The channel-switch test

Three weeks into outreach, the thought arrives on schedule: this channel is not working, maybe I should try something else. That thought is sometimes wisdom and usually panic, and the difference between the two is a test you can run on your lead sheet. This lesson is the test.

The minimum honest sample

You cannot judge a channel from ten attempts. At placeholder rates, ten outreach contacts is a fifth of one customer, so a dry run of ten tells you nothing except that you have not tried yet. My line, and I will defend it: do not judge any channel until you have logged at least fifty attempts or run three consistent weekly cycles, whichever comes later. Fifty attempts at the placeholder outreach rates is roughly one customer, which means a fifty-attempt sample that produced one customer is the channel working exactly as billed. Most people who quit a channel quit inside the sample size where quitting means nothing.

Rooms and communities get a longer runway than fifty attempts, because their currency is accumulated trust rather than message volume. A community you have shown up in usefully for six weeks is still early. Judge rooms at the season mark, not the month mark.

Fix before you switch

When a channel underperforms after an honest sample, the channel is the last thing to blame. Check the three usual suspects, in order.

Volume. Did you actually hit the quota every week, or did the sheet record bursts and gaps? Intermittent outreach produces intermittent results, and the fix is the calendar, not the channel.

Follow-up. Are unanswered contacts getting their two or three nudges, or did every lead get exactly one message and a funeral? The follow-up process is where close rates move most. One pressure washing operator in our research corpus reports closing roughly half of estimates on the spot and, with a disciplined follow-up rhythm, closing 72 to 73 percent overall, and reports that a single inbound lead can cost up to one hundred dollars, which is why they exhaust every follow-up before spending another dollar on the next lead. Your numbers will differ. The ordering of effort will not: follow-ups are cheaper than new attempts, always.

Specificity. Could your message have been sent to anyone? Then it worked on no one. Rewrite before you re-channel. A generic message moved to a new channel is the same message, dying somewhere else.

The comparison

With an honest sample on at least two channels, pull the counts from your weekly tallies and put them side by side. Two numbers per channel: customers per ten attempts, and minutes per attempt. The first says what the channel yields. The second says what it costs in the only currency a solo operator has. A channel that produces a customer per ten attempts but eats your Tuesdays may still lose to one that produces half as much on autopilot minutes.

| Channel | Customers per ten attempts | Minutes per attempt | Verdict | |---|---|---|---| | Warm network | about 1 | low | keep feeding | | Direct outreach | a fifth of one | medium | fix follow-up, then judge | | Community room | too early | low | revisit at the season mark |

Fill yours from the sheet. The table above is the shape of the decision, not the numbers.

The kill rule and its two errors

If, after the honest sample and the three fixes, a channel still produces nothing, kill it without ceremony and move its weekly time to the channel with the best customers-per-minute. The gig guide's graduation map runs the same move with platforms: the apps are the on-ramp, and the killed time flows to the direct business on the far bank. Two errors guard the edges. The false negative is killing a channel during normal variance, one quiet week inside an otherwise working channel, which the minimum sample exists to prevent. The false positive is the opposite: one lucky hit, a cousin who bought, a post that caught an algorithm, and you conclude a dead channel is alive and feed it for a year. One customer is a data point. A rate is a pattern, and patterns are what the sheet holds.

Run the whole test monthly, on the counting step of your weekly hour, not daily and not on feelings. Monthly is often enough for the numbers to mean something and rare enough to let consistency work.