Category: Findings

Our respondents make extensive use of third-party tools to engage with the fediverse. In other words, they post not only via a Mastodon or WordPress instance or their respective clients, but with anything ranging from auto-posting bots to content scheduling platforms. This post covers some of those tools and what that might mean going forward.

Generally we found that once publishers set up a fediverse presence, whether in the form of an account or a dedicated instance, their further engagement is quite light-handed. Most indicate they spend little time engaging with social media and use it primarily as a distribution platform to redirect (potential) readers to their own sites, rather than as a way to elicit commentary or feedback. Furthermore, all publishers bar one we spoke to use the fediverse as part of a broader array of social media. This, then, translates into their extensive use of automated or semi-automated tools.

Content scheduling platforms

Our respondents specifically mentioned no/low-code automation platforms such as IFTT¹ but also social media content scheduling and cross-posting platforms such as Buffer and Publr. These are used to schedule and cross-post notices of new articles simultaneously to different social media platforms. Using such tools, publishers get aggregated and platform-specific insights into how many reactions, shares, and favorites posts get. These platforms sit between the publisher and the fediverse app.

Some respondents indicated also making use of link trackers such as Dub and Kutt, either as a service or as part of the publisher’s own infrastructure. These link trackers sit between the visitor and the final website. These are used to track who clicks on links and on what social platform those links are clicked on. Although many fediverse clients do not show full URLs, such link trackers can be recognized by the addition of UTM parameters to the URL, such as:

?utm_source=mastodon&utm_medium=social&utm_campaign=mastodon

Such link-trackers are often used in combination with post-scheduling tools to gauge the effectiveness of disseminating articles across different social media.

Bespoke automations

Some publishers indicated making use of bespoke (semi-automated) means of promoting articles. For instance, we learned of an example of a publisher using custom scripts to mirror different RSS feeds of the publication. We also learned of the use of a custom script used to (re)post the most popular article of the week.

A fediverse-specific practice we encountered is what one might call “boost-cycling”. That entails first publishing a post and then boosting it with the same account after a few hours. Then in later hours or days, the post is unboosted and boosted again to reach those in different time zones. Authors with multiple ActivityPub accounts, for instance a personal account and a blog, would also boost-cycle with their different accounts to increase reach and make sure the article “fans out” well enough. This has to do with federation dynamics and reverse chronological timelines.

Take-aways for developers and designers

1 – Publishers don’t post the way you think they do

When it comes to long-form publishing, publishers are not using the clients you design. Instead, they use specialist tools that connect directly to fediverse instances over the API. 

In one case we spoke to a web editor who was not sure which fediverse software or instance their publication was on. This is of course not a shortcoming of that editor but rather an indication that these professionals work with different tooling ecosystems. They are not “on the fediverse” but rather “on the content scheduling platform”. Yet, these professionals are the way through which long-form publications and journalistic outlets make their way to the fediverse.

To cater to long-form publishers, developers would probably do well to focus on integrations with third party tooling over for instance adding custom long-form editing and writing environments in their software. However, that would also need additional research that looks in more detail in to those workflows. How are those tools used exactly? What affordances do these tools have that are lacking support in fediverse apps?

2 – A long-form publishing context is different from personal social media

Publishing is a process involving several steps and different people. Authoring and editing articles happens in writing environments like Google Docs. Material is then either copied into the CMS or handed over to the web editors. Disseminating articles over social media is done (semi)automatically or is handled by someone else entirely from the social media team. These are crucial differences between individual or personal use of social media and their use in long-from publishing environments. It is therefore wrong to assume that the author and the poster are one and the same. It is also wrong to assume materials are authored in the environment where they are posted. Finally as scheduling platforms show, the assumption that the article is “posted” in the environment where it eventually ends up also should be revised for long-form content.

While these findings hint at where to look next, turning these insights in to design strategies or product roadmaps would require additional research.

1. A particular issue mentioned with IFTT is that while it supports Mastodon, it in fact only supports the instance mastodon.social. ↩︎

Based on our conversations with media professionals we identify a key issue: publishers likely won’t want to federate full articles, instead they want to redirect traffic back to their sites. This is the case both for publishers that make use of paywalls and those that don’t. It is also the case for publishers whose content is freely distributed under creative commons licenses and those that retain full copyright.

An important context here is that all publishers we spoke to found it important to have a presence on non-commercial social media. It spoke to and reinforced their values. It offered them independence. They saw it as a way to have access to their subscribers in a way which is not contingent on third parties. They did not need to be convinced why something like the fediverse was a good idea. Instead, when it comes to the fediverse as a distribution channel, they have particular needs.

Why do publishers want readers on their own site?

Formatting and layout
Publishers spend a great deal of effort to make sure their articles look good and read well. They make frequent use of illustrations, pull quotes, sub-headings and other formatting to break up the flow of text.

Those publishers whose online outlet mirrors material also published in print expressed a desire to maintain the same quality of design. Therefore, they invested heavily in custom(ized) CMS’ that can for instance flow text around images, or display multiple images as carousels or slideshows. In some cases such CMS support interactive elements such as information visualizations where readers can explore the data. Furthermore, published articles appear as part of series or with links to other articles in the archive and these can be indicated visually.

Publishers also feature image-only publications such as comics or photo reports. Especially where those include series of images, editorial control over placement is also important. Finally when it comes to photo reports, captions are an important element.

Funding
All publishers require income to operate. Although we spoke to publishers with very distinct business models, all saw their own website, and directing traffic to it, as a fundamental reason to use social media. The most obvious cases are those where the business model revolves around subscriptions, paywalls and advertising. However, publishers who offer their materials for free and who benefit from wide circulation of their material also expressed doubts. For instance, many rely on calls to become a member or to donate in the context around a piece of content. Only federating that content means it gets detached from such a fundraising context.

Control
Redirecting readers to their own platform ultimately is a way to redirect readers to a place the publisher can influence. That has different aspects. In terms of content moderation publishers wanted to for instance control comments on posts. Either to disable them altogether or to remove harassment. They can do so on their own sites, but will they be able to on federated long-form articles? Corollary to that is that some publishers mentioned the desire to know or influence the context in which the full article is displayed.

Another aspect is that it is on the publisher’s own sites that they can get fine-grained insights in to which articles are read, where they are circulating and where readers come from. Such insights are based on analytics suites on the site, or in front of the site. Metrics, such as impressions, are available on commercial social media but not on the fediverse.

What about Starlight Princess?

Starlight Princess adalah slot video bertema fantasi anime yang dikembangkan oleh Pragmatic Play, provider ternama di industri iGaming. Sejak diluncurkan, game ini cepat menjadi favorit para pemain berkat tampilan visualnya yang memukau, gameplay yang menghibur, dan potensi kemenangan besar hingga 15.000x taruhan. Games ini dapat anda temukan di Situs Slot Resmi Taurus77 yang telah dipercaya dan bersertifikat resmi PAGCOR.¹

The above quote is pulled from a WriteAs instance. It is either SEO spam, advertisement for online casinos, or both. Long-form spam can be trivially found on many WriteAs and Plume instances, as on WordPress sites. Such content also gives you an idea who would actually like to federate full articles: Starlight Princess would.

How to deal simultaneously with both Starlight Princess and the fact that most publishers want to refer traffic back to their own site? This is a fundamental question for designers and developers of long-form support in fediverse apps is a key determinant of the utility of such a feature. The current design of many fediverse servers means that remote content is cached, viewable and searchable on different instances. It is a spammer’s paradise. At the same time, without thoughtful design of long-form support, the fediverse risks satisfying none of publishers’ needs to funnel traffic back to their own sites: formatting, support of revenue models and control over content and context.

How could long-form federation meet publishers’ needs?

However, at the same time the people we spoke to did indicate there would be interest in federating full articles under some conditions. We list some of the different things we’ve heard in this direction:

Those outlets who already adapted their content to web publishing² saw possibility for federating full articles. That means that it is likely more suited for (certain forms) blogging and newsletters, as it addresses concerns around formatting and layout.

Distributing articles to paying subscribers would be an option several respondents expressed interest in³. Ultimately the main goal of the metrics that publishers employed is to know how well their reach is. Metrics are also one of the big reasons publishers want readers on their own site. These metrics are used as educated guesses how clicks can convert in to revenue. Tying federation to revenue thus eases those funding concerns.

Aggregate metrics could make it easier to know and prove to colleagues and decision-makers that investing in this ecosystem is the right decision.

Reply controls and moderation capabilities on federated long-form content respond to publishers need for influencing or controlling the context their materials are published in or circulated in.

1. “Starlight Princess is an anime fantasy-themed video slot developed by Pragmatic Play, a leading provider in the iGaming industry. Since its launch, the game has quickly become a player favorite thanks to its stunning visuals, entertaining gameplay, and massive win potential of up to 15,000x the stake. You can find this game on the trusted and PAGCOR-certified official Taurus77 slot site”. Machine translated from Indonesian. Original at @taurus77 @retrogarde.de ↩︎
2. As one interviewee described it: “You know, a cover photo and 600 or 700 words […] the standard written article on the web as it has been forever”. ↩︎
3. Here the paid-subscriber RSS feed by 404 media is relevant prior art. It would however entail that people will not end up following publication@outlet.tld but rather subscriberUUID@outlet.tld and currently none of the systems work in this way. ↩︎