Step 5 of 6 · Channel Selection
How to choose marketing channels that pay back.
Sooner or later in most leadership meetings, someone brings a new channel to the table: an agency proposing a LinkedIn pilot, a trade publication with a sponsorship package, or a board member who read that buyers now ask ChatGPT for recommendations. Each pitch sounds sensible, and it's hard to turn any of them down, because nobody in the room can say which of the channels you already pay for is bringing in customers. So the new one gets a small test budget, and it joins the Google Ads, the SEO retainer and the spring trade show on a list that only ever gets longer.
The channels that pay back are the ones where your IMPACT Prospect already pays attention, at a cost per customer below what that customer is worth to you. In ProfitPaths® Channel Selection comes after the Prospect Magnet and the conversion path are built, so a channel has a narrow job: carry a Prospect Magnet to your IMPACT Prospect and land them on a page that converts. This page shows how to choose those channels, and how to judge each one by the signed customers it brings in, as recorded in a source of truth you control.
In ProfitPaths channels come fifth, after four decisions most companies skip (the offering, the buyer, the Prospect Magnet and the page that converts), because a channel can only be judged against a specific buyer and what that buyer is worth.
Know your ceiling first
Decide what you can afford to pay for a customer before you compare channels.
Most channel decisions get made on cost per click or cost per lead, because those are the numbers the platforms show you first. They tell you the price of attention, and they say nothing about whether that attention ever turned into a customer, which is how companies end up able to say which channel produces clicks and unable to say which one produces revenue.
The number that should decide is the most you can afford to spend to win one IMPACT Prospect, and you worked out its inputs on the IMPACT Prospect page. Once you know it, a channel with an expensive click can turn out to be the best deal you have, and a cheap one can turn out to be a waste.
How much that should be depends on the business, so there's no single rule worth copying. It turns on how much of a customer's lifetime profit arrives in the first year, how long you can wait to earn back what you spent, how much cash you can tie up doing it and how fast you're trying to grow. A customer who stays for years and pays a little each year can justify a much higher cost to win than one who brings most of their profit in the first few months, because the second kind leaves nothing to earn the cost back from.
What the contractor can afford to pay for a property manager
Its customers stay about seven years and its cash flow can carry a one-year payback, so it chose to spend up to one year's gross profit, which leaves six more years of profit after the cost is earned back. The same choice gives an owner-occupied business, worth $4,200 in gross profit a year, a ceiling of $4,200, so a channel that wins customers at $5,000 each would lose money on owner-occupied businesses and be a bargain for property managers.
earned back inside the first yearWork backwards from the customer
Turn what a customer is worth into what you can pay for a lead and a click.
At the contractor's starting rates, before the fixes on the Prospect-to-Lead Conversion page, about 1% of visitors became leads and about 1 in 9 leads signed an agreement, so a $13,680 ceiling per customer became about $1,520 per lead and about $15.20 per visit. Put in your own numbers, then test what a channel charges you against the result.
Example These are the contractor's numbers for a property management firm, so replace them with yours.
At that price, one customer costs you about $9,000, which is $4,680 under your ceiling, so this channel can pay back if it reaches the right buyer.
How many months to count depends on how long customers stay, how quickly their profit arrives and how much cash you can tie up while you earn it back; the contractor uses 12 because its customers stay about seven years. Nothing you enter here is saved or sent anywhere.
Better conversion raises every one of those ceilings, because when more visitors become leads and more leads sign, each click is worth more, which is why the Prospect-to-Lead Conversion step comes before this one. After its fixes, 4% of the contractor's visitors became leads and about 1 in 8 of those leads signed, so the same $13,680 ceiling allows about $1,710 per lead and about $68 per visit.
Choosing channels
Choose channels by where your buyer pays attention, using three questions.
Many companies pick channels because competitors use them or because everyone at a conference was talking about them, which means someone else's buyer chose your channels for you. With the buyer and the Prospect Magnet defined, the question gets much narrower.
Can you single out your buyer there?
A channel full of people is useless if you can’t reach your IMPACT Prospect specifically.
Are they looking for you yet, or do you need to find them?
Search catches people who already know they have the problem, while LinkedIn, events and direct outreach reach people before they start looking, so each needs a different job and a different Prospect Magnet.
Can you win a customer there for less than your ceiling?
Compare the likely cost per customer with what that customer is worth, not with what other channels cost per click.
Plan your channels
Run each channel through the three questions, and the planner tells you which to fund.
The contractor reaches property managers who are already searching through Google, reaches the ones who aren't yet looking through LinkedIn and direct outreach to a named list of firms, builds trust through the local building owners' association, and leaves SEO until paid search has shown which terms turn into signed agreements.
Example This is the contractor's channel plan, so replace it with the channels you use or are considering.
Your channel plan
That's a focused plan, with 2 to fund properly and 2 to test.
Each channel gets enough budget to show a result, and you can add the next one once these are proven.
“Not sure” about cost per customer counts as a test, because the only way to find out is to fund the channel for a set period and trace what it produces. Nothing you enter here is saved or sent anywhere.
For high-ticket B2B offerings, we've seen a short list of the right accounts, each with its own landing page and sales follow-up, outperform broad advertising. In that setup ads still matter, but as support for the account-based effort rather than the whole plan.
If this sounds familiar
“We’re in every channel and none of them is clearly winning.”
This is one of the most common things leaders tell us, and it usually has the same cause. Channels get added one at a time, each for a reasonable-sounding reason, and the budget gets divided between them until none has enough to produce a clear result. Five channels funded at a fraction of what each needs produce five inconclusive reports, and the team ends up arguing over which one to cut with nothing solid to go on.
Choose deliberately, and fund properly
- Rank every channel by cost per customer against your ceiling, not by clicks or leads. Where you don't know the cost per customer, mark it unknown rather than guessing, because the unknowns are what the source of truth further down this page is for.
- Fund the two or three that fit your buyer best, at a level where a result would mean something, and park the rest for now rather than trimming everything evenly.
- Give each one a set period and change one thing at a time, because when several things change at once the results become impossible to read.
- Add a channel only when the current ones are proven, and add it with its own Prospect Magnet and its own budget instead of a slice of someone else's.
Paid search, then SEO
Use paid search to learn, then let SEO build on what it proves.
Paid search and SEO do different jobs, and the order you use them in matters. Paid search gives you answers in weeks about which searches bring in buyers, which messages they respond to and which pages convert them. SEO takes months to build, but once it works it keeps lowering your cost of winning customers.
So we usually run paid search first, learn which search terms produce revenue, and then put SEO investment behind those proven terms, so paid and organic reinforce each other instead of competing for the same budget. The same thinking applies to newer places your buyers look, like AI search: start from the questions your IMPACT Prospect asks, and build where you've proven those questions turn into customers.
Judge every channel by the customers it brings in
Each ad platform grades its own results, so their totals add up to more customers than you signed.
Every ad platform reports the results it takes credit for, and each one counts generously. A platform will often credit itself when someone saw an ad and bought weeks later, even if they'd have bought anyway (often called a view-through conversion), so two platforms can both claim the same sale while your analytics tool, which sees only part of the path, tells a third story.
None of them is lying, exactly, since each is reporting which sales it touched along the way, while the question you need answered is which sales you'd have lost without it.
What we've seen in the accounts we manage
If every form fill is a conversion, including job applicants and vendors pitching their services, the platform learns to find more people who fill out forms. In our own campaigns, counting low-value actions taught the platform to find people who looked and didn't buy.
In client accounts, optimizing on a qualified-lead event instead of every form fill changed who the platform found, and it can raise lead quality at the same cost.
In one B2B ecommerce account, the ad platform reported more revenue than could be traced to its ads anywhere else, which is why we use one independent source of truth, usually the CRM, and treat each platform's own numbers as a second opinion.
What the contractor's numbers showed
The contractor's platforms claimed more agreements than it signed, and the association behind two of them got no credit at all.
Back to the HVAC contractor from the earlier pages. After the fixes on the Prospect-to-Lead Conversion page, property managers are signing maintenance agreements from three places: Google search, LinkedIn ads aimed at property and facilities managers, and the local building owners' association. Step through the contractor's first full quarter with all three running and a CRM set up to record where each agreement came from.
- No. 1Google Ads
- No. 2Google Ads
- No. 3Google Ads
- No. 4Google Ads
- No. 5Google Ads
- No. 6Google Ads
- No. 7Google Ads
- No. 8Google AdsLinkedIn, claimed by both platforms
- No. 9Google AdsLinkedIn, claimed by both platforms
- No. 10Google AdsLinkedIn, claimed by both platforms
- No. 11Google AdsLinkedIn, claimed by both platforms
- No. 12Google AdsLinkedIn, claimed by both platforms
- No. 13Google AdsLinkedIn, claimed by both platforms
Each tile is one signed maintenance agreement. Tiles outlined in red are claimed by both platforms.
Claims in each platform’s own report
Together the two ad platforms claim 19 agreements, in a quarter when the contractor signed 13.
Six agreements are claimed by both platforms, because each one credits itself whenever the buyer also saw its ads. That includes the two that started at the building owners’ association, which has no dashboard to claim anything.
The platforms and the tracking tags were each right about what they could see, and each saw only part of it. Self-reported answers matter most for the channels nobody clicks, like referrals, events and associations, because every tracking tool undercounts them and they're often the first thing cut when budgets get tight.
Set up a source of truth you control
A source of truth you can trust takes less software than most companies expect.
You need a few things working together, in roughly the order listed here, and if you already have a CRM and an ad account you have the start of the first two. What matters is the connection between them, so that a signed customer in the CRM can be traced back to the campaign, the call or the conversation that started it.
Not sure your tracking can be trusted?
Request a free audit of the campaigns you’re running now. We’ll review your ads and the tracking behind them, and show you where spend is leaking and what we’d change first.
Free, with a real person from the 5K team. Takes about a minute to request.
What you need, in order
- A CRM as the hub. Every lead and every sale lives here, with the channel it came from and, eventually, the revenue and margin it produced.
- Source tracking on every link and form. Tagged links (often called UTM tags) and hidden form fields record which campaign brought each visitor, so the CRM knows where a lead came from without anyone guessing.
- Call tracking, if buyers phone you, because calls are often your best leads and the easiest to lose track of.
- A “How did you hear about us?” question on every form and every intake call. Buyers will tell you about the colleague, the podcast or the trade show that no tracking tool can see, and although people don’t always remember accurately, the answers fill gaps nothing else does.
- Qualified events sent back to the platforms, so the ad platforms count qualified leads and sales instead of every form fill, and learn to find more of the right buyers.
- A monthly review of cost per qualified opportunity and cost per customer by channel, beside what those customers are worth.
Judge each channel by what its customers are worth
The cheapest channel per customer can be the one bringing in the least valuable customers.
Once the CRM knows where each agreement came from, put each channel's cost per customer beside what those customers are worth. Cost per customer is spend divided by customers won, and lifetime gross profit per customer is annual revenue times gross margin times the years a customer stays, minus what it cost to win them, the same arithmetic as the IMPACT Prospect page.
Example These are the contractor's three channels, using $12,000 per building from the IMPACT Offering page and the 38% margin and seven years from the IMPACT Prospect page, so replace them with yours.
Ranked by cost per customer
cheapest first, the way most budgets get cut
- Google search$2,000
- Building owners' association$3,000
- LinkedIn$4,500
Ranked by what each customer is worth
lifetime gross profit, after the cost of winning them
- LinkedIn$155,100
- Building owners' association$92,760
- Google search$61,840
Google search is the cheapest way to win a customer, but LinkedIn brings in the customers worth the most, so a budget cut on cost per customer alone would land on the wrong channel.
Use customers won from your CRM, not conversions from the ad platforms, and the margin finance uses. Nothing you enter here is saved or sent anywhere.
If the contractor ranked its channels on cost per agreement, LinkedIn would be the first to go, since it's the most expensive way the company wins a customer. But LinkedIn reaches the larger property management firms, which rarely search for a contractor and sign agreements covering more buildings, so each of its customers is worth more than twice what a Google customer is worth over the life of the agreement. A company that judged its channels on cost per lead, or even cost per sale, would have cut its most profitable one.
If nobody at your company can say what it costs to win a customer from each channel, you need three numbers to start: spend by channel from the platforms, customers won by channel from the CRM, and the annual value and margin of those customers from finance.
Tell the platforms who bought
Send qualified leads and signed sales back to the ad platforms, so they bid for buyers.
Once the CRM records which leads became real opportunities and which signed, the same record can go back to Google Ads, LinkedIn and Meta, matched to the ad click or the person, which is usually called offline conversion tracking. Without it, a platform counts the student who downloaded a guide the same as the property manager who signed, and keeps bidding for more of whatever it counted. Google's Smart Bidding optimizes for whichever conversion types you include in the "Conversions" column,1 so that list is the instruction your bidding follows, and what offline conversion tracking is walks through the setup on each platform.
Send the earliest stage sales would agree is a real opportunity, and the signed deal as well, because the first gives the platform enough examples to learn from and the second tells it which of those were worth the most. Google recommends bidding on a single stage of your lead-to-sale funnel, with a relatively short conversion delay and at least 15 conversions a month,2 which is why, at the contractor's volume, we'd send the booked site visit (the stage between a lead and a signed agreement in the funnel on the Prospect-to-Lead Conversion page) alongside each signed agreement, and the best conversion actions to set in Google Ads covers choosing them. Attach each agreement's value, too, so the platform can tell a small property management firm from a large one; setting conversion value in Google Ads to profit works that out stage by stage.
LinkedIn and Meta accept the same CRM stages through what each calls its Conversions API,3 though Meta's Conversion Leads goal currently works only with Facebook and Instagram lead forms, and its guide calls for at least 200 leads a month.4 Imported conversions are still credited by each platform's own attribution, so this teaches the platforms whom to find without settling which of them earned the sale, which is still the CRM's job.
Credit isn't the same as cause
Credit for a sale shows you the path, but it can't tell you what caused the sale.
Attribution assigns credit for a sale to the touchpoints that came before it, which is useful because it shows the path buyers took. It can't show you whether a channel caused the sale, though, since some of the people who clicked your branded search ad would have found you anyway.
The extra results a channel caused are called its incrementality, and measuring them means comparing people who saw your marketing with similar people who didn't, for example by pausing a channel in some regions and comparing results. Most companies should start with a clean source of truth and add tests like these once spend is large and spread across many channels, because that's when they're worth what they cost.
What doesn't work
Five ways companies waste money choosing and judging channels.
- 1Chasing the newest channel.
A channel being popular says nothing about whether your buyer uses it to make decisions in your category.
- 2Spreading the budget across every channel.
Each channel gets too little to show a result, so you never find out which one would have worked.
- 3Letting each platform grade itself.
If every platform's report is taken at face value, the totals won't add up and the budget goes to whoever claims the most.
- 4Counting every form fill as a conversion.
The platforms will dutifully find you more of whatever you count, including the people you don't want.
- 5Judging every channel on cost per lead.
As the contractor's numbers show, the channel with the most expensive leads can be bringing in the most valuable customers.
At every size
A small company usually has too few channels, and a large one too many.
If you run a smaller business
Pick one channel where your buyer is clearly looking, usually search, and one Prospect Magnet that solves a problem they already have, and get those working before adding anything. A small budget spread thin is the most common reason we hear small business owners say marketing "doesn't work," when in fact no channel ever had enough to prove itself.
To know whether it's working, start with three things: a simple CRM, or even a well-kept spreadsheet, that records every lead and where it came from; a "How did you hear about us?" question on every form and every call; and a monthly look at what each channel cost beside the sales it produced. That takes an afternoon to set up, and it answers the question most owners can't: which channel's customers paid back.
If you lead marketing at a larger company
You probably have too many channels rather than too few, and each one has an owner who believes in it. Rank every channel by cost per customer against what that customer is worth, fund the top two or three properly, and bring the list to leadership as a decision about where the next marketing dollar earns the most, so the conversation is about the buyer instead of about whose channel survives.
The harder part is getting finance, sales and marketing to agree on one set of numbers. Agree on the definitions first (what counts as a qualified opportunity, when revenue is credited, how margin is calculated), connect closed revenue in the CRM back to campaigns, and report cost per customer against customer value by channel. That's the report a CFO and board will trust, and the one that lets you defend the budget when someone asks why the expensive channel is still running.
Questions we hear about this step
- Are Google Ads Worth It?Google Ads are worth it when winning a customer through them costs less than that customer is worth to you in gross profit, so judge them on cost per customer rather than clicks, cost per click or Google’s conversion count.Read the answer →
- Are LinkedIn Ads Worth It?LinkedIn Ads are worth it when they reach high-value buyers you can’t reach well through search, at a cost per customer below what each customer is worth, so judge them on cost and value per customer, not cost per click or per lead.Read the answer →
- How Much Does SEO Cost?SEO usually costs a monthly retainer that rises with scope, competition and who does the work, so judge an SEO quote by the customers it should win on searches you already know convert, rather than by the size of the fee.Read the answer →
- How Do You Calculate Break-Even ROAS?To calculate break-even ROAS, divide 1 by your profit margin as a decimal, using gross margin minus any other costs that rise with each sale, and any ROAS below that number loses money on the first sale the ads bring in.Read the answer →
- How Do You Calculate Customer Acquisition Cost?To calculate customer acquisition cost, divide everything you spent to win new customers in a period, including ads, agency fees, tools and sales time, by the new customers your CRM shows you won in that same period.Read the answer →
- What Is a Good ROAS?A good ROAS is one comfortably above your break-even ROAS, which is one divided by your margin after the costs that come with each sale, so it depends on your margins and repeat buying rather than an industry average.Read the answer →
- How Do You Measure Digital Marketing ROI?Measure digital marketing ROI as the gross profit from the customers marketing won, minus what it cost to win them, divided by that cost, tracked by channel in a CRM you control rather than the ad platforms' own reports.Read the answer →
More on choosing and judging channels
Guides that go deeper on paid media, SEO, AI search and attribution.
They're in progress and will be linked here as they're published.
Paid Media Strategy: Scaling Spend Without Losing Efficiency
For planning paid search and paid social around what a customer is worth.
SEO Strategy for Growth-Stage Companies
For when you're ready to build organic search on proven terms.
AI Search Visibility: Generative Engine Optimization
For understanding how buyers find companies through AI assistants, and what to do about it.
Marketing Attribution: Finding What Drives Revenue
For choosing an attribution approach that fits your company.
Measuring Marketing Effectiveness Beyond Attribution
For when you need to know what your marketing caused, as well as what it touched.
Marketing ROI: Proving Which Spend Pays Back
For calculating return in terms your CFO will accept.
- Google Ads Strategy: structuring Google Ads around qualified buyers instead of clicks. coming soon
- LinkedIn Ads Strategy for B2B: reaching buyers by job title before they start searching. coming soon
- Performance Max: what Google’s automated campaign type does well, and where it needs guardrails. coming soon
- Retargeting Strategy: staying in front of buyers who visited but weren’t ready. coming soon
- Account-Based Advertising: using ads to support outreach to a named list of accounts. coming soon
- SEO vs. SEM (and PPC): how organic and paid search differ, and how they work together. coming soon
- B2B SEO: SEO for long sales cycles and small, specific audiences. coming soon
- Content Marketing Strategy: content built around your IMPACT Prospect’s questions. coming soon
- How Much Does SEO Cost?: what SEO costs and how to judge whether it’s paying back. coming soon
- ChatGPT SEO: Getting Recommended by ChatGPT: showing up when buyers ask AI assistants for recommendations. coming soon
- Trade Show Marketing: getting a return from events beyond the booth itself. coming soon
- Channel Marketing: selling through distributors and partners. coming soon
- Self-Reported Attribution (“How Did You Hear About Us?”): setting up the question and making sense of the answers. coming soon
- Customer Acquisition Cost (CAC): calculating the full cost of winning a customer. coming soon
- Conversion Tracking for Lead Generation: tracking qualified leads and sales instead of every form fill. coming soon
- Call Tracking: connecting phone leads to the marketing that produced them. coming soon
- Multi-Touch Attribution: what it shows, what it misses, and when it's worth it. coming soon
- Marketing Mix Modeling: measuring the effect of each channel across a large budget. coming soon
- ROAS vs. ROI: why a strong return on ad spend can still lose money. coming soon
Sources cited
- Google Ads HelpOffline conversion imports FAQsOctober 12, 2026
- Google Ads HelpConversion Values Best PracticesOctober 12, 2026
- LinkedIn Marketing Solutions HelpLinkedIn Conversions APIOctober 12, 2026
- Meta for DevelopersConversions API for CRM integrationDecember 17, 2025