Which channel earnedanother cycle?
Most channels are kept alive by sunk cost and a dashboard nobody reads. This one asks two questions a channel has to answer before it gets another quarter of your life: did it produce an action you can trace, and did it produce it at a price you set? Both, or it has not earned it.
Nobody kills a channel. They just stop looking at it.
An action you can attribute, and an action you can afford at your own ceiling. One of the two is not evidence — it is half an argument.
The shipped sample includes a channel scoring 92 out of 100 that reads TREADING WATER, because it cost $266.67 per qualified action against a $150 ceiling.
No industry average, no conversion rate, no invented multiplier. Your target, your ceiling, your definition of a qualified action.
The failure mode is never a decision. It is a channel that produced one good week in March, has cost eleven months of Tuesdays since, and has never been formally killed because nobody wants to be the one who says it did not work. The Assay says it, from your own numbers, in one column.
Raise the ceiling by a dollar. Watch a verdict change.
Assaying Paid newsletter sponsorships. Every threshold below is your number. Nothing here is benchmarked against anyone else.
Attribution confidence
Can you say the action came from here?
Repeatability
Could you do it again on purpose?
Cadence held
Did you run it as designed for the period?
Compounding surface
Does past output keep working?
Score
92/100
Band: COMPOUNDING
Cost per action
$266.67
your ceiling $150.00
Derived marks
actions 2/2 · cost 1/2
You cannot mark past these.
TRIGGER 1 — PASS
attributable — an action you can trace here
TRIGGER 2 — FAIL
affordable — at or under the ceiling you set
GATE CAPPED one trigger failed, so this cannot read COMPOUNDING however high the score is.
Verdict
TREADING WATER
Fix first: Cost per qualified action
This grades a channel, never a person. It uses your numbers only — no benchmark, no industry average, no invented multiplier. Not financial advice.
Load C-02: a 92-point channel capped to TREADING WATER because trigger 2 failed. Then load C-03: an 80-point channel capped by the other trigger, with attribution at zero. Same cap, opposite branch. C-04 is the one that matters most — it scores 61, which is a middling band, and both triggers fail, so it reads CUT IT.
This is the real output of the shipped engine on the shipped sample.
Nothing below is written by hand. It is the verbatim stdout of python3 engine.py channels-sample.csv, captured programmatically and diffed against this page before publishing.
============================================================================ CHANNEL CONTINUATION ASSAY - CCA-099 - RedHub AI LLC ============================================================================ Did this channel earn another cycle? Your numbers only - no benchmarks. Grades channels, never people. Nothing is fetched and nothing is sent. ---------------------------------------------------------------------------- CHANNEL C-01 [Founder LinkedIn posts] ---------------------------------------------------------------------------- Qualified actions 14 (your target: 10) Cost for the period $900.00 Cost per qualified action $64.29 (your ceiling: $100.00) Qualified actions vs your target 2/2 weight 22 (derived) Attribution confidence 2/2 weight 20 Repeatability 2/2 weight 18 Cost per qualified action 2/2 weight 16 (derived) Cadence held 2/2 weight 12 Compounding surface 2/2 weight 12 Score 100/100 -> COMPOUNDING Trigger 1 attributable PASS Trigger 2 affordable PASS Gate inert - both triggers passed VERDICT COMPOUNDING Fix first nothing - every signal is at 2 ---------------------------------------------------------------------------- CHANNEL C-02 [Paid newsletter sponsorships] ---------------------------------------------------------------------------- Qualified actions 9 (your target: 8) Cost for the period $2400.00 Cost per qualified action $266.67 (your ceiling: $150.00) Qualified actions vs your target 2/2 weight 22 (derived) Attribution confidence 2/2 weight 20 Repeatability 2/2 weight 18 Cost per qualified action 1/2 weight 16 (derived) Cadence held 2/2 weight 12 Compounding surface 2/2 weight 12 Score 92/100 -> COMPOUNDING Trigger 1 attributable PASS Trigger 2 affordable FAIL GATE CAPPED one trigger failed: cannot read COMPOUNDING VERDICT TREADING WATER Fix first Cost per qualified action ---------------------------------------------------------------------------- CHANNEL C-03 [Podcast guesting] ---------------------------------------------------------------------------- Qualified actions 6 (your target: 5) Cost for the period $300.00 Cost per qualified action $50.00 (your ceiling: $120.00) Qualified actions vs your target 2/2 weight 22 (derived) Attribution confidence 0/2 weight 20 Repeatability 2/2 weight 18 Cost per qualified action 2/2 weight 16 (derived) Cadence held 2/2 weight 12 Compounding surface 2/2 weight 12 Score 80/100 -> COMPOUNDING Trigger 1 attributable FAIL Trigger 2 affordable PASS GATE CAPPED one trigger failed: cannot read COMPOUNDING VERDICT TREADING WATER Fix first Attribution confidence ---------------------------------------------------------------------------- CHANNEL C-04 [Cold outbound email] ---------------------------------------------------------------------------- Qualified actions 2 (your target: 6) Cost for the period $1600.00 Cost per qualified action $800.00 (your ceiling: $150.00) Qualified actions vs your target 1/2 weight 22 (derived) Attribution confidence 0/2 weight 20 Repeatability 2/2 weight 18 Cost per qualified action 1/2 weight 16 (derived) Cadence held 2/2 weight 12 Compounding surface 2/2 weight 12 Score 61/100 -> TREADING WATER Trigger 1 attributable FAIL Trigger 2 affordable FAIL GATE FLOORED both triggers failed: CUT IT regardless of score VERDICT CUT IT Fix first Attribution confidence ---------------------------------------------------------------------------- CHANNEL C-05 [Short-form video] ---------------------------------------------------------------------------- Qualified actions 0 (your target: 5) Cost for the period $700.00 Cost per qualified action undefined - no qualified actions Qualified actions vs your target 0/2 weight 22 (derived) Attribution confidence 1/2 weight 20 Repeatability 1/2 weight 18 Cost per qualified action 0/2 weight 16 (derived) Cadence held 2/2 weight 12 Compounding surface 1/2 weight 12 Score 37/100 -> CUT IT Trigger 1 attributable FAIL Trigger 2 affordable FAIL GATE FLOORED both triggers failed: CUT IT regardless of score VERDICT CUT IT Fix first Qualified actions vs your target ---------------------------------------------------------------------------- CHANNEL C-06 [SEO blog posts] ---------------------------------------------------------------------------- Qualified actions 4 (your target: 4) Cost for the period $380.00 Cost per qualified action $95.00 (your ceiling: $120.00) Qualified actions vs your target 2/2 weight 22 (derived) Attribution confidence 1/2 weight 20 Repeatability 1/2 weight 18 Cost per qualified action 2/2 weight 16 (derived) Cadence held 1/2 weight 12 Compounding surface 0/2 weight 12 Score 63/100 -> TREADING WATER Trigger 1 attributable PASS Trigger 2 affordable PASS Gate inert - both triggers passed VERDICT TREADING WATER Fix first Compounding surface ============================================================================ PORTFOLIO ============================================================================ Channels assayed 6 Compounding 1 Reading CUT IT 2 PORTFOLIO VERDICT ONE CHANNEL COMPOUNDS This rollup takes the BEST channel, not the worst. It is the only rollup in the catalog that does. You do not need ten channels that work; you need one, and the rest of this page tells you which nine are costing you the time to find it. Cut first C-04 - $1600.00 spent, 2 qualified actions Double first C-01 - $64.29 per qualified action ============================================================================
Six weighted signals, three verdicts, and one gate that needs both keys.
Qualified actions vs your target 22, attribution confidence 20, repeatability 18, cost per qualified action 16, cadence held 12, compounding surface 12. Weights sum to 100. COMPOUNDING at 75 or more, TREADING WATER from 50 to 74, CUT IT under 50. Two of the six are derived from your numbers, not marked — you cannot mark your way past them.
At least one qualified action, AND enough attribution confidence to say where it came from. A number you cannot trace is not evidence that a channel works. It is evidence that something happened.
At least one qualified action, AND a cost per qualified action at or under the ceiling you set. Not our ceiling. Yours — which is the only one that can be wrong in a way you are able to correct.
Both triggers pass and the gate is inert — the score band stands. One passes and the channel is capped at TREADING WATER: it has not earned another cycle, but it has not disproved itself either. Neither passes and it reads CUT IT, whatever the score says. The verifier exercises all 2,916 combinations of the six marks and four trigger states and proves no single-trigger path to COMPOUNDING exists.
Every other rollup RedHub ships takes the worst row, because one broken control breaks the set. This is the exception, and deliberately so. You do not need ten channels that work — you need one. So the portfolio verdict reads the best channel, and a separate line names the most expensive channel reading CUT IT, which is where the time is actually going.
It grades a channel from your numbers. It does not supply the numbers.
- A continuation decision: does this channel get another cycle?
- Deterministic and offline — same inputs, same verdict, every time.
- A named model you can argue with: weights, bands and both triggers are on this page.
- Dateless. Re-run it next quarter with next quarter’s numbers; nothing drifts on its own.
- Not an attribution model. It consumes a confidence mark you make; it does not compute attribution.
- Not connected to anything — no analytics, no ad account, no CRM. You bring the counts.
- Not benchmarked. There is no industry average anywhere in it, and it will not invent a multiplier.
- Not financial advice, and it scores no people.
Scope note: no verdict here guarantees a revenue, growth or acquisition outcome. A COMPOUNDING channel is a channel that met two conditions you set, in one period you measured. This is not financial advice.
Anyone paying for distribution in hours, dollars, or both.
- Solo founders running three or four channels and trusting none of them.
- Bootstrapped teams choosing what to stop doing next quarter.
- Agencies who need a defensible reason to end a channel, not a feeling.
- Anyone whose reporting is full of impressions and empty of decisions.
- Teams wanting an attribution model — that is a different product.
- Anyone who cannot yet count a qualified action. Define it first.
- Evaluating the person who ran the channel. This grades the channel.
- Anyone expecting a growth guarantee. There is none here.
The Assay consumes an attribution mark. These three earn you the right to make it.
The Assay asks how confident you are that an action came from a channel. This one grades whether the report behind that confidence reconciles.
ViewMethodology upstream of the decision: whether a result is even reportable yet. The Assay is what you do once it is.
ViewWhen the Assay tells you to double a channel, this is the treatment for the machine-readable half of it.
ViewThe mechanics, stated plainly.
Kill the channel on purpose.
Not by forgetting it.
One purchase, lifetime access, 12 months of updates. $99, once.
Deterministic and offline. It uses your numbers only, grades channels never people, and guarantees no revenue or growth outcome. Not financial advice.
Sold by RedHub AI LLC · Secured by Stripe · redhub.ai