Cross-Channel Budget Pacing: When to Shift Spend Between Google and Meta

The account that performs best in a month is rarely the one with the best creative. More often it is the one that had its budget in the right place at the right time. Pacing gets less attention than creative, but it decides whether you hit the monthly target or have to explain why you missed it.

Average ROAS is the wrong number to allocate on

In benchmark reports, Google often looks like the obvious place to put money. WordStream’s 2025 benchmarks cover more than 16,000 US search campaigns from April 2024 through March 2025. Google Ads had a median conversion rate of 7.52% at a $5.26 average CPC. Meta’s lead campaigns had a similar conversion rate, 7.72%, but a much lower CPC of $1.92 and a cost per lead of $27.66, compared with $70.11 on Google, per WordStream’s 2025 Facebook data. These numbers compare the two platforms on averages. None of them tells you where the next dollar should go.

An average describes past spend. It mixes every dollar you already spent, including the cheap early conversions and the expensive later ones, into one number. That number is useful in a report for management. It does not help with a budget decision, because that decision is about the nextdollar you spend. When you pace across two platforms, the question is “which channel will get more from my next dollar?” The past average does not answer it.

What marginal ROAS measures

Marginal ROAS is the revenue from one additional unit of spend. It shows the effect of the next dollar on its own, so it reveals diminishing returns that a blended average hides. In Lifesight’s definition, it tells you how effective your next dollar will be, while average ROAS tells you how effective your past spending was. To calculate it, divide the change in revenue by the change in spend.

This can change which channel gets the money. Say you run both platforms and need to decide where an extra $2,000/day should go. Google has a much higher average ROAS. The table shows what the next $2,000 returns on each.

Illustrative figures.Google has the higher average ROAS, 4.2x against 2.3x. But its next $2,000 returns only 2.5x, and Meta’s returns 3.2x. The extra spend should go to Meta, even though Meta looks worse in the report. If you allocate by the average, you add money to a channel that is already saturating. If you allocate by marginal return, you buy the cheaper conversions that are still available.

Reading the diminishing-returns curve

Revenue does not grow in a straight line with spend. The measurement firm Measured describes three phases that every channel goes through: increasing returns, linear returns, then diminishing returns. Saturation is normal and happens in every channel. What differs is where your account is on the curve today.

You can see it in the numbers. Ruler Analytics describes a prospecting campaign where spend went from £55,500 to £60,200. Revenue grew, but the marginal ROAS on the added spend was 5.22, about half the overall ROAS of 11.05. The campaign was still profitable. It had also reached the part of the curve where each extra pound brought in clearly less than the one before. A large gap between average and marginal ROAS is the sign to look for.

When marginal ROAS falls to about half of average ROAS, the channel is close to saturation. Put the next dollar in a channel that can still grow.

For pacing, this means the following. When a platform’s marginal return falls well below its own average, extra spend there buys few additional conversions, and they are expensive. Move that money to the channel whose curve is still steep. The saturated channel has not “stopped working.” The same money simply returns more in the other channel.

The equimarginal rule: where the next dollar goes

The equimarginal principle tells you when to stop moving budget. Budget is allocated best when the next dollar has the samemarginal ROAS in every channel. While Meta’s next dollar returns more than Google’s, keep moving spend to Meta. As Meta’s spend grows, its marginal ROAS falls. As Google’s spend shrinks, its marginal ROAS rises. Stop when the two are equal. At that point total return is at its highest.

The gains can be large. In a Measured case study, a $50M budget was moved toward channels with higher marginal returns. The result was a 13.4% incremental revenue gain, about $16.3M, with zero additional spend. Moving budget you already have is often the cheapest way to grow revenue. It takes time to analyze, but no extra media spend.

There is one limit. You can only equalize marginal ROAS if you can estimate it, and each platform tends to over-report the revenue it drives. Meta and Google count conversions with their own models and attribution windows, so the ROAS numbers inside each platform cannot be compared directly. Compare the two on a shared result instead, such as blended ROAS from GA4 or your CRM, or cost per qualified lead. Estimate the marginal return from a real spend change you made. The dashboard’s default view will not give you this number. If the input numbers are wrong, the allocation will be wrong too.

For the same reason, it makes sense to manage the two platforms as one system. We cover this in running Google Ads and Meta Ads together. The two budgets affect each other even if your tools treat them as separate.

Pacing without underspending or overspending

Knowing where spend should go is one part of the job. The other part is ending the month on budget. You do not want to finish 15% under because you were too careful, or 20% over because one strong day spent more than planned. For this you need to know how each platform spends a daily budget.

On Google, the average daily budget works as a monthly limit. Google can spend up to 2x your average daily budget on any single day, but will not bill more than 30.4x that budget across a calendar month. A $100/day budget is therefore a budget of about $3,040 per month, and Google can spend it unevenly across the days. As of June 2026, Google also actively paces spend toward the full monthly amount, even for campaigns on limited ad schedules. A campaign that spent little in the middle of the month can then spend much more later to catch up. Expect daily spend to go up and down, and check it against the monthly total.

A pacing check you can run in five minutes

1. Calculate month-to-date pace. Divide spend to date by the share of the month that has passed. If you have spent 40% of the budget on day 20 of 30, you are on track to spend only 60% and underspend by 40%.
2. Check marginal ROAS on each channel. Estimate it from your last spend change. Put the catch-up spend on the platform where the next dollar returns more.
3. Adjust in small steps. Large budget jumps distort the pacing math and can disrupt platform learning (see the next section).
4. Watch the upper limit. Make sure one high-demand day does not use up the budget you need for the rest of the month.

Underspending is also a mistake. Every dollar you leave unspent while marginal ROAS is still above your profit threshold is revenue you chose not to get. Good pacing avoids both errors: going over the budget and leaving profitable spend unused.

How often to rebalance

A common habit is one large budget change on Monday morning. On Meta, this costs you. A budget change larger than roughly 20% can reset an ad set’s learning phase. The ad set then has to collect the ~50 optimization events per week it needs to exit learning again. One large weekly change can lead to days of unstable, unoptimized delivery, which is the opposite of what a pacing correction should achieve.

The better approach is to check the numbers often and change budgets in small steps. Small, frequent changes keep each channel close to its best point on the curve and do not trigger learning resets. The same approach applies to automated bidding strategies and cutting wasted spend: many small corrections work better than a few large ones.

Large weekly budget changes can reset learning. Small hourly changes keep each channel near its best point without a reset.

Doing this by hand does not scale. You would have to estimate marginal ROAS on both platforms, check month-to-date pace and move budget in steps under 20% across dozens of campaigns, every hour, in every account. That is more than one person can do manually. Adriva runs this in the background. It rebalances budget between Google and Meta every hour based on marginal return. We explain how in how Adriva rebalances budget hourly. This article describes the logic. Adriva applies it every hour instead of once a week.

Summary

To pace budget across channels, judge each channel by its marginal return. Move spend toward the channel with the higher marginal return until the two platforms are equal. Make the changes in small, frequent steps, so you end the month on budget and do not reset what the algorithms have learned.

• Use marginal ROAS for every allocation decision.
• Equalize marginal ROAS across Google and Meta to get the highest total return.
• Make small, frequent changes to stay on budget and avoid learning resets.