It is difficult to discuss enterprise software without at least mentioning Microsoft, and there is no better time than now: last week the company (briefly) became the third U.S. company, after Apple and Amazon, to achieve a market capitalization of over $1 trillion, and is currently the most valuable publicly-listed company in the world.

It is a remarkable turnaround, and, thanks to the fact that I started Stratechery just months before the exit of former CEO Steve Ballmer, it is one that I have been able to document stage-by-stage. The critical breakthrough was three-fold, and, as is so often the case, the three breakthroughs were really about the same existential question — whither Windows:

• First, Microsoft’s leadership accepted that its nature was that of a horizontal company, not a highly differentiated vertical one built around Windows.
• Second, Microsoft embraced a world where Windows was one client amongst many, and targeted its services to iPhone, Android, Linux, and Mac.
• Third, and most importantly, Satya Nadella brilliantly navigated The End of Windows internally, freeing Microsoft employees to build products that customers actually wanted, not that Microsoft needed.

The most important factor that made all of this possible, though, is that for all of the disruption that the enterprise market has faced thanks to the rise of Software-as-a-Service (Saas), Microsoft was remarkably well-placed to take advantage of this new paradigm, if only they could get out of their own way.

At least in part.

The SaaS Business Model

There are three parts of any new paradigm in technology: doing current use cases better, coming up with a new business model, and creating entirely new use cases. Microsoft, to Ballmer’s credit, was actually very early to the new business model aspect of SaaS.

Previously, enterprise software was sold on a license basis: companies bought software on a per-seat basis (or per-server or per-core basis in the case of back-end software), and when new versions of the software came out, they would potentially update — or not. Or not wasn’t great for anyone: companies would be running on out-of-date software, and vendors would not make new revenue.

What Microsoft figured out is that it made far more sense for both Microsoft and their customers to pay on a subscription basis: companies would pay a set price on a monthly or annual basis, and receive access to the latest-and-greatest software. This wasn’t a complete panacea — updating software was still a significant undertaking — but at least the incentive to avoid upgrades was removed.

There were also subtle advantages from a balance sheet perspective: now companies were paying for software in a rough approximation to their usage over time — an operational expense — as opposed to a fixed-cost basis. This improved their return-on-invested-capital (ROIC) measurements, if nothing else. And, for Microsoft, revenue became much more predictable.

SaaS and Current Use Cases

A more profound implication of SaaS, though — and to be clear, I mean software accessed over the Internet, not datacenter software paid for on a subscription basis — is how it makes current use cases more efficient for existing enterprise customers on one hand, and accessible for completely new customers on the other.

Start with enterprise customers: the reality for many industries is that their needs are variable. Sometimes they need more seats for a particular piece of software, and sometimes they need less; this is particularly pronounced in the case of Infrastructure-as-a-Service (IaaS), where computing needs may change seasonally. The brilliance of paying on a subscription basis is that a company can buy exactly what it needs, when it needs it, and no more. Microsoft, again in credit to Ballmer, was moving this way with Office 365: seats could be provisioned on a monthly basis, with no major upfront expenditure required.

That lack of upfront expenditure, though, also expanded the market: buying an
Exchange seat on Office 365 means hiring Microsoft to run your email server, something that previously needed to be done internally. Now all kinds of small-and-medium sized companies could use enterprise level software without needing their own IT departments.

Microsoft’s SaaS Challenge

This was also a challenge for Microsoft, to be sure: “hiring” SaaS providers meant it was easier to find providers that actually cared about modern use cases, particularly mobile. I wrote about this in 2015 in Redmond and Reality, in the context of cloud storage:

Once you remove the burden of support and maintenance — that’s handled by the service provider — it suddenly doesn’t necessarily make sense to buy from only one vendor simply because they are integrated. There is more freedom to evaluate a particular product on different characteristics, like, say, how easy it is to use, or how well it supports mobile. And it’s here that Microsoft products, particularly the hated SharePoint, were found to be lacking.

This is where Nadella made the biggest difference. It was notable, from a symbolic perspective, if nothing else, that his first public event was unveiling Office for iPad:

This is the power CEOs have. They cannot do all the work, and they cannot impact industry trends beyond their control. But they can choose whether or not to accept reality, and in so doing, impact the worldview of all those they lead. This is why it matters that the first public event Satya Nadella appeared at was Office for iPad. This is why it matters that Microsoft released it even though the Windows Touch version wasn’t finished. This is why it matters that Microsoft gave up the pretense of Windows Phone license payments that were already effectively zero and simply made it free.

This is the only possible route for a SaaS provider: the entire point is to host all of the infrastructure in one place, which means the greatest possible gains come from increasing the addressable market, which further means serving all devices, not simply the ones owned by one’s company. That, though, is the burden of incumbency: what is obviously right from the outside is often counter to what is obviously right when it comes to company cash flows and especially company culture.

Zoom and Being Better

Redmond and Reality was about file-sharing software, but the broader idea — that SaaS changes the plane of competition from ease-of-integration to ease-of-use — is perhaps best exemplified by the rise of Zoom. It turns out that video-conferencing software is an exceptionally difficult technical problem, and Zoom has done a better job than anyone in solving those technical challenges. It is simply better than the alternatives.

Even so, if said video-conferencing software had to be delivered via an on-premises software installation, it is doubtful that Zoom would be as successful as it has been: just as important is that signing up for Zoom requires nothing more than an email address; a paid plan takes only a credit card. This reduction in friction means that quality matters more than it ever did previously, which is why Zoom is such a success.

The challenge for incumbents, including Microsoft and also other competitors like Citrix, Cisco, etc., is that years of building their business on leveraging their existing relationships with enterprises left them vulnerable to a company like Zoom singularly focused on delivering a superior product, at least once a SaaS architecture made distribution so much easier. Make no mistake, enterprise software still requires a sales force, but it is far easier to start with customers that have already discovered and tried the product on their own than it is to sell something without any sort of pre-existing relationship.

Slack and New Use Cases

There remains, though, one final implication of a new paradigm, and this one is the most profound: completely new use cases. This was something Slack sought to highlight in their S-1, which was made public last week.

First, the company argued that Slack transforms internal communications:

The most helpful explanation of Slack is often that it replaces the use of email inside the organization. Like email (or the Internet or electricity), Slack has very general and broad applicability. It is not aimed at any one specific purpose, but nearly anything that people do together at work.

Unlike email, however, most of this activity happens in team-based channels, rather than in individual inboxes. Channels offer a persistent record of the conversations, data, documents, and application workflows relevant to a project or a topic. Membership of a channel can change over time as people join or leave a project or organization, and users benefit from the accumulated historical information in a way an employee never could when starting with an empty email inbox. Depending on the size of the organization, this might provide tens, hundreds or even thousands of times more access to information than is available to individuals working in environments where email is the primary means of communication.

Secondly, Slack argues that it changes what it means to integrate software:

Also unlike email, Slack was designed from the ground up to integrate with external software systems. Slack provides an easy way for users to share and aggregate information from other software, take action on notifications, and advance workflows in a multitude of third-party applications, over 1,500 of which are listed in the Slack App Directory. Further, Slack’s platform capabilities extend beyond integrations with third-party applications and allow for easy integrations with an organization’s internally-developed software. During the three months ended January 31, 2019, our more than 10 million daily active users included more than 500,000 registered developers. Developers have collectively created more than 450,000 third-party applications or custom integrations that were used in a typical week during the three months ended January 31, 2019. Additionally, we are currently developing low-code solutions to create integrations and workflows entirely in Slack, suitable for all users and based on a simple, non-technical user interface.

There is a two-part challenge when it comes to introducing a completely new way to work: first, you have to convince companies that the new way to work is better, and second, you have to actually help them implement it. It is here that the Internet’s impact on enterprise software is the most profound:

• First, the Internet is inherently viral, thanks to the fact that information can be transmitted with zero marginal cost. In the case of Slack, telling others about its benefits required little more than a post on social media, and over time, an invitation to a Slack team.
• Second, and related to the prior point, it is actually cost-effective for Slack to provide a free product: there is no need for a customer installation, simply a few entries in a database.
• Third, implementation is a matter of paying — and that’s it. There are no qualms about using scarce IT resources, simply a question about costs, and this decision is usually based on an originally-free implementation.

This gets at why I believe Slack is the poster child for the impact of the Internet on the enterprise software market: Zoom is in some respects a more impressive business, but its use-case was a pre-existing one. Slack, on the other hand, introduced an entirely new way to work, and based on its S-1, did so in a way that will produce a very profitable company over time (Slack is losing money, but at a far lower rate than it is growing revenue; this is a company that has leverage on its costs and will be very profitable in the future).

What Microsoft is Missing

Make no mistake: the Microsoft optimism that is driving a (near) trillion dollar valuation is justified. Azure is the biggest reason, of course, but Office 365 benefits from all of the dynamics I described above: as I noted last week, its market is increasing both in terms of current customers, new users at companies it already serves, and upselling all of those users to new functionality.

At the same time, the reason to use Microsoft is very much grounded in the past: Office documents are familiar, and Exchange remains the standard for enterprise email. The advantage of going with Microsoft is that everything works mostly as it has previously. That, though, raises an existential question that Nadella’s Microsoft has yet to answer: why would a new company, without any attachment to Microsoft-based workflows, choose Office 365?

Note that this is a separate question as to whether Teams, Office 365’s answers to Slack, is viable: distribution still matters in enterprise software, and Teams has valuable strengths that derive from its integration with Microsoft’s other products.

At the same time, even the bullish case for Teams is that it captures a segment of Microsoft’s existing userbase:

This is precisely what you would expect from a product leveraging an existing use case and an existing customer relationship. Contrast this first to Zoom, which addresses an existing use case with the need to acquire new customer relationships: Zoom had a challenge building their initial customer base, but from that base they have growth opportunities both in terms of new use cases and also deepening their engagement with their customers.

Slack’s opportunity is even more striking: by virtue of starting with both new customers and a new use case, the opportunity to absorb both existing use cases (always easier than creating new ones) and also deepening utility with existing customers is significant. That is how you get IPO graphs that look like this:

Slack is not only growing users, it is also growing its monetization of those users over time, and it is fair to expect both to continue. This is exactly what Microsoft is lacking: at best the company is transitioning existing Microsoft users to a SaaS model, and keeping them away from companies like Zoom or Slack. That, though, is not a recipe for growth in the very long run.

The Enterprise Growth Framework

You can chart these three products on those two vectors — the pre-existence of a customer relationship, and the pre-existence of a customer use case:

This is where Nadella’s Microsoft has fallen short. The company has done well to leverage its pre-existing strengths into more valuable relationships with its existing customers and a viable option for new ones, and, as I noted above, has indeed moved into new use cases; Teams clearly goes in the lower-right part of the above graph:

The problem is that to the extent Teams is successful it is because it is exploiting Microsoft’s existing customer base, not necessarily winning customers who would have never considered Microsoft in the first place. There is not nearly enough industry-leading technology (as is the case with Zoom) or innovation in new use cases (as was the case with Slack) to engender confidence that the company can grow beyond its existing customer relationships in the very long run. This is why companies like Zoom and especially Slack are so valuable: they create new customers who are primed for growth; Microsoft, meanwhile, is mostly keeping its existing customers in-house.

This, then, is Nadella’s new challenge: the company could have acquired Slack early in Nadella’s tenure, and considered Zoom, but waited too long on both. Microsoft has figured out how to leverage its existing userbase: how to increase it remains an open question.