing company with a great product-market fit. 

Revenue growth

As LTV goes up and churn goes down overall ARR should significantly increase. Additional revenue generated in the short term through accelerated growth (more than the planned growth of the company pre-acquisition) typically breaks even in a few quarters justifying the premium. This is an investment that an acquiring company makes and is funded by debt. Financing an acquisition is a whole different topic and perhaps a blog post on that some other day.

Margin improvement

This is a key metric that many people overlook. Concur has -5.3% operating margin and SAP has promised 35% margin (on-prem + cloud) to the street by 2017. To achieve this number, the overall margins have to improve and an acquiring company will typically look at reducing the cost of sales by leveraging the broader salesforce and customer base.

This is a pure financial view. Of course there are strategic reasons to buy a company at premium such as to get an entry into a specific market segment, keep competitors out, and get access to talent pool, technology, and ecosystem.

Based on this, I’ll let you decide whether SAP overpaid for Concur or not.

Disclaimer: I work for SAP, but I was neither involved in any pre-acquisition activities of Concur nor have access to any insider Concur financial data and growth plans. In fact, I don’t even know anyone at Concur. This post is solely based on conventional wisdom and publicly available information that I have referenced it here. This post is essentially about “did x overpay for y?,” but adding SAP and Concur context makes it easy to understand the dynamics of SaaS enterprise software. 

Photo courtesy: Iman Mosaad

Salesforce.com's $212 Million Acquisition of Heorku - A Sparkling Gem In Radiant Future Of Cloud And PaaS

I met James Lindenbaum, a founder of Heroku, in early 2009, at the Under The Radar conference in Mountain View. We had a long conversation on cloud as a great platform for Ruby, why Ruby on Rails is a better framework than PHP, and viability of PaaS as a business model. He also explained to me why he chose to work on Heroku at Y Combinator. I was sold on their future, on that day, and kept in touch with them since then. The last week, Salesforce.com acquired Heroku for $212 million. That's one successful exit, which is good news in many different dimensions.

PaaS is a viable business model

PaaS is not easy. It takes time, laser sharp focus, and hard work to build something that the developers would use and pay for. A few companies have tried and many have failed. But, it is refreshing to see the platform and the ecosystem that Heroku has built since its inception. Heroku did not raise a lot of money, kept the cost low, and attracted customers early on. I was told (by Byron, I think) that an average cost for Heroku to run a free Ruby app for a month was $1. They considered it as marketing cost to get new customers and convert the free customers to paying ones, as they outgrew their needs. I cannot overpraise this brilliant execution model. I hope to see more and more entrepreneurs being inspired from - simplicity, elegance, and execution of Heroku's model - to help the developers deploy, run, and scale their applications on the cloud. In the last few years, we have seen a great deal of innovation in dynamic programming languages, access algorithms, and NoSQL persistence stores. They all require a PaaS that the developers can rely on - without worrying about the underlying nuts and bolts - and focus on what they are good at - building great applications. If anyone had the slightest doubt on viability of PaaS as a business model, this acquisition is a proof point that PaaS is indeed the future. Heroku is just the beginning and I am hoping for more and more horizontal as well as vertical PaaS that the entrepreneurs will aspire to build.

Superangels and incubators do work

There have been many debates on viability of the investing approach of the superangels and the incubators, where people are questioning, whether the approach of thin slicing the investment, by investing into tens and hundreds of companies, would yield similar returns, as compared to return on traditional venture capital investment. I also blogged about the imminent change in the VC climate, and decided to watch their returns. The numbers are in with Heroku. It's a first proof point that a superangel or an incubator approach, structurally, does not limit the return on the investment. I believe in investors investing in right people solving the right problems. If you ever meet James and hear him passionately talk about Ruby, the Heroku platform, and the developer community, you will quickly find out why they were successful. Hats off to YC on finding this "jewel". No such thing as too little investment, or too many companies.

Ruby goes enterprise

I know many large ISVs that have been experimenting with Ruby for a while, but typically these efforts are confined to a few small projects. It's good to see that Ruby, now, has a shot of getting much broader adoption. This would mean more developers learning Ruby, cranking out great enterprise gems, embracing Git, and hopefully open source some of their work. I have had many religious discussions, with a few cloud thought leaders and bloggers in the past few months, regarding the boundaries of PaaS. The boundaries have always been blurry - somewhere between SaaS and IaaS - but, I don't care. My heart is at delivering the applications off the cloud that scales, delivers compelling experiences, and leverages economies of scale and network effects. To me, PaaS is means to an end and not the end. I am hoping that an acquisition of a PaaS vendor by a successful SaaS vendor will make Ruby more attractive to enterprise ISVs and non-Ruby developers.

I have no specific insights into what Salesforce.com will do with Heroku, but I hope, they make a good home for Heroku, where they flourish and continue to do great work on Ruby and PaaS. This is what a cloud and Ruby enthusiast would wish for.

Software Is The New Hardware

Today Intel announced that it is buying McAfee for $7.7 billion. This acquisition made people scratch their heads. Why McAfee?

The obvious arguments are that Intel has hit the growth wall and organic growth is not good enough to satisfy the shareholders. But this argument quickly falls apart from margin perspective. Why dilute their current nice gross margin even if McAfee has steady revenue stream? [Read my update at the end of the post]

I believe there are two reasons. The first is that the companies need a balanced product and revenue mix regardless of different margins. Oracle bought Sun and HP bought EDS. Big companies do this all the time. The second, not so obvious, reason is a recognition that software is new hardware. The processors are processors – they are a commodity any which way you look at them. It is not news to anyone that the computing has become commodity which is the basis of utility style cloud computing. Software, embedded or otherwise, has significant potential to sell value-added computing. The security solutions could fit in nicely on a piece of chip. When you drive a few miles from Intel’s headquarters to meet folks at nVidia you will be amazed to see what kind of value a software tool kit can derive from the processors.

I don’t know how Intel will execute the merger considering the fact that this is their largest acquisition ever. But, I am even more convinced that software is the new hardware. Cloud computing, data center automation, virtualization, network security, and a range of other technologies can leverage software in a chip that is optimized for a set of specialized tasks. Time to move from commodity to specialized computing until specialized computing becomes commodity. Interesting times!

Update: Romit sent me a message commenting that how McAfee will dilute Intel’s margin since McAfee’s gross margin is more than Intel. I should clarify. The assumption on the street is that the cost of capital for this purchase is about 4% and Intel expects 8% return on the investment even after paying 60% premium for the purchase. The tricky part is that how long Intel can maintain the close to 75% software margin of a software company operating inside a hardware company. When I say diluting the margin I mean diluting the overall combined margin post-purchase. The analysts are skeptical about the success of the merger and so am I. Intel has no track record of integrating large software companies such as McAfee especially after paying significantly higher than average premium. Hypothetically if Intel were to buy a company with more synergies that can leverage existing channels and can fit into their culture they could have increased the gross margin and hence the return to their shareholders.

Cisco and Juniper eyeing the long tail of consumers for their second act

Very few companies have excelled in business beyond 25 years only with their first act, a product or a business model. Some companies recognize this early on and some don't. The networking giants Cisco and Juniper seem to get this and are looking for their second act. You don't wake up one day and drastically change your business model. It's a conscious decision based on long term strategy with very focused short term execution that is required to get to the second act.

Cisco started their "human network" efforts by acquiring Linksys and Susan Bostrom completely rebranded Cisco a couple of years back. Consumerization of the brand was a big leap from an enterprise-centric organization to get closer to non-enterprise consumers. Few days back Cisco announced the Q4 results and John Chambers emphasized that Cisco would invest into adjacencies.

"..and we will use this time as an opportunity to expand our share of customer spend and to aggressively move into market adjacencies."

On the other side of the networking world Juniper recently hired Kevin Johnson as their CEO who was the president of platform and services division at Microsoft. Competing with Cisco has been challenging and Juniper did have their own share of issues in the past but let's not forget this company started during the dot com era, had a spectacular performance, survived the burst, and kept growing. But now is probably the right time to look for the second act.

For Cisco, what could the second act be? Other than the obvious long tail of consumer-centric human network strategy I see a couple of possibilities:

1) Data Center Virtualization:

The virtualization is a fast-growing market segment that has not yet saturated. The real boundaries of data center virtualization are blurry since it is a conglomeration of server, network, and storage virtualization. Customers don't necessarily differentiate between managing servers versus backing up data across data centers.

This is an adjacency that Cisco can tap into with its current investments into data center virtualization switches such as Nexus 7000, strong ecosystem, and great service organization (The service revenue is 20% of the product revenue). In fact this was speculated when Cisco announced this switch.

This could indeed strain its relationship with vendors such as IBM and make it precarious who OEMs Cisco's switches in their data centers. Companies with large ecosystem would inevitably introduce "co-optition" when they decide to sell into the adjacencies that are currently served by their partners. They will have to learn walking on a thin rope.

Virtualization with scale can lead to rich business scenarios. Imagine a network virtualization switch that is not only capable of connecting data centers at high speed for real-time mirroring and backups but can also tap into the cloud for better network analysis. The routing protocols and network topology analysis require massive parallel processing that can be delivered from the cloud. This could lead to improvisation of many network and real-time voice and data management scenarios that otherwise wouldn't have been possible. Cisco's partnership with a cloud vendor could lead to some interesting offerings - think of it as network virtualization on steroids.

2) Network SaaS:

Network Managed Services has always been an interesting business with a variety of players such as IBM, Nortel, Lucent etc. This could be one of the adjacencies that Cisco might pursue and make it a true SaaS and not just a managed service. I won't be surprised if Cisco acquires a couple of key SaaS players in near future.

On-demand and SaaS have traditionally been considered a software and utility play. The networking companies already support the data centers that provision SaaS services but they could go well beyond that to provide Networking SaaS that provisions, monitors, and maintains the networks as true SaaS offering and not just as a managed service. This could include everything from network management, security, and related services. Traditionally SIs and partners have played this role but networking companies could see this as an adjacency and jump into it since it is a natural extension from hardware to data center to managed services to a SaaS delivery. Instead of selling to a service provider who sells services to customers an effective SaaS can turn the model upside down by partnering with service providers instead of selling to them and sell to an ever growing long tail of consumers.

Technology Acquisition List

The PartnerUp has compiled a long list of technology and web acquisitions in 2007. It is not a complete list and does omit many acquisitions but it is an interesting list that shows a clear trend that the smaller acquisitions have been more frequent. I welcome this change since it allows the acquiring companies to try out the innovation at the prototype level, explore synergies, and add value to it, or toss it out. They can refine and iterate on alternatives without worrying upfront about return on the investment. I see a lot of value in big companies using their cash to try out the outside-in-perspectives and not just rely on their internal innovation engine early on.

This is certainly better than buying the "startups" at the later stage and pay through the noses due to the ridiculous artificial valuation that these "startups" get and after the acquisition not being able to figure out what to do. Skype's acquisition by eBay is an example of one of these scenarios where eBay still cannot find the synergies and monetize the acquisition.