when to kill a channel: decision criteria

TL;DR

Set a kill threshold before you start a channel (effort in, results out, by a date), and kill it when the date arrives and the threshold is missed. Channels die by thresholds, not by feelings. Write the threshold down at launch, review on the date, and shut down cleanly so a future revisit starts from notes instead of from scratch.

when to kill a channel: decision criteria

Use this when

• A channel has consumed weeks with no measurable return
• You keep a channel alive "just in case" with no owner
• Two channels compete for the same effort and one must go
• A founder asks why you are still doing something that is not working
• You are planning next quarter and need to free up capacity

Not for this skill when

• The channel is less than one full test cycle old (let the test finish)
• The problem is execution quality, not the channel (fix the execution first)
• You are killing a product feature with existing users (migration concerns apply)
• It is a paid channel mid-flight with contractual spend (different decision)

Steps

1. Write the kill threshold at launch, before any effort goes in. Effort budget, result target, and a decision date. "Give it a few months and see" is not a threshold; "40 hours and 200 qualified visits by Dec 1" is.

   channel: [name]
   effort budget: [hours or dollars]
   result target: [metric and number]
   decision date: [date]

Expected output: a threshold note per active channel. Success check: a stranger could apply it without asking you what you meant.

1. Track the channel's cost honestly while the test runs. Hours spent, money spent, opportunity cost named. Channels look free when you do not count your own time; they are never free.

   spent so far: [hours], [dollars]
   on pace for budget: [yes | no]

Expected output: a running cost line per channel. Success check: the number surprises you at least once, which means it is honest.

1. On the decision date, apply the threshold without renegotiating it. Hit means continue with a new threshold; miss means kill. Moving the goalposts on decision day converts every test into an indefinite maybe.

   decision: [kill | continue]
   evidence: [numbers vs threshold]

Expected output: a dated decision per channel. Success check: at least one channel got killed this year, or your thresholds are decorative.

1. Shut down cleanly and take notes. Remove or pause the presence, redirect anything valuable, and write down what you learned and what would change the verdict. A clean kill preserves the option to return; a messy fade leaves landmines.

   [ ] presence removed or paused, links redirected
   [ ] learnings noted: [what did not work and why]
   [ ] revisit condition: [what would change the verdict]

Expected output: a closed channel with notes. Success check: you could relaunch in a day from the notes alone.

1. Reallocate the freed capacity explicitly. Killed effort that drifts into "general work" teaches the team that killing changes nothing. Name where the hours go next.

   freed: [hours/week] -> reallocated to: [channel or project]

Expected output: a reallocation line. Success check: the receiving project can point to the added capacity.

Variant phrasing

How to know when a marketing channel is not working

Diagnostic phrasing. Steps 1 and 3: the pre-set threshold and the no-renegotiation review.

Sunk cost fallacy in marketing channels

Psychology phrasing. Step 3 is the antidote: the decision was made at launch, decision day just executes it.

Channel audit framework

Audit phrasing. Run steps 1 through 3 across all channels at once, quarterly; it is the same skill applied as a batch.

When to quit social media for a devtool

Specific-channel phrasing. Same thresholds, same shutdown checklist; note the revisit condition carefully since audiences migrate.

Why it happens

Channels fail slowly and hope lasts longer than evidence. Without a pre-written threshold, every review becomes a debate between the effort already spent and the results not yet arrived, and effort usually wins because it is vivid. The threshold moves the decision from decision day (emotional) to launch day (rational). Clean shutdowns matter because most "dead" channels get a second life when the audience or the product changes.

Edge cases / pitfalls

• Brand or community channels have non-metric value. If a channel's job is presence rather than growth, its threshold should say so ("maintain, 2h/month") instead of pretending it is a growth bet.
• Killing during a known seasonal dip misreads the data. Set decision dates outside predictable troughs, or normalize for them.
• A channel can fail while the strategy is right. Distinguish "wrong channel" from "bad execution" before the kill; the notes in step 4 are where this lives.
• Do not kill by neglect. An unmaintained channel that still ranks or refers is an asset in decay; either maintain it minimally or shut it down properly.

Provenance

Resolved from the public thread: https://vectle.com/posts/pst_d1ajA3AjOkdZYpoPyZqU-Q