The cloud content management market is large, highly competitive, and highly fragmented. It is subject to rapidly evolving technology, shifting customer needs, and frequent introductions of new products and services. Two analogous players in this space are Dropbox and Box. Although the two cloud storage companies are often compared, a closer look reveals two adoption strategies have been formed. In comparing the two companies, we can aim not simply to highlight offering a similar product but more interestingly how these two companies differ.

Challenges and Opportunities 

In looking at the cloud storage space, Box is perhaps better compared to Microsoft and Google because the company since day one has been focused on enterprise customers and meeting those specific needs, whereas Dropbox has focused on customer products. In many ways, servicing enterprise customers is a lot more attractive to software-as-a-service companies. In comparison to enterprise customers, selling to consumers can be difficult because:

• Consumers need to be convinced of the value of their data: Regardless of the fact that data is useful and unique and should be protected, the vast majority of consumers just don’t care to back up their data. It should then come as no surprise that less than 10% of people back up their data weekly.
• Consumers have multiple free options: Outside of Dropbox, there are many other free cloud storage services like iCloud for Apple, SkyDrive for Microsoft, and Google Drive for Google. On the enterprise level, you would never use a free service to backup something as important as your companies’ data.
• Consumers are hard to market to: The only way a company offering a free service can be profitable is through sheer scale, and that requires convincing millions of users o sign up, users, remember, that don’t particularly value the safety of their data in the first place. Of the few that care, fewer still would be willing to pay for it. Compare this to enterprise, and all you have to do is convince the CIO and suddenly every employee will be using your service to backup data.
• For consumers, collaboration is not as important: One of the greatest utilities of Dropbox is how you can transfer and share the data, a tool that is mostly useless to consumers. The data attributes of most consumer is that it is used by them and them alone. This makes it hard to sell premium features because most consumers once again, just don’t care for them. Enterprises, on the other hand, are defined  as a collection of people and the data they jointly collect so collaboration would be a major selling point.
• Building a platform for consumers is incredibly difficult: What makes Dropbox valuable as a service is its evolution to be a platform that other apps and platforms depend on. The problem here is that the product has to be broad enough to service a large number of users and those users are not incentivized to support the platform either via monetization or collaboration with other services. Enterprise systems don’t have these drawbacks because you are building a specific solution to a specific solution, which is much easier to monetize.

Dropbox’s Pivot to Premium Business Users

In Dropbox’s S-1, the company reported “Of our 11 million paying users, approximately 30% use Dropbox for work on a Dropbox Business team plan, and we estimate that an additional 50% use Dropbox for work on an individual plan, collectively totaling approximately 80% of paying users.”

The significant takeaway here is that Dropbox has managed to increase its user base to over 500 million. Comparatively, as of Box’s last 10-K, they ave just over 58 million registered users. Taking one step further, 17% of Box’s 58 Million users were paying customers, whereas Dropbox reports 2% of their 500 Million as being paid; that’s approximately 10 Million paying users for each, but Box’s efficiency for paying customers is far greater.

Box uses a traditional sales force and sells primarily to large companies.  Box notes in its financial filings that “Our marketing strategy also depends in part on persuading users who use the free version of our service to convince decision-makers to purchase and deploy our service within their organization”. In other words, when it comes to Box’s ideal customer, the CIO decides for everyone all at once.

Dropbox, on the other hand, reports “We generate over 90% of our revenue from self-serve channels — users who purchase a subscription through our app or website.” Dropbox has a sales team, but as it notes in its S-1, the team “focuses on converting and consolidating these separate pockets of usage into a centralized deployment. Nearly all of our largest outbound deals originated as smaller self-serve deployments.”

Both companies take very different approaches which leads to different strategies. One obvious difference is the cost of acquiring customers. While the two companies spent a comparable amount on sales and marketing in 2017 ($303 million for Box, and $314 million for Dropbox, for Box that represented 60% of revenue; for Dropbox it was only 28%. These comparison fall short however, as Box’s Sales and Marketing expense includes the support for free users, whereas Dropbox’s doesn’t. Dropbox includes those in it’s COGS. For 2017, Box’s COGS was $135 Million and Dropbox’s COGS was $369 Million, leading Box to covet a Gross Margin Percentage of 73% over Dropbox’s 67%.

The following is from Box’s S-1 which outlines it’s business model.

“Our business model focuses on maximizing the lifetime value of a customer relationship. We make significant investments in acquiring new customers and believe that we will be able to achieve a positive return on these investments by retaining customers and expanding the size of our deployments within our customer base over time…

We experience a range of profitability with our customers depending in large part upon what stage of the customer phase they are in. We generally incur higher sales and marketing expenses for new customers and existing customers who are still in an expanding stage…For typical customers who are renewing their Box subscriptions, our associated sales and marketing expenses are significantly less than the revenue we recognize from those customers.”

Dropbox was less specific in their S-1.

“As we continue to innovate and optimize our go-to-market strategy, we have successfully increased monetization for subsequent cohorts. Comparing January cohorts from the last three years, at virtually every point in time after signup, the January 2017 cohort generated a higher monthly subscription amount than the January 2016 cohort, which in turn generated a higher monthly subscription amount than the January 2015 cohort.”

What the graphic above fails to inform us is where the growth in monetization is coming from. Is there an increase in revenue-per user? Are more users transferring to a premium service? etc,.

The one advantage Dropbox has over Box is the fact that Dropbox does not need to spend increasing amounts to secure paying customers, which is why Box has such huge losses.

Overall, although both companies are commonly compared as competitors in the cloud storage space, the similarities really stop at the name. Both have crafted unique strategies for handling cloud data storage. Although there are many trials with Dropbox marketing to consumers rather than enterprise customers, I believe they have made the better choice and that’s reflected in the valuation of each company. Box stands with a market cap of $3.25B while Dropbox boasts a market cap of $11.89B, still above the $10B opening valuation.

 

On March 20, 2018, Salesforce signed a definitive agreement to buy MuleSoft for $6.5 billion. The deal consisted of $36 in cash per MuleSoft share and  0.0711 of a share of Salesforce for each MuleSoft share, which at close had a value of $44.89, signifying a premium of 36% to Mulesoft’s previous close of $33.03. While the price seems high, there is a unique opportunity to be the integration layer in enterprise software. Shorty after acquiring MuleSoft, Salesforce announced the Salesforce Integration Cloud, which is where MuleSoft’s advantage in data connections will come in handy.

Salesforce plans to use MuleSoft’s expertise to jumpstart their Integration Cloud service, which consists of three broad pieces: Integration Platform, which will eventually be spun off to MuleSoft, Integration Builder, a tool that lets clients bring together a complete picture of a customer from Salesforce tools as well as across other enterprise data repositories, and finally Integration Experiences, which is designed to help brands build customized experiences based on all the information they’ve learned from the other tools.

 

What is Application Integration?

Applications Integration or Enterprise Application Integration (EAI) is the sharing of processes and data among different applications in an enterprise. For both small and large organizations alike, it has become a priority to connect dissimilar applications and leverage application collaboration across the enterprise in order to improve overall business efficiency, enhance scalability, and reduce IT costs.

There are four different levels of applications integration:

 

The User Interface Level

At the presentation level, integration is achieved by presenting several different applications as a single application with a common user interface. This older approach to integration involves using middleware technology to collect the information that a user enters into a web page or some other user interface. Presentation-level integration was previously used to integrate applications that could not otherwise be connected, but applications integration technology has since evolved and become more sophisticated, making this approach less prevalent.

Method Level

With business process integration, the logical processes required by an enterprise to conduct its business are mapped onto its IT assets, which often reside in different parts of the enterprise and increasingly in the cloud. By identifying individual actions in a workflow and approaching their IT assets as a meta-system (i.e. a system of systems), enterprises can use applications integration to define how individual applications will interact in order to automate crucial business processes, resulting in the faster delivery of goods and services to customers, reduced chances for human error, and lower operational costs. All around good things if you run a business.

Data Level

Aside from business process integration, data integration is also required for successful applications integration. If an application can’t exchange and understand data from another application, inconsistencies can arise and business processes become less efficient. Data integration is achieved by either writing code that enables each application to understand data from other applications in the enterprise or by making use of an intermediate data format that can be interpreted by both sender and receiver applications. The latter approach is preferable over the former since it scales better as enterprise systems grow in size and complexity. In both cases, access, interpretation, and data transformation are important capabilities for successfully integrating data.

Communication Level

Underlying business process and data integration are communications-level integration. This refers to how different applications within an enterprise talk to each other, either through file transfer, request/reply methods, or messaging. In many cases, applications weren’t designed to communicate with each other, requiring new technologies for enabling such communication. These include Application Programming Interfaces (APIs), which specify how applications can be called, and connectors that act as intermediaries between applications. At the communications level, it is also important to consider the architecture of interactions between applications, which can be integrated according to a point-to-point model, hub-and-spoke approach, or with an Enterprise Service Bus (ESB).

The Benefits of ESB for EAI

In today’s enterprise infrastructure, system and application integration is more and more frequently a mission-critical concern.  The wide variety of approaches and ideologies aimed at achieving this goal are proof of this fact.  When you’re just getting started researching application and data integration solutions, it’s easy to get lost in a sea of acronyms, opinions, and confusing marketing language – that’s where an ESB (and MuleSoft) comes in.

Previous models for EAI all had a similar flavor of weakness, that it was a single point of failure for the network.  In an attempt to move away from the problems caused by a previous EAI models, a new EAI model emerged – the bus.  While it still used a central routing component to pass messages from two systems, the bus architecture sought to lessen the burden placed on a single component by distributing some of the integration tasks to other parts of the network.

These components could then be grouped in various configurations via configuration files to handle any integration scenario in the most efficient way possible, and could be hosted anywhere within the infrastructure, or duplicated for scalability across large geographic regions.

Mulesoft’s Role in the Integration Cloud

MuleSoft is a software company that designs platforms for connecting applications, data, and services. Their “Anypoint” technology, which allows developers to connect different apps together so those apps can work together and share data, is a disruptive force within enterprise software markets. MuleSoft claims more than 1,000 customers, including Coca-Cola, McDonald’s, and Spotify. MuleSoft’s solutions help companies’ established networks run faster and leaner, making them ideal for businesses that are transitioning to the cloud. MuleSoft will power the new Salesforce Integration Cloud, which will enable all enterprises to surface any data–regardless of where it resides–to drive deep and intelligent customer experiences throughout a personalized 1:1 journey.

There is a huge potential Total Addressable Market (TAM) and seemingly endless product capabilities that are light years ahead of the competition. Because of this dynamic, MuleSoft is certain to continue grabbing market share against legacy incumbents in the integration software market, which is why Salesforce is so interested in acquiring.

MuleSoft’s technology would be a good fit for Salesforce and expand its total addressable market—particularly for its force.com platform, enabling faster and more complex application development, and I believe it would also enhance its integration capabilities with legacy systems. This could give Salesforce a play in the hybrid-cloud world and potentially enhance the shift to its force.com platform. Further, MuleSoft could internally help Salesforce with integrating the data between its various applications and acquisitions. Interestingly, Salesforce Ventures invested $128 million in MuleSoft before it went public, and I suspect Salesforce is a customer of MuleSoft in some capacity.

Understanding Salesforce

Salesforce is a company that helps its clients manage their customer relationships, integrate with other systems, and build their own applications. Salesforce broadly offers these products and service:

• Salesforce Applications: Salesforce includes prebuilt applications for customer relationship management (CRM) ranging from sales force automation to partner relationship management, marketing, and customer service.
• Force.com Platform: The Force.com platform is the first platform as a service (PaaS), enabling developers to create and deliver any kind of business application entirely on demand and without software. The platform also includes easy to use point-and-click customization tools to help you create solutions for your unique business requirements, without any programming experience.
• AppExchange: AppExchange is a marketplace featuring hundreds of cloud applications created by Salesforce customers, developers, and partners. Many of the applications are free and all of them are pre-integrated with Salesforce, enabling clients to easily and efficiently add functionality.
• Consulting: Salesforce provides training, support, consulting, events, best practices, and discussion boards to help clients be successful via their Salesforce.com Community.

 

Salesforce currently claims about 20% of the global CRM software market, but the company likely senses that its strongest growth potential lies outside of its current area of expertise. As companies of all shapes and sizes prepare for the next generation of technology, firms like MuleSoft—which promise to make this evolution more painless—should be extremely beneficial.

Top 10 CRM Software Vendors & 2016 CRM Applications Market Shares, Apps Run The World, December 2017

Acquisition aside, Salesforce has a number of useful tools that are being included in the Integration Cloud including a workflow tool called Lightning Flow a new service that is designed to let Salesforce customers build workflows using the customer data in Salesforce CRM. They also have a catch-all brand for building the underlying intelligence platform and integrating it into other apps called Einstein. Salesforce also threw in some Trailhead education components to help customers understand how to best make use of these tools.

Launch

This is a typical Salesforce launch. It is probably earlier than it should be, but it puts the idea of integration out there in the minds of its customers and lays a foundation for a much deeper set of products and services down the road when MuleSoft is more fully integrated into the Salesforce toolset.

For now, it’s important to understand that this deal is about using data to fuel the various pieces of the Salesforce platform and provide the Einstein intelligence layer with information from across the enterprise wherever it happens to live, whether that’s in Salesforce, another cloud application, or some on-premises legacy systems.

 

With Facebook’s business model and history with data privacy established in the last post, we are now prepared to analyze the hearing on Capitol Hill and discuss some of the commentary surrounding Facebook and regulators. While there were a few very pointed, sharp questions (especially on the part of the House) that drove some particularly meaningful dialogue, there was also a lot of confusion and missed opportunity. Even with these conflicting comments, the net effect of the hearing was expectedly lackluster. Unsurprisingly, no new information was shared and we didn’t actually learn anything new. With this possibly disappointing fact in mind, I consider the hearing a huge success for society as whole. While we didn’t learn anything new, the very fact that Zuckerberg was in D.C. at all was perhaps more important than anything that could be said verbally. The very nature of being a company of Facebook’s size and influence and pervasiveness in all aspects of society and economy is that any decision about what should be done is inherently not regulatory or legislative, but political; which makes things very complicated very fast.

Previously I hinted that the non-event characteristic of this hearing was predicable and in no way surprising. This is only half-true, however. Anyone familiar with watching these hearings unfold are complacent with nothing actually happening because nothing ever actually happens at the hearing itself, legislation comes later. Traditionally, most hearings on Capitol Hill devolve into democrats clamoring for regulation and republicans responding that regulation stalls innovation. It’s cyclical, it’s natural, and it’s predicable. This time was a little different because there was a remarkable coordination on both parties in understanding that there was a problem that needed to be solved. Now the understanding and grasp of this problem might have been misplaced, it’s worth mentioning that everyone was talking about SOMETHING. The fact that there was legitimate discussion about what needs to be done means we’re much further down the path that most hearings of this scope. Throughout the discussion, there were a number of interesting central questions that came out. All of which can be summed up to be:

• What’s going on?
• Is it a real problem?
• Does it persist currently?
• If so, what should be done?

In trying to find the answers to these questions, the hearing was characteristically distracted by trying to answer problems that were not the actual problem. Due to some of the specifics of running the debate, there were five times Zuckerberg was able to run out the clock by describing to Congress Officials that Facebook does not sell data. This frustrating realization is precisely why I feel the length of my previous post is justified. I was very clear in explaining that Facebook does not sell data and the sooner we can all understand that, the sooner we can actually drive to the heart of the problem, which certainly exists, and have a meaningful discussion. Zuckerberg was able to evade some fierce questioning several times as a result of the false narrative that Facebook sells data being proliferated simply because it sounds good. It’s extremely rhetorical, but it’s false. Is it close to the truth? Yes, but being precise matters, especially when time is a factor.

Throughout the hearing, there were a number of issues brought to attention, which makes discussing the Facebook problem difficult because the correct response is almost certainly “which one?” To that end, I feel the best course of action is to dive into each of these issues separately in ascending importance.

Cambridge Analytica

Rather than being a central point of concern, Cambridge Analytica was a symptom of a broader problem. This is partly because it’s happened before. In 2012, The Obama Administration did the exact same thing, only they didn’t have to buy the data because you’ll remember up until 2014, it was free for developers. The only difference between it happening in 2012 and 2016 are the formalities of how exactly web developers got ahold of data. Focusing on Facebook’s involvement with Cambridge Analytica and the 2016 campaign is merely a distraction and this distraction really became obvious when comparing the focus of questioning from the Senate versus the House. If you really want want to find a solution to this challenge, the first step is to get away from partisanship. If the apex of the debate over privacy is “how did this get X politician elected”, no matter how much you detest that politician, you are inherently going nowhere.

Cambridge Analytica is only important as a vehicle for prompting these hearings, but the extent to which these issues have any legs to stand on depends on how quickly we can move past Cambridge Analytica as the centerpiece of the discussion. The House hearing barely mentioned Cambridge Analytica. Instead, the focus of conversation was much more involved in the debate surrounding privacy and why Facebook was giving away data not just the Cambridge Analytica, but all sorts of companies.

Perhaps the most apt analogy I’ve heard identifying Cambridge Analytica’s role in all of this is that of pinning the events of WWI on Gavrilo Princip’s assassination of Archduke Franz Ferdinand in Sarajevo. To pin WWI on this single incident is to completely miss the point in so much as this event is merely what sparked the war. To say that we are here to discuss Cambridge Analytica is to completely miss the point. The reason I’m so eager to dismiss any question rooted in partisanship particularly is because we need to understand the reality of the situation. The losing party who would have an incentive to follow up on this issue also has no power in Congress to actually do anything. If we want real change to come, then we need to start asking the right questions about the right problems and Cambridge Analytica’s involvement in the 2016 election with Facebook’s data is not the problem at large.

Giving Away Friends’ Profiles to Developers

Whenever you’ve logged into an app or website with Facebook, that app or website would not only get all the data on you, but also the data from your friend connections. The reason this seemly controversial issue is still towards the top is because, like before, it’s in the past. This isn’t something Facebook is doing today, so to focus on this would be to give in to a sunk cost bias. As much as I believe this was a gross negligence of user’s data, to focus on it is to miss the broader problems still going on today.

By now we’ve all heard and seen news surrounding Facebook’s audit of developers to find misuse of data that wasn’t consistent with Facebook’s terms of use. As more audits are completed, I’m sure more examples of negligence will come up. This is worrisome to me because at the end of the day, it will just add more noise to the discussion. With every new update of any misuse, the media is sure to quickly jump on the opportunity to verbally bash Facebook for yet another example of mishandling sensitive data when the argument is actually stemming from the same privacy issue which was cleaned up years ago. What can Facebook really say about what happened pre-2014? The system worked exactly as designed.

To this end there are really two remedies. One option is stronger regulation with much tighter control of user data. This remedy systematically would harm new entrants into the space and the incumbents like Facebook would benefit. Zuckerberg even admitted this when asked about how regulation of this calibre would effect the company. Zuckerberg knows that Facebook has the capital to afford compliance with new regulations, but infant companies will not be as fortunate. Likewise, if you do believe that Facebook is a problem and are rightfully skeptical of this kind of regulation, there remains another form of regulatory action which I’ll discuss when it becomes obvious.

Russian Interference with the 2016 Election

To say this is a serious issue is getting at many of the point previously made with regards to Cambridge Analytica. With this discussion, people are really just looking for an excuse to blame Facebook for what happened in the election. This idea that the Russian Government somehow managed to manipulate millions of American voters through the use of ads is a red herring thrown into the debate. The idea here is that state actors were really puppeteers in shaping the viewpoints provided. In reality, the reason these agents were effective is that they were showing people what they already wanted to hear. The exploitation of division is contingent on there being division in the first place. If we were to follow this argument to its logical conclusion, you arrive at the answer that Americans have no free agency to think for themselves. Is that really the narrative we want to paint. At the end of the day, it was the American people who elected Donald Trump into office. Until that is fully accepted and we stop trying to blame or attribute agency to other people or entities, we’re not going to get anywhere.

The issue isn’t that the Russians managed to create all this division in the US, the issue perhaps is that these platforms are thriving on division. The Russians were able to pour gasoline on the fire, sure, but so was everyone else. If it wasn’t the Russians there would have been some other entity.

Fake News

The reality here is that fake news is just confirmation biases playing out. All of the research post election paints the narrative that it had no factor on the people who shared it or saw it because they already had those beliefs to begin with. The point here can neatly be tied to the “Facebook Fallacy” that Facebook sells your data. While there might be some aspect of truth, unless we’re precise in the diction, we’re just going to get distracted. This was one of the primary frustrations I had with the hearings. The questions seemed to be focused on issues that, while on margin might have some claim, distracted the conversation from far bigger issues.

The reason this is also a non-issue is that Facebook has the incentives and the technical abilities to solve this issue. We explained before that Facebook’s market position is extremely sensitive to to user activity. If Facebook as a platform were to be known universally for spreading illegitimate news sources or headlines, users would quickly stop engaging on the platform, which is disastrous for Facebook’s bottom line.

At some point we have to grant people free agency. We can’t solve every problem for everyone about the content they consume. Sometimes letting people do what they want, even through it can be mostly bad, is the best mindset to have. There’s a dark side in going the other way; that same freedom that let’s people do what they want to do, whether it be read fake news or spend too much time on gossip or clickbait, is the same freedom that lets people start new companies to create new ideas and new concepts. When you start to cut off what people are allowed to do, it’s restricting what can be created.

Content Filters

The discussion above leads soundly into a problem that is much more of an issue worth discussing. The proliferation of sharing content is both reinforcing and polarizing because people are shown only things they agree with, they’re not getting different perspectives. If you want to question how social networks impacted the 206 election, it’s most certainly around this idea than it is about fake news, Russian interference, or Cambridge Analytica.

At the end of the day, we are dealing with people’s desires to hear what they want to hear. This problem is more intractable that I think people can appreciate. The problem here is not the availability of viewpoints, but rather the implicit insistence that people should hear those viewpoints. To that, the problem does not lie with the supplier but the end user. While it’s true that we do need shared facts, if you begin thinking of solutions, it becomes very problematic very quickly. Is the solution here really that Facebook should ultimately determine what is and what is not appropriate or recommended to view? The only logical step would have to be an independent third party filtering content, and even that has a a certain Orwellian charm to it that most people would certainly not be in favor of.

The Business Model of Social Media

There’s a realization in certain circles of society that because the business model of social media companies is based on engagement, it’s in the company’s best interest to keep people engaged as long as possible. It concerns me that this business model is fundamentally not in the best interest of consumers. It also concerns me that an unelected, unofficial singular person has the final say in what over 2 billion people do with their time – the only truly limited resource.

One of the deepest, most profound problems with Facebook is Mark Zuckerberg’s tenacity to innovate the company to his vision. The problem here is that the greater your conviction that you are doing right, the wider your blind spot is to unintended consequences. If you want to criticize Facebook, you need to realize that they truly thought they were doing the right thing when in fact, they were not. Looking at Zuckerberg’s beliefs on data and privacy makes this incredibly apparent. Facebook has always been motivated by the idea that more data, more openness, and more connections are better. Getting more data to the world was always how Facebook operated. When you are driven by evangelical zeal that you are pursuing a vision that must be enacted because it is righteous, why would you you act in a way not consistent with that goal? If you want to crystalize Facebook’s “14 year apology tour”, it’s the missionary zeal of believing the company was pursuing a vision that must be enacted.

The reality is that interconnectedness comes at a cost. Interconnectedness is what allows small businesses to advertise on Facebook, but it’s also what allows Russian agents to do the very same thing. The issue here is that being mission-driven has its own ingrained problems. The more you are driven by mission, the more you are blind to negative externalities. There’s a reason that Wall Street judges performance based on the bottom line; it’s measurable and there’s accountability. Zealots throughout history are almost defined by casting away what’s taken for granted in pursuit of a mission. This sentiment is so ripe in Silicon Valley. Perhaps more than anywhere else. The extent to which people don’t council the possibility that their product or service could be used in negative ways is frightening. This entire cycle with Facebook has been a wake up call for Silicon Valley because the narrative has shifted and there’s a sudden realization that many of the these things aren’t making the world a better place. The mission is blinding executives to this possibility and it’s only when dealing with the consequences of ignoring responsibility that people are starting to realize that we should have been thinking about this the entire time.

Facebook as a Monopoly

The discussion prior segues nicely into the realization that perhaps the reason this was allowed to even go on is that we are starting to look at a monopoly, both on the digital advertising side and the social networking side.

You’ll recall previously that there are really two regulatory approaches. Although it’s easy to consider all regulation the same, the truth s such that regulation like limitations on what a company can do is one type. My skepticism about this type of regulation has been established. This form of regulation is inherently harmful to new companies and does nothing but cement the incumbents. The other sort of regulation that I am much more favorable towards is that of anti-trust regulation because the point of anti-trust regulation is to address the situation in which free market competition is no longer a meaningful limiter or constraint on a company’s actions and behaviors. It’s goal is to restore competition, not inhibit it.

• Rules based regulation limits competition
• Anti-trust regulation restores competition

Senator Sullivan of Alaska correctly implied that rules-based regulation would make the problem worse. He stated that there are really only two options for a company as big as Facebook. One is regulation, the other is break-up. You’ll recall that throughout the hearing Zuckerberg was intensely cooperative in adhering to possible regulation because he knows as well that Facebook’s position means it will be able to afford compliance. Regulation locks in incumbents because they have the ability to bear the cost of that regulation. If true competition existed in this space, how many of the problems above would have been solved?

When Facebook unwinds its ad inventory, prices go up. That’s an indicator of a digital advertising monopoly. From the perspective of people buying advertisements space, the easiest and most convenient option is to just go to Facebook because that covers both Facebook and Instagram and in terms of social media reach, that covers pretty much everything. It’s also worth mentioning that the friction involved with going around Facebook is significant and costly. On the social media side, Zuckerberg was asked several times where else consumers could go for a similar experience and the Facebook CEO was very consistent in mentioning that there eight options in terms of how people communicate. Zuckerberg failed to explicitly mention that three of those eight (37.5%) are Facebook properties. The seriousness of these implications really illuminate when you consider how haphazard company officials were in deleting some of Mark Zuckerberg’s messages to users.

The conversation becomes complicated when you try and consider answering what Facebook is doing that is anti-competitive besides simply being really successful. Certainly an argument can be made that the acquisitions are anti-competitive but when looking at a specific response, I think it’s highly likely that a rules-based regulatory response will be passed that locks down user data and prevent sharing, which would only make the competitive situation worse.

There are certainly good things that come from having one platform. Going back to the advertiser example, the fact that there is really only one place to go to advertise to reach almost everyone is a good thing if you are a small advertiser. In some regard, if the situation were such that you had to go to a plethora of different platforms to advertise, it would only benefit incumbent advertisers because they would be far too costly for the small advertiser. The fact that an advertiser can pay Facebook one time and reach billions of people across multiple platforms is extremely beneficial. You could even make an argument that this situation levels the playing field for any company that advertises its business online.

Conclusion

The reality of the space is this: there will always be a dominant player. Instead of there being an endless cycle of breaking up larger tech firms such that another can take its place, perhaps the better option is to accept the reality of the situation and put limits and rules around them and accept their benefits such that business can innovate within the ecosystem.

 

 

 

It should come as no surprise that the big story coming out this week is Mark Zuckerberg’s appearance in front of Congress earlier this week. While a few senators had some useful input, most of the questions were either flawed in direction or understanding or both. What I mean by this is that it became very apparent that the elected officials had no framework for Facebook’s business model or how that business operates with regards to web developers, advertisers, and data, which in effect allowed the Facebook CEO to turn questions around to his benefit.

Before we discuss some of the commentary (both good and bad) from the hearing, let’s clear the air once and for all about some of the speculation.

The Economics of Data

Probably the most common misnomer of Facebook’s business model rests in the idea that Facebook sells data to advertisers. This couldn’t be farther from the truth and the sooner we can get away from this, the sooner we can get to asking meaningful questions. So why wouldn’t Facebook sell data to advertisers? Well, simply put the user data is valuable. The reality of the situation is that advertisers don’t get data from companies like Facebook or Google, rather they are allowed to advertise against the data. What I mean by this is the following: Facebook does not deliver advertisers a list of users and all the data associated with the users, contrary to what some people may think. Instead, Facebook serves ads to its users. In this way, it’s easiest to view Facebook as an intermediary. While it’s true that they know about their users, they don’t just give that up to advertisers. They act as the middle-man in this transaction and allow the ads to reach the desired users essentially by allowing the advertisers to run a query targeting certain attributes. In this way, the advertisers have absolutely no way of knowing who exactly their ads are reaching specifically because they have no knowledge of the user base. There are of course ways around this, for instance if you respond directly to an ad, the company now knows that you as a user were shown the ad, but by and large the moral of the story here, which is important to get right, is that Facebook is not selling data to advertisers.

This distinction is a crucial aspect of what makes Facebook and Google so valuable. They are in complete control of the data. If either of these companies were to sell the data to advertisers, the advertisers would quickly stop paying them to run ads. Why? Because they would have all the data themselves  to run the ads without needing Facebook or Google. This is where Facebook is a little different, however. While it’s true that Facebook 100% does not sell data to advertisers, they did give away user data to web developers. I say ‘did’ because they did so previously and are not doing it now, they shut down that aspect of the business in 2014.

What was Facebook thinking giving away their data?

If you followed the logic of the previous paragraphs, you should have a basic understanding that Facebook would never sell data to advertisers because it is precisely that data that makes the company valuable. If this is indeed the case, then the question arises: What in the world was Facebook thinking just giving away their data? I think the best way to answer this question, at least for me, comes in looking at what respects Google and Facebook are different.

Google is a perfect example of an aggregator. The suppliers line up to give Google exactly what it wants. Think about it, every publisher puts up a web map that is Google-specific. Google creates all these unique taxonomies for whatever it is that the publishes want because on the other side of this equation, Google also controls the demand- the demand of how people get to these sites. Leaving anti-trust issues aside, Google is able to masterfully leverage their user experience to acquire lots of users which in turn gives them leverage over suppliers such that they can improve the product and thereby improve the user experience and get more users You can see why Google is such a great aggregator. Advertisers give Google money which improves the product, which means more suppliers of websites want to be involved, which leads to more options and a better experience for users. It’s a perfect engine firing on all cylinders, propelling itself further and further.

I think it’s safe to say that Google has more data than anyone. They get data from all over the web, from users and suppliers alike. Every site desperately wants to be in Google. If you have a small blog talking about tech, you want to be on Google. If you have a large company, you want to be on Google. Even Amazon wants to be on Google. If you search for a product on Google, often times Amazon is the first response and Amazon is happy about that. I think it’s also safe to say that Google’s hold on data is superior to everybody else and that’s because of the way the value chain works for Google. Google is so unique because every participant is so happy and eager to give data to Google simply because it’s in everyone’s best interest to give that data to Google.

Now let’s contrast this with Facebook. Going against Google directly in this space would never work but if you create your own closed ecosystem, a walled garden of sorts, where Google can’t reach in, then maybe you can create an environment where you can benefit in the same exact way and Google can’t get involved. To make this point clear, you need to understand that Facebook and Google as search entities have no connection with each other. Whatever you type in Google never interacts with the Facebook ecosystem and vice versa.

What makes Facebook so unique is that their content suppliers are also their content consumers and because of this dichotomy, there becomes a supercharged viral loop where those demanding the content are also creating the content, which subsequently increases demand. This interaction is exactly why Facebook has the tightest viral loop in the history of business, because the suppliers and consumers are the same people. The speed at which this loop runs is unprecedented too because it’s all happening in Facebook’s platform. Because this is Facebook’s platform, they can close it off from Google and build up an incredible repository of Data that Google can’t touch and is unique to Facebook and this is exactly where the problem lies. The types of data that Facebook and Google have are different. Google has data from all over the web given to Google willingly. Whereas Facebook has an astronomical amount of personal data, ranging from your likes, comments, and interactions. But while Facebook has all of this personal data, they have a distinct scarcity of all that other data on the web. Facebook was very concentrated in the type of data it had and (once again) leaving privacy issues aside for a moment, you can see a very sound strategic thinking develop. Facebook has an excess of data in this one area and they will trade that data to get some of this missing data that it’s lacking; like restaurant data or travel data or local business data.

Let’s look at exactly how a trade like this would play out:

If Facebook does this trade with Yelp, Facebook now has its own user data and now local business data. Likewise, Yelp has its own data and now has some Facebook user data. Now Facebook trades with Trip Advisor, leaving Facebook with new data on travel. If you follow the aftermath of these trades, you realize that Yelp and Trip Advisor are made marginally better off because they now have new data with which to serve users, but if you look at Facebook, Facebook now has personal data plus restaurant data plus travel data. They get all of it. If you can now visualize Facebook at the epicenter of this data exchange where all these companies get Facebook’s data and Facebook gets the data from all the other other companies, the net result is Facebook has all of the data.

The Facebook Fallacy

The question persists, if you are sitting on the most valuable advertising business in the world and the critical component of that value is the data on users, why on Earth are you giving away the data? I think in analyzing the differences in Google and Facebook helped to uncover the answer to this question which is: Facebook wasn’t trying to be an advertising company (or at the very least if they were, they didn’t want to be). At the end of the day, Facebook wanted to be a platform and what do you need to be a platform? You need developers. And what have you got that will bring developers into the platform? You have all this personal user data. The crux of this fallacy is such that instead of focusing on what they were, an advertising company in which the correct strategy is to keep their user data safe, Facebook instead thought of themselves as a platform. Platforms are sought after as the gold standard in Silicon Valley, and once you understand this narrative you begin to understand why Facebook was giving away their valuable data. I think in looking at how this has played out, Facebook got incredibly lucky by essentially being forced to back into this incredibly successful business model as an advertising company and leaving the old platform model behind.

To be clear:

• An ad company: You harvest user attention so you can serve them ads.
• A platform: You harvest user attention such that other companies can build on top of you and you take a tax along the way.
  • Windows and apps
  • Amazon with AWS
  • Apple with the App Store

At the end of the day however, Facebook isn’t a platform, it’s an advertising business. It’s crucial to recognize that you can’t be both because they are competing with each other for attention. Either you are featuring developers or you are featuring ads. A perfect example of this is Android, which doesn’t have ads because it’s a platform.

The heart of the issue

The problem was never that Facebook was an ad company, the problem was that Facebook refused to accept that it was an ad company. Had they accepted they were an ad company, they never would have been so cavalier with the data because the incentive of an advertising company is not to give away their data. When looking at the advertising space, it would be easy to think that Facebook and Google’s dominance means that they can get away with almost anything, but the modern digital era has created an environment where this is not necessarily the case. Facebook is heavily incentivized to keep their users happy, Google is heavily incentivized to keep their users happy because if their power in the market rests in their users, the greatest threat to them is, by extension, users growing disenchanted towards the product.

My concern for Facebook doesn’t rest in its business as an advertising company and any association thereof, but rather in how Facebook keeps user attention. If we imagine for a moment that Facebook had not mistaken their business strategy and instead acted as an advertising company the whole time, I’m still concerned in the case of Facebook that the incentives for the company aren’t necessarily aligned with the long-term wellbeing of its users. Comparing Google to Facebook once more, if you want results from Google, you type the word or phrase in and Google’s incentive is to provide you with the most accurate result and then send you on your way. Facebook’s is to keep you coming back and there’s a degree of addiction in getting that incentive to work that is deeply troubling to me.

Possible Regulation

Because Google’s power derives from their place in the value chain, their long-term regulatory framework is really that of anti-trust. Facebook, on the other hand, is much more political. On an acquisition basis, Facebook should have never been allowed to acquire their way into social media dominance with Instagram (and to a lesser extent tbh). It’s interesting to think about what that political regulation might look like because the answer is almost certainly legislation stating “don’t share your data with anyone else” and this likely regulation is incredibly beneficial to Facebook because they already have all the data. Passing legislation that would ban the sharing of data means that Facebook wouldn’t need to share data with any would-be competitors.

If this narrative plays out, Facebook is essentially throwing the entire industry under a fire by cementing themselves as the ever-lasting incumbent. With no more openness, no more sharing, and no more API, the only companies that benefit are the companies that already have the data. The entire process of creative destruction innovating and taking the place of the old becomes disrupted. This is such a challenge because if you believe that Facebook is too large and too powerful and you want a competitor to come along, you need data portability if that competitor is going to stand any chance. Yes Facebook may get regulated, but that regulation almost certainly will do nothing but entrench Facebook’s position and make it that much harder for a competitor to come along, especially if that competitor can no longer take the same data shortcuts that Facebook has.

There is no perfect solution where you have perfect privacy and also perfect competition, there has to be balance. Rather than incorrectly discuss the fallacies of the situation, we should be asking much broader questions. Should one company be able to collect all this data? Or is this just the nature of data. Every one of us is sending data with almost everything that we do; whether it’s our IP address, our cookies, or the location tracking in our phones. So much data is being handed out in the virtual space that the data is inherently infinite and should we really be trying to limit the exchange of this data?

 

*Editor’s note

Because this post ran a little longer than I had anticipated, I will commit to writing a detailed account of the specifics of the hearing within the next few days. Think of this post as a primer for what’s to come. Now that we understand Facebook’s current situation a little better, perhaps we will now be better equipped to analyze the real problems plaguing Facebook and take a deeper look into how regulation will affect the social media giant.

If you haven’t watched the entire hearing, you can click here.

See you for part 2!

 

 

 

 

Earlier this week, Spotify Technologies (NYSE:SPOT) launched its public shares on the market, opening at a per-share price of $132. The music streaming giant launched with an equity price of around $26.5 billion, which lands the Sweden-based company among the top 10 largest tech IPOs ever. Needless to say, Spotify going public was a big deal, but is the company worth the hype?

SaaS Overview

The business model of most Software as a Service companies is actually quite simple: low marginal costs and high fixed costs. It’s important to be discriminatory when looking at cost structures of any business. The former (which I’ll refer to as marginal costs) are useful in identifying a company’s margins, while the latter (which I’ll call operating or fixed costs) affect overall profitability. Not surprisingly, the structure of SaaS companies that I listed above is extremely attractive. With high margins, you benefit exponentially with growth because every unit you produce and sell slowly eats away at those fixed costs, like R&D, until you make a profit. This scenario is the “perfect storm” for investors because revenues will outpace costs fairly quickly and the result means a more attractive bottom line.

Spotify’s Fixed Costs

Make no mistake that Spotify is a SaaS firm. Remember that when looking at how valuable software companies are, fixed costs really don’t matter long-term because with enough users, they will pay for themselves. In some cases, high R&D can even be a positive Key Performance Indicator (KPI). Spotify has been diligent in increasing user revenue and has managed to grow revenue 448% over the last five years.

This is impressive, especially considering the company offers its service for free with ad-supported revenues, with a secondary option to pay for a premium subscription (ad-free). This premium subscription represents 90% of the company’s revenues and has grown a respectable 180%, compared to ad-supported revenue growing around 245%. Regardless of where the revenue is generated, this is good news for the company at a first look because once a user is integrated, conversion to premium becomes a likely possibility.

With these revenues in mind, let’s look at Spotify’s operational costs, which have tracked revenue growth nicely:

Here we see that R&D has roughly tracked revenues at around 10% and Selling, General, & Administrative (SGA) expenses have tracked nicely at around 20%. Looking at this, the company’s operations will look something like this.

There’s a problem, however, and that’s Spotify’s marginal costs.

Spotify’s Marginal Cost Disaster

The reason most SaaS companies can operate under such low marginal costs is because it turns out that storing data on a server is actually quite inexpensive. The company’s individual cost per user is essentially nothing, they are not making a product which requires direct materials or direct labor, they are providing a service which has none of this. So why doesn’t Spotify fit the mold?

Boasting a Cost of Goods Sold of above 80% of revenues, Spotify has a significant problem with marginal costs. But why, the company is a media streaming application that stores data in the cloud and rents that data to users? The answer: the royalties it must pay on the music it streams. Spotify has to license the music it streams either to record labels or songwriters. This means that the company is completely at the mercy of these record labels, which can essentially set any rate to be paid and has the ability to change rates with every new contract. Taking into account these marginal costs, Spotify not only isn’t profitable, but it’s losses are growing.

Conclusion

Spotify is a unique case in terms of most SaaS companies because its marginal costs relative to revenues far exceed most zero-marginal-cost based SaaS firms. At the end of the day, Spotify and its managers have no control over the company’s marginal costs because they are determined by record labels/publishers. With this in mind, Spotify is only left with two options. Either (1) they cut back their fixed costs (which are already quite low by industry averages for technology companies), or (2) they can continue to grow their user base such that they make up for the high COGS. I think this point is best made with an example.

• Imagine Firm A has Revenues of $10/unit, COGS of $1/unit and fixed costs of $100. Each unit sold contributes to $9 to bottom line performance. In this example, the firm need only sell 12 units before the company is profitable.
• Spotify more closely represents Firm B, which we’ll again say has revenues of $10/unit and fixed costs of $100. Only this time the COGS is $8/unit. Here, each unit sold contributes only $2 of margin and the firm will need to sell 50 units before it is profitable, four times as many as Firm A!

In looking at both options, I find myself doubtful that either will occur. As I previously stated, Spotify has done a great job efficiently controlling Operating Costs with growth, and I don’t see them cutting back here. As far as reach goes, Spotify has done a great job integrating with other services and creating a strong presence internationally. Compared to Apple Music, Spotify has much more of the international market with nearly twice as many paid subscribers than Apple Music, and as the middle class of these countries develop, that means potentially more subscribers. In the U.S., however, there is a different narrative playing out. As is stands, Apple Music is primed to take over Spotify in terms of paid subscribers by summer 2018. Apple Music has grown U.S. subscribers 5% annually compared to 2% for Spotify.

Overall this leaves me with the question of where Spotify will be in the future and why they even went public in the first place. To answer this question we have to look at what public equity even does for a company – it’s either a source of funding or a way of cashing out. If Spotify were a more traditional SaaS firm and this were 1980, I would tell you that going public meant great things for the company because they would use the funding to grow operations and as the operations grow, the high margins will reap the benefits of new users and the company will report great gains. This, however, is not the case for Spotify. Make no mistake that Spotify has already done a phenomenal job expanding their operations both domestically and internationally. Spotify has no doubt been through multiple rounds of private venture funding and in going public, the owners get their pay out and everyone goes home happy.

Taking all of this into account, Spotify rests in a peculiar position. As someone motivated by how success is defined in this industry, I feel comfortable saying Spotify has been incredibly successful without being hyperbolic. Spotify has brilliantly expanded not only its core business but created its own network as well. With every home speaker that utilizes Spotify, the service is becoming more and more ubiquitous. This network offers companies precisely the sort of scalable user generation that makes profitability an inevitability. While I’ve enjoyed watching Spotify develop its network in hopes of doing just this, I find myself still skeptical of long-term sustainable growth because without a doubt, the company just doesn’t have control over its marginal costs and that may be the downfall of this company. Unless Spotify can create its own record label and source content there, I fear the music industry will bleed this service dry.

Appendix: A Final Warning

I feel that I must preface this warning by stating that I am a Spotify Premium customer and I happily pay Spotify every month for access to my favorite songs. With that out of the way, I have some fears which solidify when looking at the streaming industry as a whole and why, once again, Spotify is an outlier. Think of Netflix, for example, which similarly must license shows and movies to be essentially rented by users for a monthly price. In comparing these two services, you have to recognize a fundamental difference in how the media is consumed. With shows and movies, you may watch a classic once or twice, but thats’s it. Now compare this to Spotify, where you have a remarkable back-catalogue of songs that you will listen to on repeat over and over again. When Netflix was forced to raise prices, as a user you had a very simple decision to make: “are the few shows I watch every week worth the price increase?” If the answer is no, you would simply cancel the service and your opportunity cost is watching maybe one movie a week or one show a day. Compare this to Spotify now and the issue becomes clear. Spotify does not have the one song or album that you would miss out on, it has every song you ever listened to or want to listen to. Whereas Netflix had an opportunity cost of maybe 10 shows a week, canceling a Spotify subscription would mean you lose hundreds of listens a week. I have no knowledge of whether Spotify will be forced to increase prices like Netflix has, but I do know that Spotify gets its content from the same exact place, the media industry, and that makes me weary of price increases. If the record labels negotiate a higher price for Spotify, Spotify will be forced to charge users a higher price in return. If enough users are unhappy with this dynamic, it paves the way for another company to take the lead. I look at Apple Music in this case. Apple Music is not Apple’s primary source of revenue and that means if COGS for streaming music increases with higher rates charged by labels, Apple Music has the benefit of being able to sustain losses and not be in danger of bankrupting the company at large. At this point, it’s a war of attrition between the music streaming giants. A war of attrition that with 100% confidence I can tell you Apple will win with its cash reserves larger than the GDP of some developed economies.

I can think of no better analogy with which to launch this blog that that of a blank canvas. The empty canvas represents limitless opportunity in the hands of the artist/writer. Perhaps the next T.S. Eliot-esque ballad will be composed, qualifying the methods of disruption and value. Or perhaps the romanticized theatre of Hemingway can convincingly hides the ugly truths while highlighting the beauty of human nature.  Suffice it to say that the blank pages could be used to create and share all the things we find wonderful and just.

While I am not a literary expert in trade, rest assured every story will have a purpose and every argument will not be made with haste. The writers of yesteryear were motivated in deconstructing the complexity of the world around them and creating a story to send a message. I aim to do just that in the tech space. While there are many great sites that follow this sector closely, often times the angle is biased and context is lost. More often than not, I find myself walking away with little-to-nothing. This is precisely why I am creating this platform, to provide the world a place to visit to answer the questions that drive conversation in this space.

For now, my commitment will be to write one big-picture blog a week related to the technology sector. While my advantage lies more on the financial analysis of technology companies, I aim to take a more general approach and tell a story that paints a picture of the space.

Keep up @silicon_sector, with RSS, a weekly email update, or just bookmark this page. I look forward to making it worth your attention.