Marco's Blog

All content personal opinions or work.
en eo

Ellen Pao or the End of Internet Capitalism

2015-07-10 11 min read Essays marco

Have you been following the Internet, lately? Its self-declared home page,, has been splattered all over the news after the interim CEO, Ellen Pao, caused a major uproar and finally had to resign.

If you’ve never been there, the idea of reddit is neither new nor innovative. Its users are grouped into named categories, called subreddits. They can post entries to the subreddits, and other users can vote them up or down. When you then go to the subreddit, the entries that have been viewed most favorably are on top.

This is an old idea. Slashdot, a site devoted to geeks, has been working like this for a very long time. Digg, which seemed to be advancing towards the top of the Internet, was pretty much the same as reddit. While the former is still limping along, the latter was sold at a fire sale and turned back to irrelevance.

It’s not only news selection sites that are affected, though. The same happened to places like MySpace, which ruled the Internet for a while and then disappeared into oblivion, to be rapidly replaced by Facebook. Interestingly, while the two sites came from completely different angles, they ended up being pretty much the same in the end.

Why, then, did Digg and MySpace disappear so completely and so rapidly? The answer is patently obvious and very simple: users left the two sites in droves to go to a new community. The real question, though, doesn’t have such an easy answer. Why did the users leave in droves in the first place?

Back in the day of MySpace’s massive failure, people were talking mostly in terms of fickleness of users. It’s too easy to go to a new community, they said. The switching cost is low. Cool kids want to leave the place as soon as the parents show up. Facebook was invitation only, which made it automatically cool. Plus, it didn’t have the atrocious ads that MySpace was famous for.

There are two nuggets of wisdom to extract from that (somewhat simplistic) analysis. First, switching costs are low. That means it’s easy to find a different site that offers comparable functionality at the same price (free). This, in turn, is rooted in the fact that the cost of creating a site are small. Nowadays, with the ready availability of massive computing clouds, an entrepreneur doesn’t even have to scale up computing infrastructure: Amazon AWS and Microsoft’s Azure can do it for her.

The second nugget of wisdom is hidden in the ads that MySpace was famous for. Most cool sites become instantly uncool as soon as monetization starts, or soon thereafter. Everybody hates ads, right?

Only that there are businesses that make loads of money (cough, Google, cough) by presenting ads that are actually useful to people. It’s not true that ads make a company uncool and make users flee. I would say anecdotal evidence still bears the correlation between monetizing and decline of any major site that ended up irrelevant.

In a different essay, I wrote how Internet startups are ill-suited for the typical structure of a corporation. Instead of being factories with large capital requirements, they are workshops. They risk little capital, but lots of time. They become wildly successful or are miserable failures based on the ideas and execution, not based on resources and capital.

Still, and that’s very surprising, they are incorporated with the same logic as a steel mill. The allocation of shares (ownership, a capital thing) is what matters. Owners of shares choose the board of directors, which in turn chooses the executive, who in turn choose a cascade of employees.

Which is exactly how Ellen Pao became the CEO of Reddit. Her connections with venture capital allowed her to rise, despite her lawsuit for gender discrimination against her former venture capital employer failing miserably. I think this is a sign of a healthy environment: she sued for gender discrimination, she lost, and she’s still seen as an outstanding candidate.

In turn, this is exactly why Ellen Pao failed: first, she forced the shutdown of several of those subreddits where people where “hating” on others (fat people, in particular); second, she fired a crucial employee, Victoria Taylor, who had been shepherding one of the most successful features of the site, the series called “Ask Me Anything” (AMA for short).

Now, the Internet’s uproar was enormous. First, the ama subreddit shut down completely, followed by several other very popular subreddits. The moderators of these subreddits are all volunteers, so they had nothing to fear by their action. Which may have been the undoing of Ellen Pao: faced with the fact that, next time around, the whole site might just be off the air, it was better to let her go than to risk a repeat offense.

Ellen Pao, it should be noted, didn’t do anything horrible or terrible. She didn’t even do anything stupid or adversarial. She simply instituted rules of conduct (no fat people hating!) that are otherwise commonplace and she fired an employee that (apparently) refused to do as she was told, or at least put up resistance.

Ellen Pao’s problem is systemic. It is the same problem that hit MySpace and Digg, and an untold number of companies: once capital requirements go down and work becomes the paramount value of a company, changes naturally made under a capital ownership structure easily run afoul of the actual value of the company.

What do I mean by that? I mean that the value of a company is both in its capital assets (the servers, the connectivity, the original idea) and in the ongoing work performed (users sharing funny links and media). Our corporational structure, though, considers only the capital assets when it determines who should run a company.

This was widely visible in the past, too. In fact, the corporate raider scheme of the 80s, the Hostile Takeover, very specifically addresses the temporary imbalance between value and assets. It takes over a corporation for the specific purpose of dismantling it, because its pieces are more valuable than the whole.

Now, back in the 80s the logic was still favoring capitalism as it stood. The companies dismembered had done a poor job at managing themselves. The layoffs were painful, but new and better companies would soon follow in the footsteps of those that had been hostillely taken over. Many new jobs would be created, and the end result would be a recovery. Periods of unemployment or required relocation were simply minor and temporary inconveniences.

In the second decade of the 21st century, things are largely different. The main difference here is that there simply aren’t a whole lot of capital assets. I mean, there still are, of course, but capital is starting to be commoditized. Large infrastructure finds it easier to rent out to entrepreneurs than to come up with its own ideas. It is easier for Amazon to offer AWS services than to come up with something for the servers to do.

This requires a quite fundamental change in the way we look at companies. If the value of a company is not in its capital assets any longer, does it make sense for the company to be steered by its capital assets? The question is not an ethical one, but practical. If the majority of the value in a company is determined by, say, its workforce, isn’t it a tragic risk for the company to make its decisions based on what’s best for the capital assets?

To bring the point in perspective, here an unrelated story. Berlin, the capital of Germany, has long attracted immigrants. Many of them pooled into the city because it was more tolerant in general than other places in Germany, because there was work to be had, and because friends and family already lived there. In fact, immigration was so strong that entire neighborhoods were almost exclusively inhabited by immigrants. The most famous example was Kreuzberg, which was so beset by white flight that only a handful of citizens were “ethnically” German.

When it came to elections for the neighborhood council, though, only the handful of German citizens were allowed to vote. Decisions were made based on this handful of people’s biases, without consideration for the greater community. This meant that any infrastructure investment was consistently nixed, since it would have mostly benefited the immigrants. Also, because there were so few voting residents, results of elections would vary greatly depending on the year.

This anecdote works in two ways. First, it exemplifies the consequences of having decisions made by a minority of the value (democracy being fixed on one man – one vote). Second, the ethics of excluding the majority of the neighborhood’s population from voting are not considered, just as was the case when talking about decision making in Internet startups.

The fact is, making decisions based on a minority of the value is a really bad idea, as Digg and MySpace found out. It is a very bad business idea. It kills your business, potentially.

So, what to do about it?

Germans, the quirky innovators that they are, came up with an interesting solution after WW2. Boards of directors of publicly trader company are mandated by the government to be composed of 50% union representatives. That is, the company’s steering committee (not the executives) is controlled half by capital and half by work. The system works extremely well for Germany with all its flaws.

The main flaw is that unions are a very poor proxy for work value. They are structured around democratic principles, which gives the janitor the same voting rights as the creative director. There is nothing wrong with being a janitor, of course, but making decisions based on what’s best for the janitor seems a strange way of getting a company to move forward. In fact, I witnessed myself a company who went belly up because the CEO was best friends with the support person and made an irrational business decision to avoid her receiving more support calls.

Similarly, I also witnessed a company that made a similarly foolish decision about voting rights of owners. It was an LLC with three owners, one of whom owned 90% of the company, with the other two sharing the remaining 10% at 5% each. The way the agreement was written, voting rights were assigned per person and not by capital investment. The result was that the minority of the capital made all the decisions.

Again, this is not necessarily bad. But it skews decisions against the interests of capital. In fact, what ended up happening was that the profits from the business ended up shared in three equal parts instead of according to investment. At some point, the owner of the majority of the capital decided that was not in his best interest and had the LLC dissolved. That ruined the business for all three of them.

Fortunately, there is an easy way to solve the conundrum for many companies: consider the salaries of employees as their voting right, as much as you’d consider the size of the capital investment. That means that a company would have to figure out how much of its value is in the work performed, and how much of it is in the capital infrastructure. Then the board of directors would be composed based on those voting rights.

Things would be simplified enormously if we’d just consider the salaries at the same level as the capital. A company without any assets, that just hires people to perform work, would then be run 50% by the workers, and 50% by the capital that pays for them. If capital assets are required, then more than 50% of the company would be owned by capital.

Workers would have voting rights based on their value to the company, which is their salary. This would put undue voting in the hands of managers, who in actuality do not have any of the value, since they are coordinating in function. Also, their salaries are so much higher than those of workers that their voting would skew results in favor of management. An equitable solution to this would be to take the salaries of managers and apportioning to the people that work for them, and only give those voting rights.

For instance, if the Vice-President of Engineering has four Directors reporting, and the Directors each have five Senior Engineers (a case I am familiar with), then each of the Senior Engineers would get voting rights based on their salary plus one fifth of their Director’s salary plus one twentieth of the Vice-President’s salary (or total compensation).

One should note that there is the possibility that total salaries outstrip capital investment quickly, as soon as revenue starts flowing in. Companies in high tech typically have profit margins high enough that this would be the case. Also, the capital investment in a company’s shares is typically higher than their nominal value. Thing have to be figured out, for sure. The basic principle is sound: workers should count in the decision-making processes of a company, not just capital. The details are left to heuristics.

Which gets us back to Reddit, because the case of volunteer work is not covered by this notion. The value of the volunteer moderators is not captured in salaries. The same is true for the real value of companies like Facebook, which is the work that users put in. Nobody goes to Facebook because it’s pretty, because it’s fast, because it’s useful. People go to Facebook because there are other people on Facebook. The moment people start not to be on Facebook, the site dies.

And there is no part of the decision making process that takes that into account. And as with all things logical, this one will eventually bite. Facebook’s problem, just like Ellen Pao’s problem at Reddit, is that decision making is decoupled from value. As long as that’s the case, value is bound to eventually go down.