Fifty thousand people in a room where nobody knew who was in charge.

The hardest product I have had to explain

Dreams Quest was building Origins: The Fall of Azoria, an experiential RPG where dreams and the subconscious intertwine with the physical world, built around life force, resonance and enlightenment.

Most Web3 games are explainable in a sentence. This one was not, and the abstraction was not a marketing failure. It was the actual product. Which meant every person joining was forming an impression of something genuinely difficult to describe, mostly from other community members.

Three problems that compounded

Scale had outrun structure

Over 50,000 members across Discord and Telegram, moderated by a team with no defined ownership. New moderators and ambassadors arrived without a clear answer to “which channels are mine, and what am I responsible for?”

Leadership was seeing a filtered view

No consistent feedback loop ran from moderators on the ground to the Community Lead and core team. Recurring issues surfaced late, or not at all.

Engagement was event-driven

Activity spiked on announcements and sagged between them, because nothing recurring gave people a reason to come back.

Giving the team ownership

I supervised six moderators and ambassadors, and the first change was the least glamorous: every person got explicit duties and named channels. Not a rota. Ownership. If something went wrong in a channel, there was a person whose channel it was.

I then built the reporting line that did not exist, giving the Community Lead and the Dreams Quest team detailed, ongoing performance feedback on each moderator and surfacing community issues directly rather than in summary. Leadership stopped getting the sanitised version.

The decision I would make again

Because the concept was abstract and easy to misread, I required that moderators preview the game directly before supporting users.

Not read the deck. Not memorise the FAQ. Play it.

The reason is specific to this product: when a moderator explains an abstract concept from marketing copy, they reproduce the copy, and if the copy is what confused the user in the first place, the loop never breaks. A moderator who has actually experienced the thing can find a different door in, one that suits the person asking.

Programming, so the community had a heartbeat

I built and ran a monthly content calendar so there was always something scheduled between announcements: four formats doing four different jobs.

• AMA sessions. Access to the team
• Round Table Sessions. The community talking back
• Quiz Nights. Rewarding the people who had actually paid attention
• Tournaments. Something for the competitive core to win

Sustained monthly, throughout my tenure.

Results

A working escalation path from moderators to the Community Lead and core team, where previously there was none.

“Precious was my direct hire at Dreams Quest and I am very happy with his work during my tenure there. Precious was proactive and engaged with the project and gave suggestions for improvement and actively sought out the information he needed to do his job excellently.” Andrew Shantz · Community Design Specialist, Dreams Quest

What this work taught me

This is where I learned that community work is operations. The moderation problem was not people being bad at moderating. It was that nobody had defined what they owned. Structure fixed it, not effort.

It is also where I first hit the problem I have worked on ever since: a product nobody can explain in a sentence will be explained badly by everyone else, in public, at volume. Someone has to decide how it gets described, and then make sure the people describing it have actually used it.