Marketing Budget Allocation: The Best Split by Channel

For a US service business between $1M and $10M in revenue, the allocation that works is: spend 6–10% of revenue on marketing total, then split it 50% demand capture / 30% owned assets / 15% demand creation / 5% measurement. Demand capture is Google Search, Local Services Ads, Google Business Profile, and SEO — people already looking for what you sell. Owned assets are your site, CRM, review engine, and automation. Demand creation is paid social, retargeting, and direct mail. Before you split anything, fund each channel to its minimum viable spend — roughly $1,500/month per metro for Google Search. Half-funding four channels beats nothing but loses to fully funding two.

That last rule is where most budgets break, and almost nobody writes about it.

The Minimum Viable Spend Rule (Fund the Floor Before You Split)

Every paid channel has a spend floor below which the money does essentially nothing. It isn't a linear curve — it's a step function. You get near-zero return until you cross a threshold, then returns begin.

Real floors for US service businesses, per metro of roughly 500k–1M people:

Add those floors: $6,600/month, or roughly $79,000/year. If your total marketing budget is $40,000/year, you cannot run five channels. You can run two, properly. The correct allocation for a $40k budget is $1,800/month to Google Search plus LSAs and $1,500/month to your website, CRM, and review system — not $667/month across five channels.

A budget spread across five channels at 40% of each channel's viable floor doesn't produce 40% of the results. It produces close to zero, and it produces zero *slowly*, which is worse — you burn 6 months before you know.

The Four Buckets, and What Actually Goes In Them

Demand capture — 50%. Search ads, LSAs, GBP optimization, SEO, directory presence. This is the highest-intent traffic that exists. A person typing "emergency AC repair near me" at 2pm in July has a purchase intent measured in hours. Expect 8–15% lead-to-booked-job rates here versus 1–3% from cold social.

Owned assets — 30%. Your site, your CRM, your review generation, your follow-up automation. This bucket is unglamorous and gets cut first, which is why so many companies have a $9,000/month ad spend pointed at a website that converts at 1.8% when the category median is closer to 4.5%. Doubling conversion rate is mathematically identical to doubling ad budget, and it costs far less.

Demand creation — 15%. Paid social, retargeting, direct mail, sponsorships. Slower, more expensive per lead, but it's the only thing that grows the pool of people who search for you later.

Measurement — 5%. Call tracking, attribution, dashboards, and someone who reads them. On a $200,000 annual budget, that's $10,000 — trivial insurance on the other $190,000.

Allocate by Payback Window, Not by Channel

The framework nobody publishes: sort spend by how long the cash takes to come back, then match that to your actual cash position.

If you have under 60 days of operating cash, put 80% into the 0–30 day bucket and accept slower growth. If you have 6+ months of runway, a 40/30/20/10 spread across those windows compounds far harder. A roofing company with $80,000 in the bank and a 45-day receivables cycle has no business putting 30% of budget into a 9-month SEO play — and plenty of agencies will happily sell them one anyway. Model it against your real numbers with the ROI calculator before you commit a dollar.

Reallocate on Your Sales Cycle, Not the Calendar

Monthly budget reviews are the default and they're wrong for most service businesses. Review on a multiple of your sales cycle.

Judging a 120-day-cycle channel at day 30 means you'll kill it right before it pays. This single error — impatience mismatched to cycle length — wastes more budget than bad channel selection does.

When This Framework Is Wrong, and Who Should Not Spend This Money

The uncomfortable part, stated plainly.

Do not increase marketing spend if you can't service more work. If you're booked 4 weeks out and turning away jobs, more leads mean longer hold times, worse reviews, and a lower close rate on the leads you already paid for. The correct spend is *hiring*. A crew that adds $400,000 in annual capacity returns more than $60,000 in ads pointed at a saturated schedule.

Do not hire an agency under roughly $3,000/month in ad spend. At a typical 15–20% management fee, a $2,000 spend generates $300–400 in fees — below what any competent team can service. You'll get an offshore junior and a template. Run it yourself, or use a flat-fee setup. Our own pricing has a floor for the same reason, and if you're under it, we'll tell you.

Do not allocate to any channel before fixing speed-to-lead. Companies responding to inbound leads in under 5 minutes qualify them at dramatically higher rates than those responding in 30+ minutes. If your average response time is 4 hours, every dollar of the allocation above is discounted by more than half. Fix the phone before you fund the ads.

This framework fails for businesses under ~$500,000 in revenue. At 8% of revenue, that's $40,000/year — $3,300/month, which covers roughly one channel plus basic assets. The four-bucket split is not meaningful at that size. Do one thing: Google Business Profile, reviews, and LSAs. Ignore the rest for 18 months.

It also fails for genuinely new categories. If nobody searches for what you sell, demand capture at 50% is allocating half your budget to a search volume that doesn't exist. Invert it: 60% demand creation, 20% capture, 20% assets.

Failure modes to watch for:

A Worked Example

A $3.2M residential electrical company, 40-day sales cycle, 90 days of cash on hand.

Total budget at 8%: $256,000/year ($21,300/month).

Every channel clears its floor. At a $310 cost per booked job and an $1,850 average ticket, $256,000 buys roughly 825 jobs and $1.53M in revenue — a 6:1 gross return before delivery cost. Reviewed quarterly, matched to a 40-day cycle.

The numbers will differ for your business. The structure — floors first, buckets second, payback window third, review cadence matched to cycle — holds across trades, legal, medical, and professional services. If you want the version built on your actual close rates and ticket sizes rather than benchmarks, start with a free marketing audit and bring 12 months of revenue data.