Twitter: what was, what is & what might be

The aim of this paper is to consider three factors that contribute to the existence of the Twitter social media platform. First, reference will be made to an overarching framing in terms of the methodology adopted to develop the product. Second, the user base will be addressed and third, I will hone in on specific features of the platform informed by, in the first instance, two specific document sources namely Twitter’s New Order & Twitter is a Truth Machine to examine detail and consider what was, what is and what might be. In addition to those three factors (Fig. 1), the purpose is to unpick such detail so that a demonstration of how the affordances and constraints of Twitter might be foregrounded when consideration is made for creative enhancement, perhaps.

Overarching Methodology

It was during March 8th 2006, Twttr registered the domain name twttr.com and was released March 21, 2006 as an internal service for a podcasting company’s employees, which in turn became a public service in July, 2006. The Twitter version was launched, with the domain name twitter.com, in September 2006 and the developers/founders (Jack Dorsey, Noah Glass, Biz Stone and Evan Williams) pledged that the vision for the platform was:

A global community of friends and strangers answering one simple question: What are you doing?

An agile methodology is adopted by Twitter spanning team building, new hires/onboarding and productivity, greater detail can be found in this talk delivered by Nick Muldoon. Essentially, a combination of top-down and bottom-up policy is in place, which enables teams to generate ideas that can be implemented and aligned with the notion of iteration as opposed to a tighter framing as that of scrum, whereby top-down competency schemes ensure accountability is a shared language. A variety of structured team- groups are constructed whether specialist, generalists or a mixture of both, and often the aim is to try to use the same programming language and associated libraries, where possible, for continuity to support progression with the purpose to maintain and enhance the product for the overall good of the cause, as opposed to the hero culture that can exist among engineering teams. New hires are onboarded via a structured programme and their continuous development is accommodated by the range of courses offered by Twitter university. Regarding product outputs, scrum involves two week cycles and there is flexibility surrounding the sprint in terms of incorporating iteration too.

For more information on Agile and other methodologies such as Waterfall see my piece here about development of Ed tech software; in which the findings point to a need for these methodologies to consider user-voice more often across the development process, in order to draw upon the users’ knowledge and experiences. In sum, it appears there may be a correlation/relationship between a well-executed methodology and level of community control of the program, ‘the systemic change focus of the program, and the degree to which the program is multidimensional’ such as we find in ecological programming (Jakes, 2008). So, it goes, in the establishment of a transparent nexus for user voice and level of community control, this ought to impact on a business model that is open to evolution too. The inference is akin to the same top-down/bottom-up approach used at Twitter with teams; but with the community involved, because this type of agility will enable an ever-evolving business plan as opposed to a static plan that inevitably impacts on whether a large ‘multidimensional’ user base is retained. Incidentally, while writing this paper I came across a blog that appeared to visualise the sort of movement I explain here.

User base

It has been suggested Twitter’s most existential and enduring problem as a business: is its inability to retain a large proportion of the new users who sign up. While the longstanding goal of businesses on Twitter is to drive people off the platform and onto their website. Yet, over time Twitter’s user base has evolved to include users who engage with the public platform for news, shared opinions and humour as well as a variety of other factors that attract engagement. This means there is a growing population of users who are engaging for longer periods of time. A good example is in the number of established global education-related, professional development chats that take place weekly :- @LTHEchat #lthechat #edchat #edtechchat #EDchatDE as well as regular outreach initiatives :- @MHChat #MHChat and there is an ever emerging range of domain-groups e.g. #LeadersHour @_Future_Leaders and so on, arguably a new wave of engagement. This adds to the multidimensional user base that is retained.

So, where is the complexity in terms of focus and impact on the aforementioned? It can be proposed, and this is quite common across all software/engineering/development companies; attention and focus to detail for enhancement, zooming in on the detail at the expense of a stagnant business model that stands alone, runs risk of oversight especially to the tune of maintaining an ecological programme/process. Let’s take an example of that, in other words, let’s funnel down.

Features & Functionality: Affordances & Constraints

I’m not going to reference again, all the sources I have used to inform this paper have been mentioned once, so more detail about the following can be found in those links.

It is noticeable Facebook’s news feed algorithm automatically orders every post according to a highly-sophisticated formula that is personalized to each user’s habits, tastes and relationships. Similarly, Twitter has an algorithm that affects only the tweets at the very top of your feed. The rationale behind the introduction of the algorithm was to avoid the scroll of death and to improve relevance:

As soon as you open it, Twitter quickly collects and assesses every recent tweet from every person you follow and assigns each one a relevance score. This score is based on a wide array of factors, ranging from the number of favourites and retweets it received to how often you’ve engaged with its author lately.

Furthermore, tweets are weighted based on individual user engagement and ranked, this means the affordance for the user is that you see occasional tweets from people you don’t follow, because the ranking system shows that you’re likely to want to see them and the “In case you missed it” is part of the algorithm too.

It is worth linking to one of Sir, Tim Berners Lee’s suggestions, in his statement this month, which simply said, ‘We need more algorithmic transparency’, to be fair this concurs with Twitter’s new movement into the realm of algorithm usage/design, which aims to be transparent via Twitter development documentation.

The potential issue with any algorithmic design is the way in which the constraints operate. A good example is the said Twitter algorithm that specifically enables the user to see more of some type of tweets and less of others. I am specifically talking about the boundaries that are inherent in the If, else conditional statements used in coding, because boiled down this means this or that as an outcome when the algorithm runs. While this bounded choice of one or the other can be useful in programming, there can also be several choices built into the algorithm, and if statements are particularly powerful there can be any number of branches e.g. conditional operators in JavaScript:

a == b a is equal to b.

a < b a is less than b.

a > b a is greater than b.

a <= b a is less than or equal to b.

a >= b a is greater than or equal to b.

a != b a is not equal to b.

The complexity, of course, is in the desired metrics that are assigned to each function and the ‘new’- being machine learning/AI algorithms. The Twitter algorithm, which organises the timeline relates to measures of active users and every possible engagement and attention metric. Thus, as you can probably guess with more choices built into the algorithm this means an increase in volume of metrics, and it is fair to say when is enough, enough? Indeed, on that note, let us return to the vision of the product back in 2006, ‘A global community of friends and strangers answering one simple question: What are you doing?’

It seems timely after 11 years in business, to stop and reflect on direction for sure. A provocative summary I wish to place here, is that the very nature of sophisticated algorithms; the constraints, tied to the features and functionality of software/platform design has the potential to reduce the affordances for user engagement and this is manifest in the current example whereby the Twitter timeline runs risk of a total lack of authenticity and loss of up to the minute, real-time functionality. Returning to the notion of ecological programming, where might the focus lie in the future?

User Experience/Community

If fancy, state of the art algorithms is to be a growing presence in product design for Twitter; then parallel to that agenda needs to exist a focus on the user experience. A user develops their capabilities over time, and in the main, intuitively if the product/platform enables prompts to do so. Yet, not all users develop in this way. It seems to grow user base, and to retain a large reach; development of features in the product, which enhance user capability skills is a good idea. Why? Three reasons:

-growing use of algorithms that produce functionality/features ahead of user capabilities

-to retain a community of users there needs to be a sense of freedom, choice and control to inform a sense of ownership over their own user experience

-providing the right pitch and level of challenge via features/affordances and constraints for users to want to return and engage for longer periods of time.

Ultimately, the skill-set I am referring to is the ability to engage on a level of critical thinking. A movement beyond the zone of code that personalises each user’s habits, tastes and relationships. Given, the development of capability involves knowing and applying basic skills/routines, techniques, underpinning concepts, processes and higher order thinking skills. And, this in turn might very well enable metrics that include quality engagement to help shape discourse that is useful and valuable for all users. Metacognition is about an ability to use a third space for thought for example A+B=C (the in-between-ness, the third space for authentic, creative sense-making). A simple suggestion for metrics might be akin to the Pinterest ‘tried-it’ button, and so on. Moreover, to really keep the authenticity, timeliness of the live feed in Twitter it may be the algorithms that organise the feed could be functional in the Tweet Deck alone. Leaving the live feed unadulterated and the Tweet deck managed by the user, they choose the settings with the fancy algorithms.

To conclude, I suppose what I am suggesting is a shape shift as represented in Fig. 2: