How to allocate a marketing budget across channels

Most contractor marketing budgets are set by last year's number plus a guess. A better process has three steps. First, size the total: the Gartner 2024 CMO Spend Survey puts the average marketing budget at 7.7% of company revenue, with B2B nearer 8.4% and B2C 5.7%. Second, split it between brand building and demand capture, anchored to Binet and Field's 60/40 finding. Third, divide spend across paid, SEO, and CRO by marginal return, not by habit. Nielsen found 50% of media plans are underfunded by a median of 50%, leaving real ROI on the table. The fix is measurement, not gut feel.

Set the total from revenue, not last year

The number starts as a percentage of revenue, then bends to your margins and growth plans. The Gartner 2024 CMO Spend Survey puts the all-industry average at 7.7% of company revenue, down from 9.1% in 2023. Splits run higher for B2B at 8.4% and lower for B2C at 5.7%. The U.S. Small Business Administration recommends 7 to 8% for firms under $5 million in revenue, but only if margins sit in the 10 to 12% range. A remodeler on 5% margins cannot spend like one on 15% and stay solvent.

Then adjust for ambition. The common rule of thumb is roughly 5 to 7% of revenue to hold your position and 8 to 12% or more to grow share. Early-stage companies in their first two years often need 12 to 20% to build any awareness at all. Treat the percentage as a starting anchor, not a law. The right total is the one your unit economics can carry while still funding the channels that compound.

Split brand and performance before you split channels

Before arguing over Google versus Meta, settle the bigger split: how much builds future demand versus how much captures demand that already exists. Binet and Field's analysis of 996 IPA case studies across 700 brands found that, on average, a 60/40 split of brand building to sales activation maximizes long-term effectiveness. The ratio is not fixed. New entrants skew toward brand, sometimes 70/30, while established brands with strong recall can run closer to 40/60. The point is that performance marketing harvests demand it did not create, and starving the brand side eventually starves the funnel.

Nielsen's research underlines the cost of getting the size wrong. Its analysis found 50% of media plans are underinvested by a median of 50%, and that reaching the optimal level could lift ROI by about 50%. Underspending was heaviest in upper-funnel channels: digital video budgets fell short by 66% and digital display by 60%. Brand work is hard to attribute and easy to cut, which is exactly why it gets cut first and why disciplined teams protect it.

Where the brand-versus-performance ratio should shift:

• New or unknown brand with low awareness: lean to brand, around 70/30
• Established brand with strong recall: lean to performance, around 40/60
• Considered, high-price purchases with long cycles: more brand
• Impulse or low-price categories: more activation
• Aggressive share-grab phase: protect brand even under short-term pressure

Divide the performance budget across paid, SEO, and CRO

Inside the performance slice, three buckets do different jobs on different clocks. Paid search and social buy demand now; Gartner's 2024 data shows digital taking 57.1% of paid media, led by search at 13.6% and social at 12.2% of total budget. SEO and content build an asset that pays later: HubSpot's long-running data shows inbound leads cost 62% less than outbound, and industry benchmarks put organic search conversion near 2.4%, ahead of paid search around 1.3%. The catch is time. SEO needs six to twelve months to compound, so fund it as a fixed line, not a variable one you raid when paid gets tight.

CRO is the multiplier that makes every other dollar work harder, and it is chronically underfunded. Only about 39.6% of companies run a documented CRO program, yet firms that invest in optimization report an average ROI near 223%. Because conversion rate is the denominator under every channel's cost per acquisition, a single point of lift divides straight into paid CPA and organic CPA alike. WellBuilt treats CRO as overhead on the whole budget, not a line item competing with media, because that is the math.

Your budget is optimal only when the next dollar earns the same return in every channel. Until then, move money.

Trust incrementality, not attribution, for the verdict

Attribution tells you which click was nearby when someone converted. Incrementality tells you whether the spend caused the conversion at all. The gap is large and expensive. Analysis across 200-plus companies found last-click models overcredit paid search by 40 to 65%, mostly on branded and competitor terms that capture demand the brand already owned. The starkest figure comes from a dataset of 225 DTC geo-tests run between 2024 and 2025, where branded search posted a median incremental ROAS of just 0.70x, well below the 1.0x breakeven and the portfolio median of 2.31x.

That number is directional, not gospel; it comes from a single vendor's sample of measurement-savvy DTC advertisers, US only, and you should validate before you cut. But the lesson holds. Geo holdout and lift tests are the gold standard precisely because they survive cookie loss and cross-device gaps that broke user-level tracking. Run an incrementality test on your largest line items before you scale them. A channel that looks like your hero in last-click can turn out to be subsidizing conversions you would have won for free.

Reallocate by marginal return, where the next dollar earns most

The decisive number is marginal ROAS: what your next dollar returns at today's spend level, not the blended average you report up the chain. Channels saturate. The first $100K in a channel might return 5.0x while the tenth returns 1.8x, even though the average still looks healthy. A channel can show a blended 4x while its last 20% of spend earns under 1x. Budget is optimal only when the marginal return is equal across channels, because any time one channel's next dollar beats another's, moving money up improves the total.

This is reallocation, not new spending. Media-mix work typically trims overinvested channels by 15 to 25% and feeds underinvested ones by 30 to 50%, lifting total efficiency 20 to 35% at a flat budget. A 2024 econometric study of ecommerce brands measured a 12.9% revenue gain from exactly this kind of shift with no increase in spend. WellBuilt runs budgets on a test-and-shift cadence: measure marginal return, move money toward the steeper part of the curve, and re-measure. Set-and-forget budgets quietly overpay the saturated channels.

Hold a test reserve and reallocate on a schedule

Use the 70/20/10 frame to keep a budget from calcifying. Put 70% behind proven channels carrying current results, 20% into emerging tactics that have shown early promise, and 10% into genuine experiments that might fail. The 10% is your pipeline of future 70% channels; cut it and you stop discovering anything. Email earns its place in the proven tier on math alone, returning roughly $36 for every $1 spent by Litmus's count, which is why it rarely loses its line.

Cadence beats any single allocation. Pull marginal CPA and incremental ROAS monthly, run a holdout test before scaling any line past 20%, and rebalance quarterly rather than once a year. Search behavior, auction prices, and saturation points all move, so a split that was optimal in Q1 leaks money by Q3. The discipline is unglamorous: measure, shift toward marginal return, protect the brand and test reserves, and repeat. That loop, not a perfect spreadsheet, is what compounds.

Attribution vs. incrementality for budget decisions

Attribution Reports credit

• Assigns conversion credit to touchpoints along the path
• Cheap and always-on, but distorted by cookie loss and cross-device gaps
• Overcredits closing channels; last-click can inflate paid search 40 to 65%
• Good for pacing and diagnostics, weak as a reallocation verdict

Incrementality Proves cause

• Uses geo holdouts and lift tests to isolate what spend actually caused
• Survives privacy and tracking limits by comparing regions, not users
• Exposed branded search at a median 0.70x incremental ROAS in DTC tests
• Slower and effortful, but the right basis for scaling or cutting a line

Key takeaways

• Size the total from revenue first: about 7.7% on average, higher for B2B, then bend it to your margins and growth targets.
• Set the brand-versus-performance split before channel splits; 60/40 is the long-run benchmark, with new brands skewing to brand and established ones to performance.
• Fund SEO and CRO as fixed lines, not variable ones; inbound leads cost 62% less and a point of conversion lift divides into every channel's CPA.
• Judge channels by incrementality, not last-click; branded search can post a sub-1.0x incremental ROAS while attribution calls it a winner.
• Reallocate by marginal ROAS on a quarterly cadence, trimming saturated channels and feeding efficient ones; this alone lifts efficiency 20 to 35% at a flat budget.

SourcesGartner CMO Spend Survey (budget share and paid media channel breakdown), 2024 · U.S. Small Business Administration marketing budget guidance, cited 2024 · Binet & Field, The Long and the Short of It, IPA Databank (996 case studies), 2013 · Nielsen ROI Report, '50-50-50 Gap' underinvestment finding (50% of plans underinvested by a median 50%; digital video 66%, display 60%), 2022 · Stella DTC Digital Advertising Incrementality Benchmarks (225 geo-tests), 2025 · Last-click overcrediting analysis across 200+ companies (MCP Analytics), 2024 · HubSpot inbound vs. outbound cost-per-lead and CRO ROI statistics, plus WordStream organic vs. paid conversion-rate benchmarks, 2024 · Econometric study of ecommerce reallocation and media-mix saturation curves, 2024 · Litmus email marketing ROI ($36 per $1), 2024