How to attribute which growth sources actually retain

How to attribute which growth sources actually retain

Not all joiners are equal — a member from a targeted niche post behaves nothing like one from a giveaway. Counting joins by source is easy; the playbook is measuring which sources retain.

Step 1: tag the source
Use invite-link tracking (Discord shows per-invite join counts; Telegram supports named invite links). One link per channel, post, or partner.

Step 2: build a retention table by source
For each invite source, compute week-1 and week-4 retention of its cohort, not just raw joins. This is where it gets uncomfortable.

What the data usually reveals
High-volume sources (giveaways, big shoutouts, list placements) tend to retain worst — they attract incentive-seekers and curiosity clicks. Low-volume targeted sources (a relevant niche thread) often retain 2-3x better per joiner. Volume and quality frequently trade off.

Step 3: reallocate
Optimize for retained members per unit of effort, not joins per unit of effort. A source delivering 500 joins at 5% week-4 retention loses to one delivering 100 at 30%.

The caveat
Invite attribution breaks when links get reshared, and sample sizes for small sources are noisy. Treat small-source retention rates as directional.

Open question: should you ever chase vanity growth knowing it dilutes the active ratio, purely for the social proof of a big member count?