Content language · English

The Command Loop: Gate, Rates, Weekly Decisions

Operator notes · 2026-11-03 · No hire guarantee · Digital download

Most job searches fail as operations problems, not talent problems. You apply. You wait. You refresh. You invent a story about “the market.” The board grows. Nothing decides. A Command Loop is the smallest closed system that forces a decision every week: gate what enters Applied, keep stage names honest, read rates by channel, then scale / change / stop in writing.

This is not motivation. It is ops. If you leave Sunday with vibes and a prettier spreadsheet, you held a status meeting. Status meetings are how pipelines rot politely.

The four moves (say them in order)

Gate — nothing becomes Applied until the Application Quality Gate passes.
Stages — locked vocab: Sourced → Applied → Screen → Loop → Offer → Closed.
Rates — Screen/Applied, Loop/Screen, Offer/Loop, signal events per hour, by channel.
Weekly command — scale, change channel/motion, or stop—with one-line reasons.

Skip Gate and your rates lie. Skip rates and your weekly “strategy” is mood. Skip the weekly write-up and next Monday is a random restart of last week’s spray.

Move 1 — Gate before Applied

Sourced is allowed to be messy. Applied is not. An Application Quality Gate is a short, ruthless checklist: file hygiene, JD fit, keyword mirror, honest metrics, channel + next action. Fail required items → stay Sourced. No Easy Apply theater. No “I’ll fix the resume after I click submit.”

Why operators care: ungated spray poisons the denominator. Screen/Applied collapses, so you “optimize” by spraying harder. Referrers notice. Signal per hour dies while you feel busy. Gate protects metrics and people who might put their name on you.

Operator rule: Three Gate fails on the same resume_version → freeze Applies. Repair evidence and targeting first. Volume is not a strategy when the packet is broken.

Move 2 — Locked stages (no creative renaming)

Sourced → Applied → Screen → Loop → Offer → Closed

Onsite or video is a format under Loop—not a new stage. Autorejects are not Screens. “Recruiter liked my LinkedIn” is not Applied. Stale and Cold are timer states with next_action rules, not feelings—see Stale & Cold Pipeline Rules.

If two people cannot classify the same row the same way, your funnel math is fiction. Lock the words. Argue about decisions, not vocabulary.

Move 3 — Rates that change next week’s hours

Counts without rates are a museum. Minimum scoreboard (from your sheet only—no invented industry benchmarks):

Screen / Applied (overall and by channel)
Loop / Screen
Offer / Loop
Signal events / hour (Screens + Loops + meaningful human replies; autoresponders don’t count)

Tag every Applied with a channel: job_board, company_site, referral, recruiter_inbound, community, outbound_hm. Compare weeks in writing. “Feels worse” is not a row. For the worksheet shape, see Job Search Funnel Metrics.

Evidence still matters under the hood: an Evidence Bank keeps resume bullets as views over proof, not adjective spam. Gate without proof is checkbox cosplay.

Move 4 — Weekly Command Review

Protect 30–40 minutes the same weekday. Inputs open before you start: kanban, metrics, touch log, targeting, debriefs. Output is three forced decisions with one-line reasons—scale, change, or stop—per channel and for the hunt as a whole. Detail lives in Weekly Command Review.

Examples of honest stops: a board channel with adequate sample and dead Screen rate; a Priority C spray habit that never produces signal; a resume_version that keeps failing Gate. Stopping is a command, not a moral failure.

What the Loop is not

Not a hire guarantee. Process control ≠ outcomes.
Not a Notion aesthetic project. Core uses HTML and XLSX; other kits use the formats listed on the product cards.
Not a leetcode subscription with stages sprinkled on. Interview sprint kits deepen Screen/Loop execution; they do not replace Gate.
Not fake conversion claims. Your rates come from your rows. SAMPLE files in kits are fictional samples, not testimonials.

Minimal Day-0 install

Write floors (cash / TC) so Offer panic has a precommitted floor.
Cap Priority A/B targets (≤12 is a workable starting rule of thumb—adjust to your capacity).
Import a kanban CSV; filter by stage and next_action_date.
Run Gate on the next real application before you mark Applied.
Put Weekly Command Review on the calendar.

If Screens are already booked, deepen interview execution next. If Sundays are mushy after Core is running, deepen diagnosis. If the role family is locked and evidence is thin, add one role vertical—not five. Start with Core; add Interview / Offer / Role kits when you need depth.

When interviews or diagnosis join the Loop

Once Gate and stages are live, bottlenecks rename themselves. Screens on the calendar → deepen interview execution (STAR harvest, mocks, same-day debriefs)—do not abandon Gate to “prep harder.” Sundays that produce no channel decision → deepen funnel/channel diagnosis until scale/change/stop is forced in writing. Role-family evidence thin after Core is running → add one vertical Gate/evidence pack, not a stack of every role kit. Offers arrive → floors and scripts, not a new spray campaign. The Loop stays the spine; specialty kits attach to named pain.

Close the loop every week

Monday’s work should inherit Sunday’s decisions. If you scaled referrals, the hours show up as warm touches—not more Easy Apply. If you changed motion, the Gate and packet change before volume returns. If you stopped a channel, the board shows Closed with a reason, not ghost rows.

That is the Command Loop: Gate protects the denominator, stages keep language shared, rates make tradeoffs visible, weekly writing reallocates hours. Busy without that loop is just expensive hope.