You scroll down the feed and click into a few channels, trying to build mental connections between what you're seeing.
You endeavor to make your first imposition on the network: your very first post.
You nervously wait.
You finally get a notification: networkp has commented on your post.
Congratulations—you've left the starter area.
It's been about a month since I joined River. In that time, I've written about what River is here and here. But I'm tired of being a beginner River user.
I want to be an elite River power user™.
After slamming my head trying to think of ways to elevate my River game, I came across this interview. After some scavenging, I found what I needed to reconstruct and visualize the greater River network. I went ahead and did that:
Visualizing user interactions as networks
That's all of us, in relations (not in the biblical sense). Without a direct following mechanic between users (nodes), there are a few ways to visualize user-user interaction (edges):
Comments - both how many comments user writes and how many they receive on a post in a channel they own
Follows - how many channels a user follows and how many users follow a channel another user owns
Shared interests - users which follow >3 of the same channels
Basically, ingoing and outgoing interactions, ignoring connects. I reconstructed these interaction networks and took a peek.
Clustering users by interaction patterns
I am not going to directly give numbers on how large River is. On the user side, River is not a metrics-based platform so it feels invasive and rude to turn the lens around. Just know that if I stop posting, the brown-nosing didn't work.
That said, I'll comment briefly and vaguely on size.
Based on the number of active users, there has to be a large population of lurkers. If that's you and you are interested in imposing your vision on River, it's best to join now while there is still ample opportunity to make your voice heard.
River's current size creates this window. Clearly, River is well past a hypothetical Dunbar's number. It is impossible to know every single user but simple enough to recognize superusers. Yet, River hasn't quite hit the critical mass necessary for recognizable subcommunities organized around specific topics to emerge. The network is sparse at the channel-level where topic focused interaction like that would occur.
There are groupings of users who interact with each other but they haven't necessarily calcified. Many of the representatives in each group have significant cross-group interactions. I used the Leiden algorithm to visualize groupings based on commenting patterns:
2 representatives from each grouping:
Group 0: networkp, missmayhem
Group 1: spideynate, fallwinter2002girl
Group 4: kikomiro, <a deleted account>
Right now, I'm not sure what's driving these groupings—whether these users joined and interacted on the platform around the same time, if they are bonding over shared interests, if this represents time zones, or if these are real friend groups. Whatever the case, cross-community nodes, users who act as bridges between these groups, will have more significance as groups delineate over time.
We are all three hops away from @networkp
If you've posted, you know River protagonist, @networkp. He is, by several engagement metrics, the most connected node in the River network. When visualizing how users relate through comments on posts, 66% of the user base is at most three connections away from networkp. The rest of the users, the 33%, are mostly inaccessible and have never posted.
You also probably know @hardliver (and maybe learned of River through her socials like I did). If networkp is driving the most engagement, she is receiving the most.
No big suprises there. Both are affiliated with the site. Who is the closest person in terms of community impact who is not directly part of the River team?
Based on comments and follows, it is one of these people: @missmayhem, @800cherries, or @fallwinter2002girl (I actually don't know if they are completely unaffiliated with River). I didn't adjust these metrics for time since they joined so the temporal dynamics of the cumulative engagement I'm looking at is unaccounted for. Whoever joined first had a head start.
How do people use the platform in general? To investigate general patterns in usage, I performed unsupervised clustering on the portion of the user base showing activity with variables for # of channels they follow, # of users that follow a channel they own, # of comments they made, and # of comments they received:
Yes, I know the heatmap isn't very informative given how hardliver and networkp are such huge outliers (literally the two clusters at the bottom). But the underlying data suggest a strong culture of reciprocity on the platform. There is a two-way pull in the network where engaging with a user generally results in them engaging more also.
There are also users who attract more engagement than they themselves give, generally found in cluster 4 where activity is lower across the board. For less-connected users, this lopsided engagement ratio may signal content quality—since they lack the "engagement floor" that highly-connected users enjoy from network position alone, their engagement spikes are more likely driven by the content itself. This metric is likely even stronger at the per-post level, identifying standout individual contributions rather than consistently high-quality users.
However, high receiving-to-giving ratios could also indicate:
- Controversial or provocative content
- An unusual format for a post
- Off-platform reputation driving on-platform engagement
- Strategic timing or channel selection
These users represent an interesting retention challenge. They're creating content others value but not deeply embedded in the network through reciprocal engagement. Rather than directly incentivizing higher output (which might dilute quality), the platform could study what makes their posts successful and create conditions for similar content to thrive—or simply ensure these users feel valued despite lower engagement output.
Refining strategy in light of new data
While still curated by the channel owner, the last highly visible form of top-down transmission on the site, the blog tab, is now populated by user posts. Cultural transmission of "River values" is gleanable from the main feed or certain channels but again there isn't really anything holding people to them. The culture is technically up for grabs.
Where traditional platforms have tried to impose top-down moderation, River's team keeps finding additional ways to proverbially offer the wheel to the community. In some ways, it feels like the community has declined this (passenger princesses). It's not that no one wants to step up and really push to gain local prominence, it's just 1. not that easy and 2. not that clear how to accomplish this.
So.
Back to this question of terraforming River. We revisit this question with a bit more data in hand. I don't think it is terribly hard to figure out what specific actions need to be taken so I'll focus on the nuance.
With new understanding of the network's topology, you might find which levers to target engagement towards. The most connected nodes? Or maybe the periphery? You might think the strategy is then to post at the highest quality possible, as frequently as possible—crafted around those specific levers.
You would be wrong.
Again, the platform is built on cultural standards. People are looking at the feed and trying to read the room, checking the few existing engagement metrics to figure out what will be well-received by the community. Completely normal.
Producing polished and resource-intensive posts is not the current cultural standard. Ultimately, creating a large enough imbalance in content quality is subversive and can briefly detract from overall participation in content creation. This lack of resilience to imbalances is a size-dependent effect. At River's current size, you need a scene, not an exhibition. People need to hang to create a scene. If people don't think they can hang, they won't stick around. It's not that you can't pull up and flex once and a while. You just have to also sprinkle some shitposts in too.
Content diversity and quality, yes, but if you want hands on the wheel, you have to write (or make long-form videos). Personal blog-style visual posts with short form writing balances attention-holding, branding, and influence. Exceptional visual posts largely influence other visual posts. Written posts can influence both written and visual posts (embedded images) and have the capacity to influence user behavior in more directed ways. Like @checkthreetimes points out, "writers still run the internet". No two ways about it.
Comments are a pro-social, low-effort way to build recognition. It works. Yet, mass commenting has its limits. You cannot directly networkp and expect networkp results. River is ostensibly networkp's job. Just as you cannot show up at a business and start running the place, this too is, well, not entirely off the table—it just needs to be done in a very particular way.
Bottom line—how difficult is it to overturn existing network culture?
Through engagement, moderately difficult.
Through a sudden influx of coordinated users, easy.
Prognosticating
Useful social networks are built on weak ties and strong interactions. You want breadth of connection but depth of engagement. Elimination of weak interactions in favor of strong ones creates natural friction that preserves culture—in theory.
River is also a physical space after all, grounding the digital in reality. These are advantages. But they don't make River immune to the forces that destroy online communities.
What would be the consequences of unmanaged growth? With rapid growth, forums and communities very frequently splinter, devolve, or drop off without some form of moderation, the cruel fate of most community-based online spaces. A few weeks ago, I would say this was River's eventual destiny too.
In this aspect, obscurity is both a gift and a curse. One of the main draws of River is its current size. The act of joining was colored by the predatory behaviors observed on other platforms. As a dark forest, River "cap[s] the downsides of looking bad and the upsides of our best jokes by virtue of a contained audience." Preserving insularity as a core feature necessitates careful planning on how growth is approached.
This feeds into something I've mentioned before—how size of a platform impacts how a platform feels. We intuitively know this. Twitter feels different from a subreddit. The point is that River's current design is highly subject to drastic changes in user experience with different user numbers, contributing to the evolution of the platform as new kinds of people arrive.
Let's consider 4chan. Like River, 4chan has virtually no content moderation—only blatantly illegal content gets removed, and even that inconsistently. Yet like a roach, it has survived over two decades. How? 4chan benefits enormously from its reputation as an internet gutter. The platform's toxicity is the moat, filtering for a specific type of user while repelling everyone else. This creates remarkable cultural cohesion within its boards, even at massive scale. No algorithm to game, no follower counts, and no moderation. Just brutal self-selection.
River faces the inverse problem. Its branding is intending to attract thoughtful, creative people. Unlike 4chan, River can't rely on toxicity as a filter to manage growth. The platform needs some mechanism to preserve its character without alienating the collaborative ethos that makes it work.
Besides moderation, the other commonly applied mechanism is curation. Moderation says "you can't do that." Curation says "look at this instead." Intrinsic to both is that they can breed resentment; however, public sentiment about moderation is way, way worse. People who feel ignored or not catered to push back, especially in situations where moderation is added post facto. Communities that grew up on certain norms don't take kindly to new rules. Though River remains moderation free, recent updates experiment with varying the levels of existing top-down curation (like the blog tab) but the overall balance required is likely delicate and not at all clear.
Another platform that uses light curation in the absence of moderation is Are.na. The Are.na comparisons are unavoidable as are allegations of River being an Are.na clone. I personally wasn't in the know about Are.na until after I joined River so I tried it out (as a super, elite River power user would). While the platforms overlap almost completely in design features, the user experiences are markedly different.
I'm not sure why I was so surprised. Despite both making sneakers, Nike and Adidas ultimately sell different, independently popular products.
A lot of this is driven by minute UI differences that have an outsized impact on behavior. Engagement is buried under several additional user actions and obfuscated from user visibility. Part of this is branding choices. While I'm still trying to put my finger on which features put in the most work, it is undeniable these differences add up to emphasize the respective platforms' philosophies and functions:
Are.na is an organizational system.
River is a self-disclosure system.
This is also why Are.na scales well across different sizes: a well-organized archive remains useful at any size while a large conversation becomes chaos without intimacy.
Will River become Are.na at high user counts? No, it will instead grow closer to resemble the platforms that dominated the 2010s. You won't be able to see every new post—that completionism dies first. Culture will still matter but it'll be culture evolved beyond what exists now. You won't remember most usernames, so posting your face or having a distinctive content style becomes more important. At scale, engagement gets more uneven—the comfort of "nobody gets much attention, so I'm fine" disappears. Winners win bigger. The main interactive hub shifts from the main feed into channels. These predictions need to be interpreted in the context of how platforms fit into the current digital culture and how River exists precisely as a response to mainstream platforms.
The question isn't whether River will change—it will. The question is whether it can preserve its value proposition using its many curation mechanisms throughout that transformation. The branding shouldn't need to change but the curation mechanisms might need to. In the absence of new features which would alter the current dynamic, focus should be turned to cultivating existing users or attracting small-scale influencers, those who are committed to growing the scene. We are seeing the early experiments with this: River-exclusive content. Cross-promotion across other platforms. Attempts to be the first home-grown River influencer. I do think River's viability as a way to attract attention on the greater internet, not just share in small circles, needs to be proven for substantive growth.
I only have a vague sense of what a favorable outcome for the River team would be. I don't know if they are fixated on growth. But with no growth, current network topology is more fragile than you might think. Cultivating multiple highly-connected users makes the network robust—robust to variance in individual participation. This isn't just competition for centrality; it is distributed infrastructure. networkp's liberation comes from the network no longer needing him to hold it together.
*i want to preemptively thank the river team for not telling me i couldn't do this