Facebook is here to stay

An interesting thing about being plugged into the startup world is the frequency with which one sees “new ideas”. Many of them, alas, are just old ideas being rebranded or old ideas that the person proposing them is not aware of.

Facebook’s widely documented travails, with legislators inquiring into Facebook’s role in the 2016 US Presidential elections – and possibly more – on both sides of the Atlantic, mean there is once again a swell of entrepreneurs clamouring to make a “new Facebook”. Within just the last week, one tweeted he wanted to make a privacy-first, open source Facebook. Another in a closed founders’ group described a portfolio company, which is a combination of “Facebook and Gmail.. no spying, no ads, secure social platform, where you own the content and the network .. no fake users, ever”.

Which reminded me of another interesting thing about being an old hand on the web: we know what hasn’t worked, exactly why it didn’t work, and all that it may take to make it work.

Anyone remember Diaspora? It was founded on three principles: decentralisation, freedom, and privacy. But as social networks go, how many people do you know who use Diaspora? I admit I know none.

Then there was ‘Ello, which I am somehow still on, though yet to figure out how it adds any value to my life. May be I am in the minority. There are and have been others such as Path (uh!), Google+, MySpace, and Peach (ok, I admit ignorance).

Successful ones include Whatsapp (also owned by Facebook), which is embroiled in its own version of “fake news” controversy (link may require registration) in the world’s largest democracy. Others are Slack, with its purpose built groups, and Quora, the knowledge community with over a 100 million users.

But none has yet matched Facebook’s near-total control of the conversations among the global user base of connected, mobile, literate people.

An early adopter, I was on SixDegrees.com somewhere around 1997. Many people discussing these networks have not heard about it though a Wikipedia page exists, as it should, for the super early mover in the social networking space. My observations through the years say that we all want privacy, no-ads, control over who sees our content, no fake names or bots, anonymity when we want to overshare or share stuff we should not be sharing, ability to join or leave groups, ability to engage safely, and the freedom to take our data when we leave. However few of us really know what it takes to build something like that. Fewer still know what it takes of us to ensure all of that on the platforms we use.

Let’s take the privacy and control pitch. Facebook has excellent, granular privacy controls. I should know. I operate my Facebook like a walled garden. To work it, however, you have to be alert to how your life is organised, make appropriate lists, and then be diligent about controlling who sees what out of your shared material. As I have written elsewhere, user motivation to figure out your product is a huge – and unhelpful – design assumption. The product and the UI need to be simple to use and easy to figure out.

Google+ made much noise about making privacy and control simpler and easier but Google really did not succeed at converting a sizeable chunk of its Gmail user base into Google+ users. This highlights the importance of switching costs and network effects in building a successful social network.

Then there is that bugbear of fake names, bots, and anonymity. Quora is by far and away my favourite community since 2010. It started off well. With a “real names” policy. That is how so many of us participate and write there with our real identities. Monitoring and policing real identities is a job and a half. There is a reporting feature, which active users, Top Writers, former moderators, and topic gnomes use heavily. It is difficult for users to keep on reporting if they keep seeing the reported users come back with changed names, or bots and sock puppets returning with a vengeance to vandalise content. It reduces the SNR on the website and can damage the culture of the community rapidly. It takes lots of people. Or machines. To keep bots, fake names, and anonymous usage under control. Oh, and real names mean nothing. On LinkedIn, people use real names and their employer names and yet there is open and inbox harassment, racist remarks, trashy comments on people’s posts.

Which brings me to the problem of “using the web while <insert minority type here>“. Did I say Quora is my favourite community on the web? But I am asocial as hell there. Others cannot comment on my content, nor can they message me. How did that come about? I was enticed to Quora by the purpose of sharing information and learning. At one point, there was a massive growth through influx of users, who did not understand or care about community policies, brought their own cultural artifacts and assumptions, and had a considerable impact on the experience in the community. I was quick and preemptive in battening down the hatches but others continue to suffer – Muslims, Jews, black people both Afro-Caribbean and African American, LGBTQIA persons, and women in general. This is a problem that no social network has licked yet though noises are often made in this regard.

And then there is the money question. In 2015, Facebook spent a reported $2.5Bn on capex including data centres and other infrastructure. LinkedIn is owned by Microsoft which has almost $90Bn in cash, Whatsapp is owned by deep-pocketed Facebook, Quora is a unicorn valuated at around $1.8Bn. The point is that even if the perfect product is created, and somehow users can be enticed to switch in huge numbers, creating and running a social network at scale is expensive business. One must not forget that users are used to “free”, so promises such as “no ads”, no data mining etc will lead to inevitable questions about the monetisation model.

All this can be summarised as “barriers to entry” in the social networking space.

And yet, every other day, there is an aspiration to create a new social network.

Back to Facebook then. With over 1 Bn users, Facebook is no longer a “social network”. Especially if information — fake or otherwise, and I am including Whatsapp’s challenges in this — is the stock-in-trade, Facebook fits the classical definition of “utilities” and, at the moment, it is also a natural monopoly. It is not a public utility. If, however, its externalities are anything to go by, including its impact on democracy and the information asymmetry created by its machine learning algorithms, it needs to be regulated. Those arguments have been made repeatedly over the last few years. One of my favourite commentators, danah boyd, wrote about it in 2010.

So how might Facebook be regulated? And will that reduce or increase the “barriers to entry” for new networks?

One of the best models of utilities regulation is the British one where regulation is seen as a second choice to a well functioning market. It focuses on consumer choice, competition, and forward-looking incentive regulation. Forward-looking what?

There lies the rub. None of us is paying for using Facebook. In fact, if pricing were introduced at this point, there will be an almighty uproar because, to many, it is like an “essential product” now.

If no consumer is paying to use Facebook, is it really a “market”?

As definitions go, we are in uncharted waters.

More importantly, how will regulating ensure or improve consumer choice or competition?

It is structural barriers, and consumer behaviour challenges, not Facebook, that prevent alternative social networks from achieving the same roaring success it has achieved.

In other words, unless regulators break Facebook up perhaps into consumer and business networks or force Facebook to shut down, or Facebook boldly starts charging fees, eroding its user base and reducing its own power, or an earth shattering paradigm emerges in economics and business regulation, Facebook is here to stay.

It may be forced to become more transparent, and build better governance like other listed entities. But it is here to stay.

 

%d bloggers like this: