What five months of social data taught us about growing a Web3 brand

We started tracking Holdex's X and LinkedIn performance systematically in December 2025. Five months later, the numbers told a story we did not expect.

Impressions on X grew 570% over three months. Profile clicks grew 636%. Followers grew 7.3%.

The gap between reach and growth is not a fluke. It is a signal — and once we understood what it was pointing at, the entire approach to social changed.

Broadcasting is not growth

Our first assumption was straightforward: post more, grow more. The analytics broke that assumption quickly.

The biggest single-week impression spikes did not come from posts. They came from comments. When we spent time replying to conversations on X — 10 to 15 substantive replies per session, across accounts with 10K to 100K+ followers — impressions jumped 114% in a 14-day window while the number of new posts actually dropped.

The mechanism makes sense in retrospect. When your account engages in a real conversation, your name appears in threads that already have an audience. A reply on a post with 500 likes is distribution. A post that nobody sees is not.

We were treating social media as a broadcast channel. It is actually a conversation channel that occasionally rewards broadcast.

LinkedIn is a different audience, not a different format

The second assumption we broke was that X and LinkedIn just need different lengths.

LinkedIn analytics made the problem visible. The posts that consistently outperformed everything else were general job opportunity posts — one reached 1.2K views and 22 comments. But job seekers are not our audience. DeFi founders, institutional investors, and Web3 operators are.

The content was attracting the wrong people because it was optimised for the wrong platform. LinkedIn's algorithm rewards discussion prompts and structured formats — PDFs, carousels, arguments that invite a response. Copy-pasting an X post onto LinkedIn is not a strategy. It produces shorter content that the platform did not ask for and the audience does not engage with. Writing a LinkedIn-native piece on the same topic — different format, different angle, built for discussion — is a different thing entirely.

We stopped recycling. LinkedIn now runs on its own editorial stream: three to five posts per week, Tuesday through Thursday, formats built for the platform, with active replies to every comment within 24 hours.

That last point matters more than the format. Comments dropped 44% during a period when replies were skipped. The algorithm reads silence as disengagement and reduces reach accordingly.

The profile is the landing page

With impressions up 5× and profile clicks up 6×, we expected follower growth to follow. It did not.

The conversion gap — explosive reach, slow follows — points at one thing: the profile itself. A visitor who clicks through is interested. If they do not follow, the profile did not close.

This is a landing page problem dressed as a social media problem. The bio needs to answer clearly who you are and who should follow. The pinned post needs to be the strongest signal of what following is worth. The cover image needs to match the brand message.

None of that is about content strategy. It is about what happens after the content works.

Video changed the graph in one day

On February 2, 2026, we published the first video post on LinkedIn. The impression graph spiked the same day.

We had not planned a video strategy. We tested one post and read the result.

Video on LinkedIn gets treated differently by the algorithm — more surface area, autoplay in-feed, stronger engagement signal when watched through. The data is clear enough that video is now a standing format, not an experiment.

The same logic applies on X. Images and video receive a 2× algorithmic boost compared to text-only posts. We knew this in theory. The analytics confirmed it in practice and changed what we schedule by default.

What five months actually changes

The summary is not a set of tactics. It is a shift in mental model.

Social media for a Web3 brand is not a content calendar. It is a system of engagement signals — replies, profile quality, format choices, timing — that the algorithm reads and distributes based on quality indicators it does not publish but consistently rewards.

You can post every day and grow slowly. Or you can comment thoughtfully, reply to every response, and build a profile that converts the attention you earn.

The data does not suggest posting less. It suggests that what you do between posts matters as much as the posts themselves.