The social interactions on a platform are driven far more by the technical design of the platform than they are by the people who populate them. Over time, the userbase of a platform will be shaped by those technical designs, acting in a cycle which tends to exacerbate whatever behaviors the platform encourages.

One type of technical decision that influences platforms is the design of their access control system for posts; these decisions end up fueling a sometimes surprising amount of the way the platforms are used. In social media, I’ve primarily used three major platforms: LiveJournal, Facebook, and Twitter. Each has had different capabilities over time, but even beyond capabilities, even more minor variations can change how users feel about the product.

LiveJournal Era

Journalling platforms like LiveJournal – most popular in the early 2000s – were among the first to build an explicit “social graph” of bidirectional relationships. Because these platforms had built up social graphs, it was (to the best of my knowledge) one of the early platforms to support the idea of “friends-only” content: content which was only accessible to the people with whom you had explicitly built a bidirectional trust relationship. (Earlier platforms sometimes had access to “password protected” posts, but those passwords were not tied to a social graph.) In addition to the basic ability to limit posts to “friends”, it was straightforward to create posts that were accessible to specific subgroups of your social sphere via “friend groups”. At least in the social spaces I used, it was routine that people would use multiple different levels of access based on trust level among their friend groups. The ability to have strict control over who had access to which content was a fundamental platform feature, and made a huge difference in what types of content were shared, and how they were shared.

Shift to “Status Updates” with Twitter

The evolution of social platforms in the mid-2000s towards being more focused on “status updates” and microblogging were also associated with a move away from these highly specific content controls. While LiveJournal had the ability to choose to make each post public to the internet; to just friends; or to specific subsets of friends, Twitter has never had a similar option. Instead, your entire timeline can be public or “protected” (limited only to people who you follow, or who followed you at the time that you flipped it to protected) – there is no ability to share only some updates with a smaller set of friends. While Facebook had this feature, the security level of the content you were posting was not originally particularly visible – it was only available as a small icon next to the post, rather than being visibly present. (LiveJournal included both a default security setting you could change, and included the security in the same area as the “Now Playing” UI widget that we all loved to fill in with whatever was on Winamp back in the day.)

These differences in how the platforms allow you to share your content make huge differences in the social structures around the platforms. For example, Twitter’s lack of options for post-specific visibility produce side effects across the service.

  • The only way to make part of your content “go away” is to delete it. There is no way to keep a historical record of the things you’ve written before unless you’re willing for them to be public for everyone to see. Combined with a tendency for attackers to trawl through older content looking for things to call out as problematic, it makes using the platform riskier for those who suffer harrassment campaigns.
  • With no ability to separate “public” from “more intimate” content, users of the service fall back to using multiple accounts with different social groups. This might apply not just to “Public vs. Private”, but in some cases even more broadly to “public” vs. “fandom” vs. “furry” accounts, sharing content related to different aspects of your life.
  • For those who do choose to engage in “locked” profiles, there isn’t a straightforward way (other than blocking individuals) to eliminate who has access to your content. If you decide to make your timeline private, your current followers can still see your content. This is important and valuable for many folks, but is distinctly different than how services with a post-specific visibility would typically be expected to behave – especially as those relationships do not have to be bidirectional to begin with. (If I am an abuser, if I follow you, I can still see your content after you protect your tweets.)
  • Switching to protected tweets can significantly impact ongoing conversations. (This is interrelated with Twitter’s “Directly integrating other users content into mine” method of operation for tools like quote tweets.) This is usually for good reason – again, protecting tweets is one of the few tools that users do have for protecting themselves from harassment – but can also be more frustrating when the person previously engaging was acting as a public figure (e.g. a reporter).
  • In general, the lack of “private” posting means that in Twitter’s social graph, bidirectional connections are very much not the norm: there’s no reason for me to follow a person to grant them access to my content, so the follow relationship is only related to the content you want to see, rather than being intended to grant social access.

Facebook’s evolution

Counter to Twitter’s extremely limited access controls, over time, Facebook has grown more obviously visible mechanisms of controlling content access controls. The ability to expose content to specific groups was added in at some point, and later, the content visibility – what access control each post was being osted as – moved from being indicated only by a small icon to be very visible UI widget near the posting button. At this point, Facebook has options somewhat close to what LiveJournal had in the early 2000s – built slowly, in large part to address users privacy concerns with more subtle UIs. Throughout the 2000s, Facebook’s sticky access control setting interacted poorly with low-visibility access control indicators: users would occasionally post a single post public and forget about it. The result was they they might find out months later that they’d been posting public content during the interim. (Much of this got blamed on Facebook intentionally changing defaults, but frankly, I think that it was almost always just a case of bad UI rather than changing defaults.) The distinction here between updating the default to match the most recent post, vs. requiring a specific configuration change to update the default (as LiveJournal had) was also significant.

Later on in the 2000s, Dreamwidth – a LiveJournal fork started after LiveJournal’s sale to a different company – introduced an additional concept, splitting the “reading list” and the “Friends list”. This split allows for more accurate social graph modelling: under LiveJournal’s model, it was impossible to add someone to your “friends page” – the feed of posts that you were reading – without granting them access to your “friends-only entries” of your own. A longstanding complaint from users of LiveJournal, this was added in the Dreamwidth fork, such that you could separate content you just wanted to read from users to whom you wanted to grant access.

Facebook, on the other hand, does not have this option: While it does have “I would like to remain friends but not read your content”, it doesn’t have a good way to implement “I would like to read your content, but not grant you access to my posts”. This lack of functionality likely stems in large part from the overall design of Facebook to move “publishers” away from person-style Facebook accounts to Facebook “pages”. Page owners are not granted any additional read access, so for content which is posted by brands on their pages, users can consume it in their feed (usually by indicating interest via the “Like” button) without granting access. That said, this isn’t universal: small town politicians, tech policy folks, and more sometimes use their Facebook accounts as an alternative publishing platform under their own names, and in those cases, the lack of a distinction between “Reading” and “Friending” can be an issue.

Technical Decisions are Not Neutral

Some of the differences created by these distinctions in how access controls are managed are minor – for example, where to indicate the current visibility of a post. However, these choices can have an outsized effect: on the perception of the platform, and how much it respects privacy; on the behavior of the users and how they use the access controls available; and even on who the users are willing to call “friends”. This is just one way where we can see that the technology is not value neutral: each of these choices can have an effect on who uses the platform, for what purpose.