Metrics and go-to-market
The Bullseye framework
Brainstorm many channels, test a few cheaply, then focus on the one that works. Give every test a kill criterion.
Gabriel Weinberg and Justin Mares (Traction) formalize the "one channel" advice as the Bullseye framework. The failure mode is permanent parallelism — always exploring, never compounding.
Concept
Channel-market fit means a path reaches your ICP at a CAC you can bear. Bullseye is how you find that path without betting the company on five untested motions. See channel-market fit.
The rings
- Brainstorm ideas across many channel types (do not dismiss unfamiliar ones). See the catalogue.
- Rank by expected impact, cost, and time-to-learn.
- Prioritize a short inner ring (often 3).
- Test in parallel with small, time-boxed experiments.
- Focus on the winner; revisit when it saturates.
Trade-off
Parallel small tests learn faster early. Serial focus wins once a channel works. Permanent "exploring" is how you get zero compounding channels.
Walking through a cheap test
A good test has a kill criterion in advance: "If after $3k and 3 weeks CPC-to-paid CAC is above $800, stop paid social." Without kill criteria, every channel becomes a zombie budget line.
Cheap tests look like: 50 personalized outbounds, one landing page plus $500 of ads, three partner intros, a single conference booth shared with a peer, an integration MVP that five design partners use.
Expensive tests look like: hiring five AEs before inbound exists, or a six-month brand campaign with no conversion instrumentation.
Weinberg's DuckDuckGo story: they cycled SEO, content, social, PR and BD across years as each phase's channel plateaued — same company, new bullseye rings.
Common mistakes
- No kill criterion. Zombie channels live forever.
- Expensive tests first. Do not hire a sales team to "see if outbound works."
- Never focusing. Parallelism is for learning. Scale is for the winner.
Check yourself
What belongs in a test design before you spend?
A kill criterion: spend, time window and the CAC (or reply rate) that means stop.
Why test three channels in parallel, then pick one?
Early parallelism learns faster. After a winner, serial focus is how you compound.
Key takeaways
- Brainstorm widely, rank, test a short inner ring, then focus.
- Every test needs a kill criterion in advance.
- Cheap tests are 50 outbounds or $500 of ads. Expensive tests are headcount and brand campaigns.
- When the winner saturates, run Bullseye again.
Further learning
Traction — introductory chapters (PDF excerpt)Traction-1-3.pdfleananalyticsbook.comNext, why winners stop working.