Your Channel Mix Is Probably Backwards. Here's How to Fix It.

T

The founder of a 22-person commercial cleaning company walked us through his marketing budget last fall: $18K a month into Google Ads, $4K into LinkedIn, $800 into email. He'd been running that split for three years. When we asked why LinkedIn got $4K, he said, "It's where our clients are." When we asked why Google got $18K, he paused. "It's where we started."

That pause is the whole problem. Most channel allocations aren't decisions. They're deposits. Money flows to wherever it flowed last quarter, compounded over time by whoever managed the account before you. The channel that launched your first campaign in 2021 is still receiving the majority of your budget in 2025, not because it's your best channel, but because inertia is a powerful force and audits are uncomfortable.

The cleaning company's situation was not unusual. Their ICP was facility managers and procurement leads at mid-size office buildings. LinkedIn, at $4K, was reaching them directly. Google was pulling in mom-and-pop residential leads who couldn't afford a commercial contract. They were spending 22x more on their worst channel.

Channel mix rarely gets rebuilt because nobody's job is to challenge it. The paid media buyer optimizes within whatever budget they're handed. The SEO team builds content for whatever keywords they're chasing. The budget allocation itself sits in a no-man's land between the founder and the marketing lead, and neither one wants to own the conversation that says: we've been wrong about this for two years.

Trendy platforms make this worse. TikTok Ads became a mandatory line item for businesses whose buyers haven't opened TikTok since 2022. Meta Advantage+ gets budget because the platform says it performs. Performance Max gets budget because Google says it covers everything. These tools are not wrong. They're wrong for certain offers at certain price points sold to certain buyers. That distinction requires a strategic frame the platform will never provide.

The result is what we see repeatedly: a $6M professional services firm with 60% of spend in a channel that produces volume but the wrong volume. High CTR, low close rate, frustrated sales team. The dashboard looks fine. The P&L tells a different story.

A real channel audit runs in a specific order. Most teams run it backwards: they start with platform data, look at what's performing, and call that strategy. That's optimization, not strategy. Here's the sequence that produces different answers.

• 01Step 1: Map ICP decision behavior. Where does your buyer actually research? What do they read, watch, search, or ask peers before committing budget? Not where they scroll. Where they decide.
• 02Step 2: Score offer fit by channel. A $120K annual contract is not a Performance Max offer. A $180 monthly subscription might be. Match conversion mechanism to purchase psychology.
• 03Step 3: Pull your actual channel attribution data. Not last-click. Assisted conversions, time to close by channel, deal size by acquisition source. GA4 and your CRM together, not separately.
• 04Step 4: Identify the gap. The channel your ICP uses to make decisions is almost certainly underinvested. The channel that generates volume is almost certainly over-invested.
• 05Step 5: Apply the 70/20/10 split. 70% to the primary proven channel, 20% to the secondary channel you are actively developing, 10% to a deliberate single-variable test.

The 70/20/10 framework is not new. It comes out of portfolio investing and Google's own internal budget philosophy. What's new is applying it to channel mix with ICP behavior as the anchor, not historical spend. That shift changes every number in the allocation.

A high-ticket, high-trust service purchase does not behave like a $49 SaaS subscription. It requires repetition, credibility, and the right context. Putting a $250K consulting engagement into Performance Max is the channel strategy equivalent of selling a custom BMW at a Costco demo station. The audience isn't wrong. The context destroys the conversion.

Offer fit has three dimensions worth scoring before any budget decision:

• Trust threshold. How much credibility does your buyer need before they'll act? High-trust offers need channels that carry editorial weight: LinkedIn thought leadership, long-form content, referral networks, podcast placements.
• Decision timeline. A 90-day sales cycle needs channels that work across multiple touches, not channels optimized for immediate conversion. Email and retargeting earn their budget here. Cold display does not.
• Positioning specificity. If your offer is highly differentiated, you need a channel that can carry nuance. A 15-second pre-roll ad cannot explain why your methodology is different from the six competitors a buyer is also considering.

When you score each channel against these three dimensions and cross-reference with your ICP behavior map, the misallocations become obvious. Not directionally obvious. Specifically obvious, as in: you'll see the exact dollar amount sitting in the wrong place.

“Your biggest spend is probably in your second-best channel. And your best channel is starved.”

GA4 data-driven attribution is better than last-click. It's still not the full picture. The full picture requires pulling deal size and close rate by acquisition source out of your CRM, cross-referenced against GA4's assisted conversion report. Most teams have never done this because it requires more than one platform login and about four hours of clean thinking.

What this analysis typically shows: the channel generating the most leads is not the channel generating the most closed revenue. There is almost always one channel, usually email or organic search or a specific referral source, that closes at 2x to 3x the rate of the volume channel. That channel is almost always underinvested because its lead volume looks small in the dashboard.

A 14-person IT services firm we audited had three years of HubSpot data. Paid search was generating 68% of their inbound leads. Their referral channel, which they were spending almost nothing to cultivate, was closing at 41% versus paid search's 9%. When we ran total closed revenue by source, referrals had produced more revenue in three years than paid search, on a fraction of the investment. The paid search budget was not wrong. The proportion was.

This is not a rare finding. It is the most common finding. Strategy built from ICP-first thinking surfaces this pattern consistently, because it starts with behavior and offer fit instead of platform metrics.

The audit itself takes two weeks if you move with urgency. The reallocation takes one budget cycle to implement cleanly. Here's what the 90-day window looks like in practice for a $5M to $15M service business.

• Week 1 to 2: Pull HubSpot or Salesforce deal data and tag every closed deal with its first-touch and last-touch acquisition source. Map the delta between lead volume and closed revenue by channel.
• Week 3: Run the ICP behavior interview. Five to eight calls with your best clients. Ask specifically: what did you read, watch, or search in the 90 days before you contacted us? The channel they name is your primary channel.
• Week 4: Score each active channel against offer fit dimensions. This is a spreadsheet exercise, not a philosophical debate. Each channel gets a score. The lowest-scoring channel with the highest spend is the budget that moves.
• Month 2: Shift 15% to 25% of spend from the over-invested channel to the primary channel identified in the ICP interview. Do not make this shift all at once. Ramp the receiving channel to avoid efficiency cliffs.
• Month 3: Run the 90-day comparison. Closed revenue per dollar, not leads per dollar. That is the only metric that validates the reallocation.

This is not a framework you can delegate to a junior account manager or run through an AI prompt. It requires someone who can hold the ICP behavior data, the offer fit logic, and the attribution output in the same frame and make a call. That is the judgment part. The AI can accelerate the data pull, the tagging, and the pattern recognition. It cannot supply the conviction.

For teams that want a faster read on whether their current mix is defensible, the DIY-or-Agency Quiz surfaces the structural gaps quickly. It's not a substitute for the full audit, but it's a useful pressure test before you commit to a quarter of spend on a channel you haven't challenged.

The businesses that win the next 18 months will not be the ones that added the most channels. The channel-expansion playbook is a race to diffusion. Spread budget across six platforms, run thin in all of them, and wonder why CAC is climbing. That's where most $5M to $20M operators are right now.

The operators who outperform will be the ones who committed depth to the one or two channels where their ICP actually makes decisions. Depth means enough budget to dominate the format, enough frequency to build recognition, and enough patience to let the channel compound. Spotify didn't win by being on every platform. It won by going all-in on the one behavior that mattered: on-demand streaming. Channel strategy is the same bet. Pick the right behavior. Go deep. Stop spreading thin.

If you're a marketing director at a service business and you cannot answer "why is our budget split this way" with a documented ICP-behavior rationale, the split is wrong. It may be accidentally right, but you don't know. That's a strategy problem, not a media problem. Fixing it starts at the strategy layer, before you change a single campaign setting.