/focus

name: focus
description: "Use this skill when a solo founder wants a gut check on how to spend time or money: 'is this worth my time,' 'should I keep doing X,' 'I'm not getting results from X,' '80/20,' 'should I sign up for this,' 'should I buy this,' 'should I run ads,' 'should I hire someone,' 'should I raise money,' or 'which feature should I build next.' Returns one verdict and a plan for this week. Not for testing whether an idea has demand (use validate) or for finding the founder's overall stage (use journey)."
metadata:
  version: 1.0.0

Focus

Judge one activity, one decision, or one feature choice against the founder's stage. Return one verdict and one plan for this week.

Step 1: Get the facts

Ask only what you cannot infer. Two or three questions:

• What exactly is it? Not "marketing", but "50 cold emails a week to dental offices."
• How many paying customers do you have?
• For an activity: hours per week, and results so far. For a decision: the cost, and whether you can undo it.

Then use the mode that fits.

Mode A: An ongoing activity

Apply two tests.

1. **Ceiling.** Could this activity move the business at this stage, even if done perfectly? SEO with zero customers, ads without a known conversion rate, and integrations for a product nobody uses all have a low ceiling. 2. **Execution.** Is the founder doing the strong 20%, or spreading effort across all of it? Examples: 100 generic emails instead of 10 researched ones, or four platforms instead of one.

Verdict:

• **Kill:** low ceiling, or a much better alternative exists.
• **Refine:** right activity, spread too thin. Cut to the 20%.
• **Pivot:** right goal, wrong tactic for this stage. Name the tactic to switch to.

Mode B: A one-time decision

Examples: a yearly contract, a large purchase, a first contractor, a paid course, ads, raising money.

Ask five questions: 1. Did a paying customer ask for this, or need it? 2. Can the founder do it by hand or for free first? 3. Can it be undone? Yearly contracts, hires, and investor terms are hard to undo. 4. What does it do to monthly profit? 5. Does it bring the next customer milestone (10, then 100) closer?

Verdict: **Do it**, **Do a smaller version** (name it), or **Do not do it**. For raising money, also use **finances**.

Mode C: Which feature next

1. Name the one problem each candidate solves, and how many paying customers have it. 2. Build what 3 or more paying customers need. One loud request is an anecdote. 3. Pick one. Cut it to what ships in a week, and use **build**.

Say no when a feature serves fewer than 1 in 10 customers, adds complexity for the rest, or a good-enough workaround exists.

What pays off at each stage

| Activity | Before 10 customers | 10–100 | 100+ | |---|---|---|---| | Direct outreach | High | Medium | Low | | Being useful in communities | High | High | Medium | | Content and SEO | Low: too slow | Medium: start | High: compounds | | Paid ads | Low: unknown conversion | Low to medium: small tests | High: scale what converts | | Referrals | Low: too few users | High | High | | New features | Only if customers ask | Low: fix what exists | Medium | | Bug fixes | High | High | High | | Onboarding polish | Low: onboard by hand | High | Medium | | Automation and tooling | Low: manual is fine | Medium | High | | Pricing changes | High: most founders underprice | High | Medium | | Hiring | Low | Low to medium | Medium |

The common trap: choosing what feels scalable (SEO, ads, automation) before anything works by hand.

Output

Verdict: <Kill / Refine / Pivot, or Do it / Smaller version / Do not, or the chosen feature>
Why: <2-3 sentences: the ceiling, the cost, the alternative>
This week:
- Stop: <specific>
- Start: <specific, with the 20% version>
- Time or money freed: <amount, redirected to what>
Trap to avoid: <one>