700k+ views in the last 24 hours, so there are suddenly a LOT of new people here.
probably a good time to explain what i actually do ๐ค
iโm Abhishek.
- growth + community lead at @DabbaNetwork, working across GTM, community, content, partnerships and ecosystem growth
- ex-software dev who somehow ended up on the distribution side of tech
- helped 3 AI startups figure out their early GTM, positioning + distribution
- built @CodeSamaaj, a community/network of 2,500+ devs
- worked on 40+ ecosystem partnerships across web3
- organised 10+ IRL events across South India with 1,000+ people attending
- grew this account to 25k+ followers and 55M+ impressions, mostly by figuring out what actually gets people to stop scrolling
- talent scout at @SuperteamTalent, helping find good people for startups
most of what iโm interested in sits somewhere between product, distribution, community and marketing.
basically: figuring out how good products get noticed, talked about and adopted.
Do nice people still exist?
people who treat others with respect and donโt need a reason to be nice to someone else?
I donโt even get Good Mornings nowadays ๐ฅฒ
5 places where users publicly beg for a feature:
- feature-request forums on other products (old threads with 100s of "+1"s)
- public roadmap boards (upvote counts = demand ranking)
- competitor github issues labeled "enhancement"
- 3-star reviews on playstore/app-store
- "is X dead" / "X alternative" threads on reddit or quora
you'll find your first 100 users there.
Clay went $1M โ $100M arr in 2 years:
- 6 flat years before the spike
- never churned an enterprise customer
- enterprise NRR >200%
- invented the "gtm engineer" job title so users sold it for them
overnight success that took 8 years ๐
by the time you hit the 100 users milestone, you already know:
- a pretty landing page without a channel is just a pdf
- your next 1000 users are already complaining somewhere
- "we need a growth hire" usually means "we haven't talked to 20 users"
- distribution compounds slower than product demos feel
- the founder who answers support tickets learns gtm faster than the one who only posts launch day
you need to do this today!!!!
> paste a competitor's last 50 g2 reviews into claude
> ask for the 5 most repeated complaints + the exact words buyers use
> put those words in your landing page headline this week
> do this every few weeks and track the changes in website visitor behaviour.
eventually youโll land on the best converting landing page๐ซถ๐ป
bhai, tuesday mornings hit different:
- calendar looks full
- interns (grok bots) are ready with this weeks play
- DMs are flooded
- coffee is ready, but am I? (doesn't matter)
and btw now im @thatGTMdude ๐
Killer insight from an article I was reading..
pocket fm doesn't rewrite an entire ad just because it's underperforming. they first identify exactly where it failed.
- under 2.5% ctr โ rewrite the hook
- under 55% 3-sec through-play โ rewrite the cliffhanger
if it fails both, a creative director steps in.
that's the whole QA loop.
steal users from your competitors:
> pick 3 competitors
> every monday, paste their changelog + pricing page into claude
> ask: "what got killed, repriced, or limited this week?"
> if something broke for their users, draft a 1-line migration offer the same day
break the rules...