Choosing marketing channels is mostly an exercise in refusal. Every channel works for somebody, few work for you, and running six badly beats running two well in exactly no cases.
How marketing channels differ from each other
The useful distinction is not paid against free. It is whether the channel captures demand or creates it.
Search, marketplaces and directories capture people already looking. Social, display, video and outdoor create interest in people who were not. Email, referral and reputation work on people who already know you. Consequently each answers a different problem, and a business with the wrong problem in mind picks the wrong channel.
The paid ones share a single test, and paying only when something happens is the plainest description of it.
Start with where your customers already are
Not where you enjoy spending time, and not where a competitor is loudest.
- What did your last twenty customers use to find you?
- Where do people in your trade ask each other for recommendations?
- Does anybody actually search for what you sell?
- Which channel produced your best customer rather than your cheapest?
Those four questions eliminate most options in an afternoon. Meanwhile the enquiry form question asking how somebody heard about you costs nothing and answers the first one permanently.
How many marketing channels to run
Fewer than you want. Two done properly is a reasonable target for a small business, three for a larger one.
The reason is arithmetic rather than principle. Each channel needs enough budget to produce results you can read, enough time to build and enough attention to improve. Split a modest budget four ways and every channel sits below the threshold where anything is learnable.
Add a channel when an existing one is genuinely maxed out, meaning more spending stops producing proportionate results. That is a real signal rather than a feeling.
The order that usually works
Capture before create, and owned before rented.
Cover the demand that already exists first, since it pays soonest. Then build something you own, such as an email list or a reputation, because it keeps working when budgets stop. Only then spend on creating demand, which is slower and harder to measure.
Reversing that order is possible and it takes longer and costs more, which most businesses discover about eight months in.
Judging marketing channels fairly
Channels are not comparable on a single number, and forcing them to be produces bad decisions.
A search click and a social click cost differently because they carry different intent. An email costs almost nothing and only reaches people you already have. Referral produces the best customers and the least volume. So compare each against what it is for, and against its own previous performance, rather than against each other.
Telling a poor channel from a neglected one
This is the distinction that saves the most money, and almost nobody makes it.
A poor channel has been given a fair budget, a decent landing page, reasonable creative and enough time, and still produces nothing. A neglected one was set up in an afternoon, pointed at a homepage, and never revisited.
Most channels declared dead were never actually tried. Before dropping one, check whether it received the same care as the channel you like.
Attribution across marketing channels
People use several channels before buying, and every platform claims the sale.
Expect the reported totals to exceed reality. Pick one source of truth for decisions, use the others as explanation, and pay attention to the blunt questions: did total enquiries rise, did branded search rise, did revenue move.
Where a channel creates demand that another captures, the credit lands in the wrong place. That is the single biggest reason good channels get switched off.
Reviewing the mix
Once a quarter is enough, and the review should be short.
Which channel produced the customers worth having? Which absorbed effort without result? What changed in the business that makes a channel more or less relevant? Then make one change rather than five, since changing everything at once means learning nothing.
Book the review in advance. A quarterly slot in the calendar is the difference between a mix you chose and a mix you drifted into.
When to drop one of your marketing channels
When it has had a fair run and the cost of a customer stays above what one is worth. Or when the audience genuinely moved, which happens more often than anybody plans for.
Dropping well means stopping properly rather than letting it decay. Turn it off, note what you learned, and move the budget somewhere it can reach the threshold that makes it readable.
Keep the note. Marketing channels come back into fashion, audiences move, and a paragraph explaining why you stopped in 2024 saves the same experiment being run again from scratch.
A workable starting mix
For most small businesses: one channel capturing existing demand, plus one owned channel building something that lasts.
That usually means search of some kind, paid or organic, alongside email or reputation. It is unglamorous, it fits the attention a small business actually has, and it beats a plan naming six marketing channels that nobody has time to run.
Write the two down and give each an owner and a budget. Channels fail more often from nobody being responsible than from being the wrong choice.
Frequently asked questions
How do marketing channels really differ?
By whether they capture demand or create it. Search and directories reach people already looking. Social, display and video create interest in people who were not. Email and referral work on people who already know you.
How many should I run at once?
Two properly for a small business, three for a larger one. Each needs enough budget to produce readable results, enough time to build and enough attention to improve. A modest budget split four ways teaches you nothing.
Which should I start with?
Capture before create, and owned before rented. Cover existing demand first because it pays soonest, then build something you own such as a list or a reputation, then spend on creating demand.
Can I compare channels on cost per lead?
Not fairly. Different channels carry different intent, so their costs and conversion rates are not comparable. Judge each against what it is for and against its own previous performance rather than against the others.
How do I know a channel is genuinely poor?
It has had a fair budget, a decent landing page, reasonable creative and enough time, and still produced nothing. Most channels declared dead were set up in an afternoon and never revisited, which is neglect rather than failure.
Why do my channel numbers add up to more than my sales?
Because people use several before buying and every platform claims the sale. Pick one source of truth for decisions, and watch the blunt measures: total enquiries, branded search and revenue.
When should I add a channel?
When an existing one is genuinely maxed out, meaning further spending no longer produces proportionate results. That is a measurable signal rather than a feeling that you should be doing more things.
What is a sensible starting mix?
One channel capturing demand that already exists, plus one you own. For most small businesses that means search of some kind alongside email or reputation. It is dull, and it fits the attention a small business actually has.