April 26, 2023

What Journalists get wrong about Twitter an Mastodon

It's a bit like watching someone repeatedly getting conned into playing a shell game. But hey, what can you do, right?

There’s this funny thing, I keep hearing from Journalists over and over again, when comparing Mastodon and Twitter.

We get way more engagement on Mastodon than on Twitter, but Mastodon only sends us a a fraction of the visitors and nothing ever goes viral.

Paraphrased

In other words: being on Mastodon is nice, but if your job is about getting attention (to get those ad dollars rolling), just stick with Twitter. Case closed, right? Well, not exactly. Time for an analysis.

When rendered as a bar chart, access statistics look a bit like being shown the middle finger, don't they?
I’ve had server meltdowns from going viral in the past. Whenever that happens, the logs always show exaclty the same pattern: it starts as a small trickle of extra requests that grows to a flood within minutes, peaks for about an hour, then dies down to baseline again over the next couple of days. The peak itself is short lived and flanked by a sharp rise and fall. Sometimes, there’s an echo the next day, but usually it’s just a single spike.

Most referals will only request a single page. That’s it. Veni, vidi, goodbye. They just bounce 99% of the time without any further interaction.

Now, it would be easy to argue (and probably a lot of people do): “So what? Bandwidth is cheap! If I have to entertain 99 window shoppers before someone comes in and buys something, then I’ll just have to go viral more often!”

It is common manager knowledge, after all, that even a tiny profit margin will become lucrative, if you can jsut figure out how to scale things up enough. And that’s exactly where the error in thinking lies here: the assumption that scaling is possible.

The game is rigged

Things do not “go viral” on CSM (Corporate Social Media), they are made viral.

No social media platform is build without moderation tools that allow for boosting or throttling of content (not even Mastodon is an exception to this), ultimately allowing the platform’s owner to focus or divert attention at will.

Due to federation, attention/traffic from Mastodon is (mostly, but not entirely) organic. Things hardly go viral there because the requirement for that would be to post a piece of content that is so valuable and universally agreeable that it becomes an imperative for everyone to boost. This is a rare beast and the average daily news just doesn’t cut the mustard in the Fediverse.

CSM, on the other hand, knows all to well that journalists, bloggers and website owners in general just love to see those traffic spikes in their analytics, as well as feel the dopamine kick that comes along with it. It’s an itch to scratch, there’s money to be made from that and they have the means to deliver.

The game, all CSM are playing, is to burn venture capital on attracting users with the promise of a free service that finds them the content, they are truly interested in. Of course, a free service will never pay the bills. That’s where the business user comes in, who needs to be convinced that the social network can deliver eyeballs to show products to, drive sales and earn a commission fee for it.

Of course, like every good carny game, this one is rigged as well…

Carnies employ a marvelous array of rigged devices and deceptive methods that allow the operator to determine who wins and who looses a game. They may string a player along just enough to keep him interested or even allow a clueless sucker to win a giant teddy bear (one of the cheaply produced kind) that is then paraded across the grounds, inadvertently advertising, what is otherwise an unbeatable game (sounds familiar, yet?)

Let’s say, the average CSM user expects the search/suggestion algorithm to be right 95% of the time. But the algorithm actually manages a whopping 98% percent. Where does the surplus go? Allocating it to the user has no benefit and, in fact, will only raise his expectations. Instead, the smart strategy is to fill those 3 extra slots with content from a business user, who is next in line for winning a giant teddy bear.

The regular user, who is confident in the algorithm, checks (for curiosities sake) all suggested content, finds that the 3 placed pieces do not interest him after all, but attributes them to the 5% percent tolerated error rate and simply bounces, resulting in exactly the kind of webserver statistics, shown above.

Proofing foul play here is virtually impossible because it is cleverly hidden in the error margin.

The game is built on dreams and unrealistic expectations

With Mastodon, attention is genuine, based on actual interest and scaled proportionally (aka “the real thing”).

CSM, on the other hand, more or less sends a human bot army to artificially inflate a business user’s analytics. In a nutshell, this comes down to running an ad campaign in his server logs in the hope of convincing him to:

  • Believe that CSM can drive huge volumes of targeted traffic, if he plays the game.
  • Bring in friends&family to play the game with them.
  • Contribute more free content to the CSM in exchange for a disproportionate small number of supposedly targeted eyeballs.
  • Actually pay for targeted traffic.

(the classic carny methods of attracting interest, stringing along, grabbing the money).

Business users, who are in the ad business themselves (aka journalists), are handed the short end of the stick in this game. Paying for targeted traffic is obviously not an option (unless traffic resells for at least 50 times its cost - based on a 2% click through rate). The alternative is equally bad: junk out your own inventory/labour in order to build a CSM presence and gamble on going viral once in a while without realising that the return value (much like the giant low quality teddy, made in China) largely consists of disengaged users, who are not merely ad blind, but simply close the browser tab right away because they got tricked into looking at content they were not really interested in to begin with.

CSM does not outperform Mastodon. Much like a carny, it merely uses giant teddy bear numbers to dazzle everyone into believing that they drive attention when in reality, it’s just people locked into an engagement game.

As a journalist, one can make all kinds of arguments to be on CSM, but in the long run it just slowly spirals down to selling your house and renting back an apartment.