Channel Mix Optimization: Allocating Budget Across Platforms

Your channel mix is probably based on whoever presented the best ROAS deck

Most brands allocate budget to the platform that reports the best ROAS. Meta shows 4x, Google shows 6x, TikTok shows 3x -- so Google gets the most money.

The problem is obvious once stated: each platform grades itself. And each platform's grading system is optimized to make itself look good. Google's 6x ROAS includes branded search (people who were going to buy anyway). Meta's 4x includes view-through conversions (people who scrolled past your ad). TikTok's 3x includes engaged views (people who watched 6 seconds of a video).

Comparing these numbers to determine channel allocation is like comparing test scores where each student wrote their own test. The student who made the easiest test looks smartest.

Channel mix optimization requires a common measurement framework that evaluates every platform on the same basis: incremental contribution to business outcomes.

The problem with platform-first allocation

Each platform claims the same conversions

A customer sees your TikTok ad Monday, clicks your Meta ad Wednesday, and converts through a Google search Friday. All three platforms claim the full conversion:

• TikTok reports 1 view-through conversion
• Meta reports 1 click-through conversion
• Google reports 1 click-through conversion

Your total conversions across platforms: 3. Your actual conversions: 1.

At $200K/month total spend across three platforms, this overlap typically inflates total attributed conversions by 40-70%. The resulting budget decisions are based on fictional math.

Platform algorithms optimize for their own ecosystem

Meta's algorithm optimizes to show your ads to people most likely to convert -- as measured by Meta. This means Meta preferentially targets users who are heavy Meta users and who convert in patterns Meta can track. It doesn't optimize for customers who discover your product on Meta but convert through Amazon, physical retail, or other untracked channels.

Each platform's optimization creates a self-reinforcing bubble. Meta finds "Meta converters." Google finds "Google converters." The platforms appear to work in isolation, but in reality, customer journeys cross platforms constantly.

Building a cross-platform measurement framework

Step 1: Establish a single source of truth

Pick one system as your conversion source of truth. This should be your ecommerce platform (Shopify, WooCommerce), CRM, or a server-side attribution tool. Every conversion is counted once in this system, regardless of how many platforms claim credit.

The rule: Platform-reported conversions are directional indicators. Server-side verified conversions are the decisions-making data.

Step 2: Calculate channel-level incrementality factors

Using holdout tests or other incrementality methods, determine what percentage of each platform's claimed conversions are truly incremental.

Based on aggregated data across ecommerce brands:

| Channel | Typical Incrementality Factor | Meaning | |---------|-------------------------------|---------| | Meta Prospecting | 0.55-0.70 | 55-70% of claimed conversions are real | | Meta Retargeting | 0.15-0.35 | 15-35% of claimed conversions are real | | Google Non-Brand Search | 0.60-0.80 | 60-80% are real | | Google Branded Search | 0.10-0.30 | 10-30% are real | | Google Shopping | 0.55-0.75 | 55-75% are real | | Google PMax (brand excluded) | 0.45-0.65 | 45-65% are real | | TikTok Prospecting | 0.35-0.55 | 35-55% are real | | YouTube | 0.40-0.65 | 40-65% are real | | Display/Programmatic | 0.10-0.25 | 10-25% are real |

These are ranges. Your brand's specific factors depend on brand awareness, product category, and competitive landscape. Incrementality testing gives you your precise numbers.

Step 3: Calculate incremental CPA and ROAS by channel

Multiply each platform's reported conversions by its incrementality factor, then recalculate CPA and ROAS.

Example for a brand spending $200K/month:

| Channel | Spend | Reported Conv. | Inc. Factor | Inc. Conv. | Inc. CPA | Inc. ROAS | |---------|-------|---------------|-------------|------------|----------|-----------| | Meta Prospecting | $70K | 1,400 | 0.62 | 868 | $81 | 2.1x | | Meta Retargeting | $25K | 800 | 0.22 | 176 | $142 | 0.7x | | Google Non-Brand | $35K | 600 | 0.72 | 432 | $81 | 2.3x | | Google Branded | $15K | 500 | 0.18 | 90 | $167 | 0.5x | | Google Shopping | $30K | 700 | 0.65 | 455 | $66 | 2.8x | | TikTok | $20K | 300 | 0.45 | 135 | $148 | 1.2x | | YouTube | $5K | 40 | 0.55 | 22 | $227 | 0.8x |

Step 4: Reallocate based on incremental metrics

The table above reveals clear reallocation opportunities:

Reduce: Meta Retargeting (0.7x incremental ROAS), Google Branded (0.5x), YouTube at current scale (0.8x)

Increase: Google Shopping (2.8x incremental ROAS), Google Non-Brand Search (2.3x), Meta Prospecting (2.1x)

Test further: TikTok (1.2x -- potentially viable with optimization), YouTube (may have unmeasured halo effects)

The reallocation process

Don't move everything at once

Shift 15-20% of budget per month. Major reallocations trigger algorithm relearning across all platforms, which can cause temporary performance dips.

Month 1: Cut retargeting by 40%, cut branded search by 50%. Reallocate to Google Shopping and Meta Prospecting.

Month 2: Monitor results. If total conversions held (or improved), cut retargeting and branded search further. Begin scaling TikTok if halo effect metrics are positive.

Month 3: Fine-tune. Adjust based on the new equilibrium. Run incrementality tests on the channels that received more budget to verify they maintained their efficiency at higher spend.

Account for channel interactions

Channels don't operate in isolation. Cutting Meta Prospecting may reduce branded search volume (because fewer new people discover your brand). Cutting YouTube may reduce Meta retargeting performance (because fewer people have been warmed up by video).

Monitor downstream effects of every change:

• When you reduce Meta Prospecting, watch for branded search volume declines
• When you reduce YouTube, watch for Meta retargeting and prospecting efficiency changes
• When you increase TikTok, watch for branded search volume increases (the halo effect)

These cross-channel interactions are real and meaningful. A channel with "low" incremental ROAS in isolation may contribute significantly to the performance of other channels.

Advanced: portfolio-level optimization

The efficient frontier

For sophisticated teams, portfolio-level optimization treats the channel mix as an investment portfolio. Each channel has an expected incremental return and a variance (risk). The optimal portfolio maximizes total incremental return at an acceptable level of variance.

This means:

• Diversification matters. Concentrating 80% of budget in one channel is risky even if that channel has the highest ROAS. Algorithm changes, policy shifts, or competitive dynamics can degrade any single channel quickly.
• Low-ROAS channels may earn a place. A channel with 1.5x incremental ROAS that doesn't correlate with your other channels provides diversification value. If Meta's performance drops 30% next month, having 20% of budget in non-Meta channels provides stability.
• Test budgets are investments, not costs. Allocating 10% of budget to experimental channels (Connected TV, podcasts, Reddit, influencer) is portfolio insurance. One of those channels may become your next growth driver.

Recommended allocation ranges by business stage

Early stage (under $50K/month total):

• 70-80% on 1-2 proven prospecting channels
• 10-15% on retargeting
• 5-10% on testing new channels

Growth stage ($50K-$250K/month):

• 50-60% on proven prospecting (2-3 channels)
• 10-15% on retargeting (tightly managed)
• 15-20% on scaling channels with proven incrementality
• 10-15% on testing

Mature stage ($250K+/month):

• 40-50% on proven prospecting (3-5 channels)
• 5-10% on retargeting
• 20-30% on brand and awareness channels
• 10-15% on testing and emerging channels

Notice that retargeting's share decreases as the brand matures. This is intentional: mature brands have higher organic demand, which means retargeting captures more organic conversions and has lower incrementality.

Frequently Asked Questions

How do I handle a channel that has high ROAS but low incrementality?

This is the branded search and retargeting dilemma. The channel looks great in platform reports but contributes little incremental value. The right approach is gradual reduction with monitoring. Cut spend by 20-30%, wait 3 weeks, and check total business conversions. If total conversions hold steady (organic absorbs the paid traffic), cut another 20-30%. Continue until you find the minimum spend level that maintains competitive defense (for branded search) or addresses genuine abandonment (for retargeting). Most brands can reduce these channels by 50-70% with minimal revenue impact.

Should I allocate more to Meta or Google for ecommerce?

It depends on your product and consideration cycle. Products with strong visual appeal and impulse-buy potential (apparel, beauty, home decor) tend to produce higher incremental ROAS on Meta because scroll-stopping creative drives discovery. Products with clear search intent (replacement parts, specific tools, comparison-shopping categories) tend to produce higher incremental ROAS on Google because users actively search for them. Most ecommerce brands benefit from 40-60% Meta, 30-40% Google, and 10-20% across TikTok and other channels, but your incrementality data should determine the exact split.

How do I allocate budget when entering a completely new channel?

Start with 5-10% of your total budget for 8-12 weeks. This is enough to generate meaningful data without risking significant waste. Set clear success criteria before launching: "If TikTok produces incremental CPA below $80 after 8 weeks, we'll increase to 15% of budget." Run an incrementality test during the trial period (geographic holdout works for any channel). If the channel meets your criteria, scale gradually. If it doesn't, reallocate the test budget. Never scale a new channel past 10% of total budget without incrementality validation.

---

Go Funnel uses server-side tracking and multi-touch attribution to show you which ads actually drive revenue. Book a call to see your real numbers.