Twitter’s conversation model isn’t a model for growth


Twitter’s conversation model isn’t a model for growth

Skip ahead to: Down the rabbit hole

Twitter, often bookmarked as a tech company in the league of other successful silicon valley tech companies, has left much to be desired since its IPO in November 2013. The last decade has been one of the largest growth phases for tech stocks, with Netflix’s total 10-year trailing return (dividends reinvested) stacked around 3620% and Amazon returning 1230% at the end of 2019. For comparison, the S&P 500 has returned around 249%. Twitter, however, has largely trended lower and sideways, struggling to break out past its highs in earlier years and currently hovers around a market capitalisation of $30 billion. 

Network effect and the future value of a platform’s user base is a major metric in assessing the value of tech and social media platforms like the ones mentioned above, and it also happens to be the factor distinguishing the likes of FAANG from a platform like Twitter. If, hypothetically, Facebook and Twitter had the same number of users and the same revenue at this point in time – would we land at a similar valuation for both? The answer is likely no, and the explanation for that lies in the growth trajectory of Twitter. 

The future potential of users on these two platforms is tied to the platforms’ ability to engender platform loyalty, most importantly to create repeat patterns of participation. A majority of Twitter users appear untethered to the platform, this is largely due to the nature of participation. This is something we can unpack further to better understand Twitter’s structural weakness currently inhibiting its growth. 

The lowest common denominator conversation proposition 

In April, the New York Times asked Twitter’s founder, Jack Dorsey, what he felt were the prominent mistakes made by Twitter as it evolved into a global social platform. 

“The disciplines that we were lacking in the company in the early days, that I wish we would have understood and hired for,” he said, were “a game theorist to just really understand the ramifications of tiny decisions that we make, such as what happens with retweet versus retweet with comment and what happens when you put a count next to a like button?”

Without this expertise, he said he thought that the company had built incentives into the app that encouraged users and media outlets to write tweets and headlines that appealed to sensationalism instead of accuracy. At the time, he noted, he struggled to envision the app’s potential social implications — and what those design decisions might mean for “how people interrelate with one another, how people converse with one another.”

Jack Dorsey’s acknowledgement of a preference for headlines and sensationalism as opposed to accurate, authentic conversation on the platform is what we really want to hone in on. Even more so than concerns around accuracy, there is an alarming tendency for Twitter conversations to lean into the most available appeal – the lowest common denominator on the platform. While the game theory behind features plays a sure role in this, we can trace Twitter’s current state back to its origins – focusing on two factors that have created a stale form of participation on the platform. 

Factor 1: Private to public spectrum 

Twitter in its original form as developed by Jack Dorsey, was designed as a platform for users to interact with friends, sharing life updates and so on. Given that other platforms catered more directly to “updates” with friends, the public status feature became the forefront of the platform rather than meaningful conversation between known parties. The same users who were prolifically using Facebook were also the initial users of Twitter, many of them would have envisioned Twitter as offering something different to Facebook – namely a public forum. In fact, the Twitter platform design itself seemed to indicate this, as the private circles on Twitter were not easily identifiable – the following list is not your friend list.

Factor 2: Persistence of the conversation 

Where Twitter was viewed as a public forum, the status updates on Twitter then become more a representation of the user’s public persona. The broader nature of the public persona requires status updates that engage in signalling that would appeal to the broadest audience, with the aim of amplifying the content as far as possible. If the aim is amplification for “success” on the platform, the content has to be simplified to appeal to the lowest common denominator. 

With such content, the user’s intent is not to make consumer decisions as you might on Instagram or social connections as you might on Facebook – but to receive affirmation and validation. Twitter easily tended toward headlines, not conversations. When users engage with headlines and they themselves are creating headlines, the conversation is less focal and less persistent – instead, a barrage of transient headlines takes its place. Content becomes momentary as millions of people attempt to continue to amplify over-simplified thoughts for platform success (i.e. validation in the form of likes, retweets, followers etc.) 

Lowest common denominator conversation model

These two factors give rise to a “Lowest common denominator” focused conversation model, one motivated by amplification to the broadest audience possible. Then the forms of content start to converge on the following types:

  • Relatable – stereotypes/comedic/memes
  • Informational – weighted heavily with opinions 
  • Informational – sensationalism (Tesla stock, Coronavirus case numbers etc)
  • Controversial – misinformation, cancel culture etc

It’s in the extreme of these forms that there is an ability to achieve virality, or what users on the platform define as successful signalling. Even when the content seems to centre around thought-provoking topics, the purpose tends to focus on creating impact, where the forms above are prioritised. 

The issue around this form of “successful signalling” is further exacerbated by influence, that someone with influence can compound on the ability to signal. The Twitter blue checkmarks, which were meant to verify prolific personalities in the real world, became a symbol of influence and status and quickly became confused with a stamp of veracity. 

In the absence of a huge followership, a blue checkmark or time to create a personality, a large portion of users became observers of headlines rather than the creators. This has created a polarised user base where a handful of creators get the majority of the positive feedback loop. What this means for the majority of users is that they become a more transient user base, participation is merely the observation of momentary trends. 

Twitter then as compared to Facebook, has a hard time sustaining a tether to its users as it has become a public forum. Facebook has closed circles, the private element, which in itself provides a positive feedback loop for participation on any scale.

Then how can Twitter begin to rectify the transience of its users and demonstrate the future potential of its user base? Despite the slew of features that have been released by Twitter to try and promote “healthier conversations” and greater participation, these features have not shifted the dial meaningfully. The issue may be that the features introduced such as Fleets (like Instagram stories) and audio tweets are superficial in nature and do not directly resolve the ‘lowest common denominator conversation’ model. This does not bode well for how the market might perceive their growth trajectory over the next few years. 

Down the Rabbit Hole

1. How headlines change the way we think

The psychology behind headlines has been helping publishers and journalists hook readers and propagate perceptions since the start of news publication. Much like first impressions and book covers, headlines are a publication’s primary sales pitch. However, the news is also responsible for fact-based reporting (regardless of how elastic in nature the ‘facts’ may be) and headlines need to find a balance between the sales pitch and an appropriate representation of the article they frontline. The disintegration of this balance can be detrimental to not just journalistic integrity, but the fabric of how we interact with current affairs and how they, in turn, shape our own ideologies.

“By drawing attention to certain details or facts, a headline can affect what existing knowledge is activated in your head. By its choice of phrasing, a headline can influence your mindset as you read so that you later recall details that coincide with what you were expecting.”

The New Yorker explores the study of Australian psychologist Ullrich Ecker as he tests “how slight—and slightly misleading—shifts in headlines can affect reading.”

“The headline, it turns out, had done more than simply reframe the article. In the case of the factual articles, a misleading headline hurt a reader’s ability to recall the article’s details… Inferences, however, remained sound: the misdirection was blatant enough that readers were aware of it and proceeded to correct their impressions accordingly…

In the case of opinion articles, however, a misleading headline, like the one suggesting that genetically modified foods are dangerous, impaired a reader’s ability to make accurate inferences. For instance, when asked to predict the future public-health costs of genetically modified foods, people who had read the misleading headline predicted a far greater cost than the evidence had warranted.”

Ecker, having noted the unequal distribution of detrimental effects associated with slightly misleading headlines, across factual and opinion based articles, replicated the study with tweaks to assess the “discrepancies were between the headline and the image”.

“Here, too, Ecker found that initial impressions both mattered and were not easily corrected. When the photo matched the headline, the criminal received more negative ratings, and the victim more positive ones. If, however, the headline diverged from the photo, the victim was rated more negatively when the headline had been about the criminal… Initial expectations of who would be pictured affected subsequent ratings—even though, theoretically, the misperception had been corrected twice, in the text itself and in the caption.”

Ecker’s findings are justifiable causes for alarm on two fronts: 

“First, misinformation appears to cause more damage when it’s subtle than when it’s blatant.” 

A cause for concern, because misinformation here is more likely from editorial carelessness than malicious intent. And second: 

“What Ecker’s work shows, though, is that with the right—or, rather, wrong—headline, reading the article may not be enough. Even well-intentioned readers who do go on to read the entire piece may still be reacting in part to that initial formulation.”

The New Yorker concludes, “it’s not easy to both be interesting and accurate”, but the results of the study show how damaging headlines can be when ‘interesting’ is prioritised over ‘accuracy’. In an era where the balance between the sales pitch and journalistic integrity appears to skew towards the sales pitch, there appears to be no straightforward reconciliation to this dilemma in the near future. 

Source: How Headlines Change the Way We Think – The New Yorker

2. Clubhouse: Status-signalling in a crowded room

“When a platform is able to intrinsically tie signal distribution to content creation it creates a sticky cycle of creation and consumption that feeds back into traction and network effect for the platform. This is the philosophy today’s social media giants have built their billions-wide user bases on. Clubhouse with a primarily audio approach to social interaction poses challenges to the status quo model and by virtue of this will face some challenges of its own…” 

Authentic conversations versus status signalling: a delicate balance

Given the nature of challenges Clubhouse will face, there are – broadly speaking – two directions the platform could opt for. In the first route, Clubhouse could rely on users importing already created social capital from other platforms, focusing on amplifying this signal during the limited period of live participation…

The second route would focus on the creation of social capital and its retention. Clubhouse could solve for the transient nature of its content and social capital by creating mechanisms that would allow users to signal their popularity with other users – essentially indicating their value add as a chat participant. This could be something akin to a point system within a live environment, think live upvotes or likes…”

Source: Clubhouse: Status-signalling in a crowded room – 4th Quadrant


We welcome thoughtful discourse on all our content. If you would like to further explore or discuss any of the ideas covered in this article please contact our editors directly.
Contact Details