Superpowers for decisions and hiring
The ten runs for when the next move is a choice, not a task.
The steps
- Which of these you want — Five of these attack a plan and they are not interchangeable. A decision note weighs options you already have. A scope review cuts a plan against a deadline. Inversion lists how it fails. A pre-mortem assumes it already failed and looks for what is fragile. A position analysis asks whether the ground is worth holding at all. Pick by the question you are stuck on; running two of them on the same plan is a fair way to break a tie.
- Write a decision note — Use it when a decision keeps being reopened because it was never written down.
**You get** — the decision restated — which is often the whole value — weighted criteria with the reason for each weight, a scoring matrix, three scenarios, and a recommendation rather than a list.
**Stays yours** — deciding. It recommends, and names what would make the recommendation wrong; the choice is signed by you.
Run by the **Advisor** · ~6 min.
- Review the company — Use it when there is more going on than you can hold in your head.
**You get** — what advanced with evidence rather than activity, what stalled and for how long, and what needs a decision from you with options — written to be blunt, since a reassuring review has failed.
**Stays yours** — the decisions it surfaces.
Run by the **Founder** · ~10 min · sharper with Stripe, Google Analytics, Shopify.
- Write the investor update — Use it when the update is due and the what-did-not-work section is the one you are avoiding.
**You get** — metrics with their change against last period, what worked, what did not at the same length, and asks specific enough to act on.
**Stays yours** — sending it. Anything it could not measure is marked unmeasured rather than filled in with a plausible number.
Run by the **Founder** · ~8 min · sharper with Stripe, Google Analytics.
- Review a plan's scope — Use it when the plan is ambitious and the date is not moving.
**You get** — every item sorted keep / cut / defer with a reason, what the plan forgot — migration, support load, training, documentation — and the prediction of what it will look like at the deadline if nothing is cut.
**Stays yours** — cutting. It says what to drop and why; removing it is a conversation with whoever owns the plan.
Run by the **CTO** · ~6 min.
- Find where you can be the only one — Use it when you are competing on the same axis as everyone and it is turning into price.
**You get** — the positions nobody occupies, each tested for defensibility — what stops the biggest player copying it next quarter — and what has to be given up to hold it.
**Stays yours** — taking one. It names what has to be given up to hold each position — giving it up is the decision.
Run by the **Advisor** · ~8 min · sharper with Firecrawl.
- Map your automation roadmap — Use it when you have decided to automate and the order matters more than the list.
**You get** — the binding constraint named first, the work sequenced in waves where wave one pays for wave two, payback per initiative, and what should not be automated because judgement or a relationship is the actual product.
**Stays yours** — building it. This sequences a programme; the individual automations get costed by the automation audit.
Run by the **Advisor** · ~10 min.
- Work out how this fails — Use it when a project matters enough that you want its failure routes before its plan.
**You get** — every route to failure across four kinds — execution, demand, people, and succeeding into a worse problem — ranked by likelihood times damage, each with a guardrail to put in place now rather than a resolution to be careful.
**Stays yours** — putting the guardrails in.
Run by the **Advisor** · ~8 min.
- Prepare a hire — Use it when you are about to post a role and everything after posting is undecided.
**You get** — what success looks like by month six, an advert written around what the person will do, an interview grid probing past behaviour rather than opinions, a scorecard with weak / adequate / strong per criterion, and a 30/60/90 plan.
**Stays yours** — posting it and running the interviews.
Run by the **HR Manager** · ~8 min · sharper with Greenhouse.
- Screen the applications — Use it when there are more applications than anyone will honestly read.
**You get** — every application read and scored against a grid built from the brief before reading, ranked interview / maybe / no with the CV quoted, the candidates a keyword filter would have lost flagged, and the rejection written.
**Stays yours** — the interviews. It never judges on name, age, gender or address — and ignores them where a CV shows them.
Run by the **HR Manager** · ~14 min · sharper with Greenhouse, Gmail, Outlook or AgentMail.
- Screen every applicant — Use it when applications land one at a time and the first read either never happens or happens differently depending on who is free that afternoon.
**You get** — a webhook URL to paste into your ATS or application form. Every application that hits it is checked against the requirements you wrote, quoted line by line, and returned as an internal evaluation sheet naming the one fact the recommendation turns on and what would change it.
**Stays yours** — the decision, and what may never count. It recommends and never rejects — no applicant is ever told they are out by this, and the only thing it may send them is a neutral acknowledgement.
Run by the **HR Manager** · ~2 min · sharper with Gmail, Outlook or AgentMail, Greenhouse or Lever.
You should see
You can name the one run that produces the thing you were about to make by hand.
The OpenLabor Academy