When to Kill a Channel (And Explain It to Your CFO)

Q: How do I know when a marketing channel is dead?

Three diagnostic signals tell you a channel is done. The 3-cycle test: three optimization cycles within 90 days produce no measurable CPL lift. The audience tax: CPM rises 3 quarters in a row while audience size stays flat, which means you're paying more for less reach as the pool saturates. The pipeline disconnect: the channel's contribution to closed-won revenue has dropped 30% or more over 2 quarters even though spend hasn't moved. One signal alone is noise. Two signals together is a kill candidate. Three signals together is a kill.

Q: What is a 30-day wind-down plan and why do I need one?

A 30-day wind-down is a 3-phase exit plan that lets a channel die without taking pipeline with it. Days 1 to 7: stop new spend, transition warm leads to active channels, notify the agency. Days 8 to 21: keep landing pages live but cut top-of-funnel traffic, let in-flight leads convert. Days 22 to 30: archive assets, redirect URLs, hand the budget to the channel that's earning it. Without a wind-down, sales asks why demo requests dropped, the agency wants to know what happened to the retainer, and the CRM has hundreds of mid-funnel leads that nobody owns.

Q: How do I get a CFO to approve killing a marketing channel?

Use a 4-line CFO brief. Line 1: the signal that triggered the kill, with the supporting metric (e.g., 3-cycle test failed; CPL up 22%). Line 2: the spend in question, with the date the channel went into the red. Line 3: the wind-down date and the protected pipeline number. Line 4: where the budget reallocates and the expected lift. Marketing asks 'should we kill this channel?' Finance asks 'what's the cost of running it for one more quarter?' The 4-line brief answers both, with numbers.

Q: What's the difference between a channel that needs optimization and one that needs to die?

A channel that needs optimization responds to changes. Cycle 1 of optimization gets the obvious wins (creative refresh, audience tightening). Cycle 2 catches the residuals. Cycle 3 either confirms a new floor or confirms there's nothing left. A channel that needs to die has been through 3 cycles inside 90 days with no measurable lift, the audience pool is saturated (CPM up, audience flat), or the pipeline contribution has dropped while spend held. Optimization keeps a working channel working. The kill decision applies to channels that have already failed the optimization test.

William DeCourcy · May 11, 2026

$340,000 a Year on a Channel That Stopped Working 9 Months Ago

A B2B team was spending $340,000 a year on a channel that had stopped converting 9 months earlier.

They didn't kill it because nobody had run the math.

That's a zombie channel. Spend keeps going out. Performance flatlined last quarter (or 3 quarters ago).

The channel runs because killing things is harder than letting them limp along, and because nobody on the team is incentivized to be the one who said "shut it down."

This post breaks down 3 signals that tell you a channel is done, the 30-day wind-down plan that protects the pipeline while you exit, and the 4-line CFO brief that turns the kill into a finance conversation.

Key Takeaways

3 signals diagnose a dead channel: the 3-cycle test failure, the audience tax, and the pipeline disconnect. One signal is noise; two is a candidate; three is a kill.
The 30-day wind-down has 3 phases (stop spend, let in-flight convert, archive and reallocate). It's the difference between killing a channel cleanly and watching pipeline drop with it.
The 4-line CFO brief turns a 6-week argument into a 3-minute approval. Signal, spend, wind-down date, reallocation. No deck.
Most teams over-optimize dying channels. 6 of every 9 optimization cycles on a saturated channel are emotional optimization wearing analytical clothing.
Frame the kill decision as a finance decision. Finance asks "what's the cost of running this channel one more quarter?" That's the framing that gets approval.

Signal 1: The 3-Cycle Test

Three optimization cycles within 90 days, no measurable CPL lift. The channel is done.

The mechanic: cycle 1 gets the obvious wins (creative refresh, audience tightening, bid strategy adjustment). Cycle 2 catches the residuals. Cycle 3 either confirms a new floor or confirms there's nothing left to find.

A SaaS team optimized their LinkedIn ads 3 times in 90 days. Same audience, fresh creative, tighter targeting. CPL didn't move.

The 4th cycle wasn't going to move it either, but they ran it anyway.

That's the failure mode. Most teams give a dying channel 9 cycles. Six of those are emotional optimization wearing analytical clothing, because cycle 4 is when the team starts arguing about whether the channel is "really" done or "just having a bad quarter."

The discipline: 3 cycles, hard cap. If cycles 1, 2, and 3 produce no measurable lift, the channel has failed the test. Move to signal 2.

Signal 2: The Audience Tax

CPM rises 3 quarters in a row. Audience size stays flat. Every dollar buys less reach than the dollar before it.

That's the audience tax. The pool is saturated.

The math gets uglier each quarter because the audience doesn't refill on your timeline. It refills on theirs.

A retail brand watched CPM rise 18% per quarter for 4 quarters on Meta. Audience didn't grow, and creative, bid strategy, and targeting were all fine. The pool was just done.

You have 3 options when you hit the tax.

Expand the audience definition. Low odds, big lift if it works. Try this once.

Cut spend by 50% and let the pool reset. Slow, painful, sometimes works. Try this twice if the math says the channel could come back.

Wind it down and reallocate. The right answer if option 1 didn't work and option 2 has been tried.

If your last 4 quarters look like this and option 1 hasn't worked, you're at option 3.

Signal 3: The Pipeline Disconnect

The channel's contribution to closed-won revenue has dropped 30% or more over 2 quarters. Spend has held steady or grown.

This is the most important signal because it's the one teams miss. Top-of-funnel metrics (clicks, impressions, MQLs) often hold up while downstream conversion craters.

The leads still come in. They just don't close anymore.

A fintech company saw lead volume from a channel hold flat across 6 quarters. SQL-to-close rate on those leads dropped from 19% to 7% over the same window.

Spend didn't change. Revenue contribution from that channel fell 63%.

The channel was generating leads. It just wasn't generating the right leads anymore. The audience that was clicking through had shifted, and the new audience didn't convert.

The diagnostic: pull last 6 quarters of channel-attributed closed-won revenue. Divide by spend in the same quarter. If the per-dollar revenue contribution has dropped 30% or more over 2 of the last 4 quarters, you're looking at a pipeline disconnect.

The Cost per Revenue calculator runs this math for you. Plug in spend and closed-won by channel; it tells you the per-dollar revenue contribution and flags channels where the trend has rolled over.

For more on why CPL alone isn't enough to make this call, see Stop Measuring Cost Per Lead. Start Measuring Cost Per Revenue Dollar..

How Many Signals Before You Kill?

One signal alone is noise. CPL going flat for one cycle could be seasonality.

CPM rising one quarter could be an auction shift. Pipeline contribution dropping one quarter could be a sales-side issue.

Two signals together is a kill candidate. Schedule the wind-down conversation.

Pull 6 quarters of data on both signals. Confirm the trend isn't a measurement error or an attribution problem.

If your attribution is unreliable, fix that first. A bad attribution model can make a working channel look dead. See Your Attribution Model Is Lying to You for the diagnostic.

Three signals together is a kill. The channel has failed the optimization test, the audience is saturated, and the pipeline contribution has fallen. There is no fourth optimization cycle that fixes that combination.

The 30-Day Wind-Down

Most teams skip this step. They flip the channel off in a Monday meeting and move on.

Two weeks later, sales is asking why demo requests dropped, the agency wants to know what happened to the retainer, and the CRM has 400 leads in mid-funnel that nobody owns.

The 30-day wind-down prevents that. It's a 3-phase plan that lets a channel die without taking pipeline with it.

Phase 1 (Days 1 to 7): Stop the Spend

Pause new ad spend. Notify the agency or in-house team running the channel. Pull the warm leads (anyone who entered the funnel in the last 30 days) and route them to active channels for nurture.

This is also when you tell sales. They need 7 days' notice that demo volume from this channel is going to drop, so they can plan capacity.

Phase 2 (Days 8 to 21): Let In-Flight Convert

Keep the landing pages live. Don't drive new traffic to them. Let the in-flight leads (the ones already in the consideration window) convert through their normal cycle.

This phase is the biggest source of pipeline protection. A typical B2B sales cycle is 30 to 90 days.

The leads who entered the funnel in the last 60 days haven't closed yet. If you turn the pages off on day 1, those leads stall and disappear.

Days 8 through 21 lets them convert. The cost of keeping a static landing page live for 2 weeks is near zero. The cost of losing 60 in-flight leads is real money.

Phase 3 (Days 22 to 30): Archive and Reallocate

Turn off the funnel. Archive the creative, the landing pages, and the campaign assets (you'll want them if the channel ever comes back, or as reference for the post-mortem).

Redirect any URLs that are getting organic traffic to the closest live equivalent.

Hand the budget to the channel that's earning it. The reallocation should happen on day 22 so the receiving channel can spool up before the dead channel's contribution to pipeline runs out.

The wind-down plan exists so a channel can die without taking pipeline with it. Build the plan before you need it. By the week you decide to kill, the exit plan should already exist.

The 4-Line CFO Brief

A 4-line brief turned a marketing kill decision into a 3-minute finance conversation.

The marketing team had been arguing for 6 weeks that the channel was done. The CFO kept asking why spend hadn't moved.

Both sides were right. Then someone wrote the brief.

Line 1: The signal that triggered the kill, with the supporting metric.
        (Example: "3-cycle test failed; CPL up 22% over 90 days.")

Line 2: The spend in question, with the date the channel went into the red.
        (Example: "$340K annual; ROI inverted in Q3 of last year.")

Line 3: The wind-down date and the protected pipeline number.
        (Example: "Wind-down complete by [date]; protects $1.2M in-flight pipeline.")

Line 4: Where the budget reallocates and the expected lift.
        (Example: "Reallocates to YouTube + paid search; modeled +14% CPR.")

4 lines. No slides. No 30-page deck.

The CFO signed off in 3 minutes.

The brief works because it answers the question the CFO is actually asking: "what's the cost of running this channel for one more quarter, and what does the alternative look like?" Marketing tends to argue about whether the channel is working. Finance asks the cost-of-continuation question, and the 4-line brief answers it with numbers.

Failure Modes (What Goes Wrong If You Skip This)

Killing without a wind-down. You drop demo volume by 30% the week the channel goes off. Sales hits its number anyway because in-flight leads are still closing, but the next quarter's pipeline cliff is real and the executive team doesn't see it coming.

Killing without the CFO brief. The conversation runs 6 weeks. Marketing argues the channel is dead; finance argues the spend is small enough to keep going. Neither side has the framing the other is asking for, and the decision either doesn't happen or happens too late.

Killing on one signal. You shut down a channel that was just having a bad cycle. The cycle was seasonal. You spent 6 months rebuilding the channel from scratch when you could have ridden out the soft quarter.

Optimizing past the 3-cycle limit. You give a dying channel 9 optimization cycles. Cycles 4 through 9 cost you the budget that could have gone to the next channel. The team gets demoralized because nothing they do moves the needle.

Skipping the reallocation step. You kill the channel on day 30 and the budget sits in a holding account for 6 weeks while procurement moves it. The receiving channel can't spool up. You lose the pipeline contribution from the killed channel and don't replace it for a quarter.

Knowing when to kill is half of the discipline. The structured exit, the ready-to-go reallocation, and the 3-minute CFO brief are the other half.

Run the Math This Week

Pull 6 quarters of channel-level data: spend, CPL, CPM, audience size, MQLs, SQLs, closed-won revenue.

For each channel, run the 3-signal check.

The 3-cycle test: did the last 3 optimization cycles produce a measurable CPL lift? Yes or no.

The audience tax: has CPM risen 3 quarters in a row while audience stayed flat? Yes or no.

The pipeline disconnect: has per-dollar revenue contribution dropped 30% or more over 2 of the last 4 quarters? Yes or no.

Channels with 0 or 1 yes are healthy or recoverable. Channels with 2 yeses are kill candidates; schedule the conversation. Channels with 3 yeses are dead; build the wind-down plan and write the CFO brief.

Run this analysis quarterly. The teams that kill channels on a quarterly cadence have lower zombie spend than the teams that wait until someone notices. By the time someone notices, the channel has usually been dead for 9 months.

Frequently Asked Questions

How do I know when a marketing channel is dead?

Three signals together: the 3-cycle test fails (no measurable CPL lift over 3 optimization cycles in 90 days), the audience tax shows up (CPM up 3 quarters in a row, audience flat), and the pipeline disconnect (per-dollar revenue contribution down 30% over 2 quarters). One signal is noise; two is a candidate; three is a kill.

What is a 30-day wind-down plan and why do I need one?

A 3-phase exit plan that lets a channel die without taking pipeline with it. Days 1 to 7: stop spend, transition warm leads, notify the agency. Days 8 to 21: keep pages live but cut top-of-funnel traffic, let in-flight leads convert. Days 22 to 30: archive assets, redirect URLs, reallocate the budget; without the plan, sales loses 30% of its demo volume the week you kill the channel.

How do I get a CFO to approve killing a marketing channel?

Write a 4-line brief: the signal, the spend and date the channel went into the red, the wind-down date and protected pipeline number, the reallocation and expected lift. The CFO is asking "what's the cost of running it one more quarter." Answer that question with numbers.

What's the difference between a channel that needs optimization and one that needs to die?

A channel that needs optimization responds to changes within 3 cycles. A channel that needs to die has been through 3 cycles in 90 days with no measurable lift, has a saturated audience pool, or has a pipeline contribution that's fallen even though spend held.

Optimization keeps a working channel working. The kill decision applies to channels that have already failed the optimization test.