The new alpha in GTM is turning every conversion metric into an agent loop.
These agents will persistently work towards a verifiable goal and won't stop until that condition is met.
So you can set up daily agent loops to:
- increase outbound reply rates
- SEO / GEO impressions
- Make sure deliverability score stays healthy
Your job is to set up the infra that claude code can run uninterrupted and get out of it's way.
Just cracked +20'000 LinkedIn Followers.
It brought us +1k SignUps for our SaaS.
Here is the 7 steps I'd follow to do it again:
1) Clean up your profile
2) Define your 4 content pillars
3) Post each pillar once a week
4) Add 1 lead magnet / week (drives insane growth)
5) Comment under creators in your niche
6) Automate outbound to qualified post engagers
7) Scan for trending formats and emulate them
Just FYI it took me 3.5 years to get there and the last 10k followers came in a time span of 1.5 months.
So things take time to compound
The difference between 55% reply rate vs <1% is the list.
Here are the 5 data sources I mine
1. Pull commodity data from Apollo, BlitzAPI or Prospeo
2. Find lookalikes with DiscoLike, Ocean and Techstack with TheirStack
3. Add signals from PredictLeads or BlitzAPI
4. Scrape post engagers, profile viewers and followers on LinkedIn, yours and your competitors.
5. Use your own signups and opt-ins
Layer 1 is fine for high volume email, but I advice all our customers to start with the 5 and 4.
Vercel, Supabase and fly sucks.
We churned and are now fully self-hosting.
- it's so much more efficient
- and cheaper
Agents just want access to raw compute.
AI cooked coding. Vercel and Supabase wrap AWS. AWS wraps the compute.
Skip the wrappers, Fly too, and go straight to bare metal.
We're now on OVHcloud bare metal and just bought 3 more boxes for our coding agents.
Secure your capacity now. Compute is going through the roof (Hetzner already doubled their prices this summer)
Am I wrong?