Performance Marketing

Ad Budget Allocation Across Channels: A Profit-First Method

Allocate ad budget across Google, Meta and other channels by marginal profit, not average ROAS. A step-by-step method with a worked example and checklist.

TL;DR
  • Average ROAS tells you how a channel did on all the money it already spent. Budget decisions are about the next dollar, so allocate on marginal returns: what one more step of spend adds in incremental contribution margin.
  • Set a profit floor first (your break-even ROAS on contribution margin), then move budget in steps toward the channel where the next step adds the most profit, and away from the channel where the last step lost money.
  • Reallocate on a fixed rhythm, monthly for most stores, using incrementality tests and spend step tests rather than platform-reported ROAS alone.

Why do average ROAS numbers lead to bad budget decisions?

Most budget meetings start with a table of channels and their return on ad spend. The channel with the highest number gets more money. It sounds logical, and it is usually wrong for two reasons.

Averages hide the curve. Every channel has diminishing returns. The first dollars in a campaign reach the easiest buyers: people searching your brand, recent visitors, the best-matched audiences. Each extra dollar reaches people who are a little less likely to buy. A channel can show a strong average ROAS while the last few thousand dollars you added barely broke even.

Platform ROAS is not incremental. Platforms count conversions they touched, not conversions they caused. Retargeting and branded search look excellent because they sit close to purchases that would often happen anyway. Upper-funnel channels look weak because their effect shows up later and through other channels.

So the question to ask is not “which channel has the best ROAS?” It is “where does the next $10,000 produce the most incremental profit, and where did the last $10,000 produce the least?”

What do you need to know before moving a dollar?

You need three numbers per channel. None of them come straight from an ads dashboard.

1. Your profit floor

Start with contribution margin per order: revenue minus product cost, shipping, payment fees, discounts and expected returns. From that, calculate your break-even ROAS:

Break-even ROAS = 1 / contribution margin rate (before ad spend)

A store with a 40% contribution margin before ads breaks even at a ROAS of 2.5. Below that, each sale costs more in ads than it earns. Our break-even ROAS calculator does the math for your own margins.

2. An incremental view of each channel

Adjust platform numbers toward reality. The best evidence comes from incrementality testing: holdouts, geo tests and lift studies. If you spend enough across several channels, marketing mix modeling adds a top-down view. If you have neither yet, use a store-level check: compare total revenue and new customer counts in periods when spend changed, alongside your marketing efficiency ratio.

3. The shape of the curve

For each channel, estimate what happens at the next spend level, not the current one. That is the marginal return. The next section covers how to get it.

How do you estimate marginal returns for each channel?

You will never know the curve exactly. You can get close enough to make better decisions than averages allow.

  • Spend step tests. Raise or lower one channel’s budget by a set step, hold everything else steady for two to four weeks, and measure the change in total store contribution margin and new customers. The change divided by the extra spend is your marginal return for that step.
  • Platform forecasts. Google Ads Performance Planner forecasts how results change at different budgets. It supports Search, Standard Shopping, Performance Max, Demand Gen and App campaigns, and since March 2026 no longer supports planning for Display and Video. Treat its output as a starting hypothesis, since it forecasts platform-measured conversions, not incremental ones.
  • Your own spend history. Plot weekly spend against weekly incremental results for each channel over the past year. Flattening at higher spend levels is the diminishing return showing itself.
  • Impression share and audience limits. On search, when you already capture most of the available impressions for profitable queries, extra budget mostly buys pricier clicks. Low search impression share on profitable non-brand terms is often where room remains.

Keep the step size consistent across channels, for example $5,000 or $10,000 a month, so you compare like with like.

Worked example: allocating the next $10,000 (illustrative numbers)

A store spends $120,000 a month on ads. Contribution margin before ads is 40%, so break-even ROAS is 2.5. The team has run step tests and a retargeting holdout over the past quarter.

Channel Current monthly spend Platform ROAS Marginal ROAS on next $10k Contribution from next $10k Profit after the $10k
Google Search (non-brand) $40,000 4.1 2.6 $10,400 +$400
Meta prospecting $60,000 2.9 3.0 $12,000 +$2,000
Retargeting (Meta and Google) $15,000 8.2 1.2 $4,800 -$5,200
YouTube test $5,000 1.6 2.8 $11,200 +$1,200

Contribution from the next $10,000 is the incremental revenue from that step times 40%. Profit is that contribution minus the $10,000.

Read it from the right-hand column, not the platform ROAS column:

  1. Retargeting has the best platform ROAS and the worst marginal return. The holdout showed most of its sales would have happened anyway. Cutting $10,000 there costs about $4,800 in contribution and saves $10,000 in spend, a net gain of about $5,200.
  2. Meta prospecting has an ordinary average but the best next step. It goes first in line for new budget.
  3. YouTube looks weak in-platform but adds profit at the margin, because its sales show up in branded search and direct traffic. It earns a larger, measured test, not a blind jump.
  4. Search is close to the floor. Hold it, and work on query quality before adding money.

The result: move $10,000 from retargeting into Meta prospecting this month, run a second step on YouTube next month, and re-measure. Total spend stays the same while profit rises.

Our guide to retargeting without waste covers how to cut retargeting spend without losing the sales it actually drives.

How should new customers change the math?

First-order break-even is a strict floor. Many stores can afford to acquire new customers at a loss on the first order, because repeat orders pay it back. The risk is using that idea to justify any spend.

Use a bounded rule. Allow a channel to run below first-order break-even only when:

  • It brings mostly new customers, verified with your own order data, not platform labels.
  • Your repeat contribution margin from those customers, within a set CAC payback period such as 6 or 12 months, covers the gap.
  • The cohort data supports it. If customers from that channel repeat less than average, the allowance shrinks.

For example (illustrative numbers), if a new customer earns $30 of contribution on the first order and an average of $25 more in repeat contribution within 6 months, you can pay up to $55 to acquire them and still pay back within 6 months. Paying $60 does not pay back in that window, however good the ROAS looks.

On Google Ads, customer lifecycle goals can help by bidding higher for new customers, or only for new customers, in Search, Performance Max, Shopping and Demand Gen. That keeps acquisition budgets from drifting toward existing buyers. Profit on ad spend, explained in our POAS vs ROAS guide, is a useful way to report all of this in one number.

How often should you reallocate, and how fast?

Monthly works for most stores in the $1M to $50M range. Weekly moves react to noise; quarterly moves leave money on the table.

A few practical rules:

  • Move in steps, not leaps. Large sudden budget changes make results harder to read and can unsettle automated bidding. Several smaller steps tell you more about the curve.
  • Mind daily pacing. Google Ads can spend up to twice a campaign’s average daily budget on a given day, while keeping monthly charges to no more than 30.4 times that daily budget. Judge Google spend over the month, not on a single day.
  • Separate tests from reallocations. A channel under an incrementality test should not also get a budget change mid-test.
  • Plan around seasonality. Marginal returns rise in peak demand periods because more people are ready to buy. Shift budget toward those weeks, and expect returns to fall back after.

Monthly budget allocation checklist

  • [ ] Contribution margin and break-even ROAS updated with current costs and return rates
  • [ ] Platform ROAS adjusted with the latest incrementality or holdout results
  • [ ] One spend step test completed or in progress on at least one channel
  • [ ] Marginal ROAS estimated for the next step up and the last step down on each channel
  • [ ] Budget moved from the lowest marginal profit step to the highest
  • [ ] New customer share and CAC payback checked from your own order data
  • [ ] Google Ads spend reviewed on a monthly basis, not daily
  • [ ] Results reported as incremental contribution margin and MER, not platform ROAS alone

FAQ

Is there a standard split, like 70% prospecting and 30% retargeting?

No fixed split holds across stores. The right mix depends on your margins, audience sizes and where each channel sits on its curve. Rules of thumb ignore all three. A small store with a narrow audience may spend very little on retargeting, while a store with heavy site traffic can spend more before returns fall. Let marginal profit decide.

What if I cannot run incrementality tests yet?

Start with what you have. Use your own order data to check how many orders per channel come from new customers, watch store-level revenue when budgets change, and discount channels that sit closest to the purchase, such as retargeting and branded search. Then plan one simple holdout, usually on retargeting, because it is cheap to test and often over-credited.

How much of the budget should go to testing new channels?

Set aside a fixed share for tests, sized so the store can afford to lose it if the test fails. Give each test a clear success threshold in marginal contribution, a set duration and a stop date. A test channel that clears the threshold earns its way into the main budget in steps, like any other channel.

Turn this into a plan for your store

The Free Profit Audit scans your store, sizes your biggest profit leaks and builds a 90-day plan.

Free Profit Audit