Choosing channels

Channel choice is where marketing budgets are won and lost, and most of it is decided by familiarity rather than evidence.

Start with capture, then create

Saturate demand capture before funding demand creation. If people already searching for your category cannot find you, spending to make more people search is subsidising a competitor.

"Saturated" is measurable: impression share in paid, share of the winnable queries in organic. When those are near their ceiling and you still want growth, you are in creation territory whether you like it or not.

Channels have ceilings, and they arrive quietly

Every channel saturates. The signal is rising cost per outcome at flat volume — you are paying more to reach the same people. Teams usually respond by increasing budget, which accelerates the decline. The correct response is to open the next channel before the current one tops out, because a new channel takes months to work.

Pick channels from the audience, not the trend

The question is not "should we be on X" but "where does this specific buyer already spend attention, and can we be useful there". A channel where your ICP is absent is not cheap, it is worthless at any price.

Two practical filters: can you actually make the format well and repeatedly, and can you sustain it long enough to learn anything. A channel you can only half-serve will underperform one you can commit to.

Payback period decides how much risk you can take

A business that recovers acquisition cost in one month can fund slow channels. One that takes eighteen months cannot afford experiments that take a year to read. Channel strategy is downstream of cash, and this is the constraint most plans ignore.

Test small, but test properly

An undersized test tells you nothing and gets used as evidence anyway. A test needs enough budget, enough time to clear the learning period, and a decision rule agreed in advance: what result would make us stop, and what would make us scale.

What trips people up

• Adding a channel because a competitor did. You do not know their economics.
• Judging a creation channel on last-click. It will always lose that comparison.
• Spreading budget across six channels so that none reaches the threshold where it works.

You have got this when

You can say, for each channel you run, whether it is capture or creation, whether it is near its ceiling, and what would make you stop funding it.