Special Edition X10. Search & Discover 🔎⚡
Search and Discovery in a Multi-Node Future
We all have heard the talk about a multi-node future. One where BitClout.com is no longer the front-end and we instead interact with many front-ends — each one existing like it’s own little social media platform, all pulling content from the BitClout blockchain.
You could jump from one node that’s image-focused like Instagram…
…to another node that only shows posts in a specific language…
…and then back to your core node that shows everything but tailors your feed to your preferences.
In a special edition from @Dylsell published in these pages, he explored how those nodes might be linked together and how a user might navigate between them. The simple version is that a reddit-like model, where users have a central identity/profile/interface and post to multiple subreddits, could solve a lot of the potential issues that might arise in a multi-node future. At the end of his article, @Dylsell asked a few critically important questions about a future with multiple nodes:
Where’s the front door? What’s the front door? How do prospective users learn about the product or its potential without it? Are we expecting that they find each node’s site on their own?
Those questions align with my belief that it would be extremely risky to try to sunset the current front-end and hand it off to the community before we have a critical mass of users and adoption of new nodes. But even with mass adoption and well-established new nodes, I have to wonder how discoverability will work on this new system.
Even today, while things are still pretty centralized, discoverability on BitClout is very challenging. Finding specific users, content on a specific topic, or different nodes is not streamlined. That’s just going to get worse when all of these things are spread across even more nodes.
One of the other reasons this topic jumped out at me was @Dylsell’s apt comparison to Reddit — a platform that has had a notoriously hard time nailing search and discovery. In the 8 years I’ve used Reddit, the ongoing joke has been that it’s easier to search google for content on Reddit than it is to use reddit’s own search tool. I still see those jokes being made to this day.
I suspect that this is, in large part, because searching across multiple subreddits (or nodes) is more like searching the web and a lot less like searching traditional social media.
Still, there’s a lot that’s right about Reddit’s approach to search. The main part that I think we need to adopt for BitClout is searching across different layers of the platform. When I talk about layers, you can think about it in terms of the different ways you can group content together:
- Individual posts
How you navigate those layers is critically important for new users who are trying to find their community and discover relevant content. Right now new users have to scroll through Global, try to find people they want via username, or use a 3rd party search service. That’s not ideal. Those 3rd party search services are great — CloutaVista has massively expanded our ability to search BitClout in the past few weeks. But, they still require you to leave the platform and then navigate back — that’s a lot of friction. They also only look at search through the lens of posts, not users or nodes. So you can quickly and easily find mentions of a keyword in a post, but you will have a hard time finding a community of people who post about the keywords you care about. For instance, users posting popular clouts about NFTs or nodes that mention keywords like art and photography.
The consequence of relying on our current search options is a lower retention rate. As a new user, it’s hard to see the value of a social network if you can’t easily make social connections.
I think we need to attack the problem early — while we’re still a mostly centralized platform with most users on one core node. The opportunity to set norms and steer the ship is now 🚢. Once things have splintered into multiple nodes, commanding an entire armada of ‘ships’ will be much harder.
The majority of this can be solved with two levers: in-app search and hashtags.
The first and most critical step is having in-app search that goes beyond just looking up a username. This can be a well-integrated 3rd party service, but it needs to be expanded to look at every layer of content: nodes, users and posts.
Those layers are a core concept and can be a little tricky to understand. If you use a search engine like @CloutaVista and you search for a keyword like “NFTs” you get a list of posts that use the keyword “NFTs”. That’s useful if you’re actually looking for posts that mention NFTs. But it’s not as useful if you’re looking for the people that post most frequently about NFTs or for nodes that focus on NFTs.
Instead, I’m proposing that a search for a keyword should essentially run three different searches simultaneously:
- Posts most relevant to the search
- Users most relevant to the search
- NFTs most relevant to the search
- Nodes most relevant to the search
Search shouldn’t just direct you to an individual post. It should point to the people and communities that are most relevant to whatever topic you search for. That’s a big deal for a brand new user trying to navigate and find their own community on BitClout.
Having multi-layered search doesn’t guarantee that search will work flawlessly. Case and point: Reddit search has offered this for a while and it still is much maligned by its users.
To achieve truly relevant results we need content that is labeled. BitClout isn’t a search engine company and I’m assuming it doesn’t have the resources to build a linkback/authority based search system for the blockchain. The alternative? Have users do some of the work themselves. A solid hashtag system creates an easy way to group content by subject matter. This lowers the burden on users; less keyword stuffing and more organizing unique communities around a hashtag that can be filtered across nodes.
That filtering function is important because hashtags could be a very easy back-reference when other users search for nodes. In this system, a node that filters out everything but content using #NFTs would rank higher when somebody runs a search for “NFT”. That’s another solid step towards new users being able to find their community more quickly and efficiently. Which in turn is the best way to illustrate BitClout’s value to new users.
Last but not least, hashtags allow folks to expand their “Following” feed to focus on topics, not just people. That means you can still discover new creators while keeping your feed relevant. If you’re a brand new user starting on Day 0 and you don’t know anyone on the platform, you can still follow a hashtag you care about and fill your feed with the creators who share your interests. That’s another big push towards illustrating the platform’s value to a new user.
Since BitClout isn’t focused on user growth (yet), these issues might seem like they’re a lower priority. But I actually suspect that they matter now more than ever. With @diamondhands listing onboarding flow as a priority post-NFTs this seems like something we need if we’re going to prove that BitClout can be a platform that serves all creators.
@JoshuaCottrell and @BitsTODAY