When to Increase Spend and When to Hold Budget

· Tutorials/Guides

When to Increase Spend and When to Hold Budget

"It's working, so spend more" is the most common budget decision we see, and it is not actually a decision. It is a reaction to a good week, made without checking whether the account has room to absorb more money without getting less efficient in the process.

Scaling and holding are both legitimate calls. The problem is that most teams make the call based on 1 number, usually ROAS or revenue from the last 7 days, when the real answer depends on 3 or 4 signals working together. Get this wrong in either direction and you either leave growth on the table or torch efficiency chasing volume the account was never ready to support.

Here is the framework we use, and the 2 different reasons CPMr can be telling you to hold even when every other number still looks fine.

The Scale Signal

3 conditions need to line up before we recommend increasing budget on a campaign or account. MER needs to be stable or improving over a trailing window long enough to smooth out weekly noise, not just strong in the most recent 7 days. CPMr, cost per 1,000 people reached, needs to be flat or declining, which tells you the account still has room to reach new people without paying an escalating price to do it. Impression share needs genuine headroom, meaning you are not already winning close to every available auction in your target audience.

When all 3 line up, scaling is close to a low-risk decision. The account has proven efficiency, has not saturated its current audience, and has real auction room left to buy. That combination is rarer than most budget conversations assume, which is exactly why checking for it matters instead of assuming a good week means all 3 are true.

The Hold Signal, and Why 1 Number Isn't Enough to Read It

CPMr climbing is the clearest single signal that tells you to hold rather than scale, but it splits into 2 completely different problems that call for 2 completely different fixes.

CPMr is CPM multiplied by frequency, the true cost of reaching 1 unique person rather than the cost of 1,000 impressions. If the same person sees your ad 5 times, that is 5 impressions and 1 person reached. CPM alone stays flat in that scenario while CPMr keeps climbing, which is exactly why CPMr catches a problem before CTR drops and well before ROAS moves. By the time those platform metrics show trouble, CPMr has usually been elevated for weeks already.

When frequency is rising and CPM is holding flat, your budget is cycling through the same audience on repeat. That is audience saturation, and the fix is exclusions and campaign consolidation to force the algorithm back into unreached inventory, not more budget. When CPM is rising and frequency is holding flat, you are facing auction pressure from increased competition, a different problem entirely, and no amount of exclusion or consolidation fixes a price war you are simply losing. Same headline number climbing. 2 different diagnoses. 2 different responses, and scaling budget into either one without first identifying which cause is driving it just accelerates the wrong fix.

Why "It's Working" Isn't a Strategy on Its Own

A strong ROAS week tells you the campaign converted well against the audience and creative it had available that week. It does not tell you whether that same audience has room to absorb 2 times or 3 times the budget without the efficiency collapsing, because a good week and genuine headroom are 2 different facts that happen to look identical on a dashboard.

We have seen accounts double budget off a strong week and watch CPMr spike within 10 days, because the campaign was already close to saturated and the extra spend just accelerated the same audience getting hit more often. Check headroom before you check the mood of the last 7 days. The mood tells you the campaign is healthy right now. Headroom tells you whether it can stay healthy at a bigger number.

Building This Into a Weekly Cadence, Not a Quarterly Guess

The accounts that scale well treat this as a weekly check, not a quarterly budget meeting. Every week, pull MER trend, CPMr trend broken into its CPM and frequency components, and impression share for your top-spend campaigns. Most weeks the answer is hold steady. Some weeks it is scale with confidence. Occasionally it is pull back before a saturation problem gets worse and more expensive to unwind.

Building this cadence before Q4 matters more than at any other point in the year, because Q4 volume amplifies whatever state the account was already in. An account with real headroom scales into Q4 efficiently. An account already showing saturation signals in September gets that problem multiplied by November spend levels, at a moment when it is far more expensive to fix.

Scaling and holding are both good decisions when they come from checking MER, CPMr, and impression share together instead of reacting to a single strong or weak week. When CPMr is climbing, diagnose whether frequency or CPM is driving it before you choose a fix, because exclusions solve saturation and do nothing for auction pressure, and the reverse is just as true.

Run this check against your top campaigns this week, before you finalize Q4 budget increases. The 10 minutes it takes to pull MER, CPMr, and impression share together is cheaper than finding out the hard way in November that the account never had the headroom you assumed it did.

Scaling on a good week and scaling on real headroom are not the same decision.


Back to all articles