It's one of the most common messages we get from owners running their own ads. The campaign was returning 4x at $50 a day, they moved it to $100, and within a week it was closer to 2x. The instinct is to assume Meta broke something. It almost never did.
When ROAS falls after a budget increase, one or more of five things is happening. Some are real and some are an illusion created by how Meta counts sales. Each has a different fix, so it pays to know which one you're looking at.
- Extra budget buys your next-best customers, so the return on the added dollars is lower than your average. That part is normal.
- Big budget jumps can push an ad set back into the learning phase, and results wobble until it settles.
- Sales from new spend can take up to seven days to be credited, so a budget change judged after two days always reads low.
- In a small audience, more budget mostly means the same people seeing the same ad more often.
- The real scaling lever is new creative, not a bigger number in the budget field.
1. You bought the easy customers first
At any budget, Meta shows your ads to the people it predicts are most likely to buy. At $50 a day it only has to find the most likely handful. At $100 a day it has to go further down the list, to people who are a little less ready, a little less interested, or a little more expensive to reach. Your average return falls because the extra dollars buy weaker results than the first dollars did.
That's why average ROAS is the wrong number to scale on. What matters is the return on the added spend, often called marginal ROAS. A simple example:
| Daily budget | Daily revenue | Average ROAS | Return on the added spend |
|---|---|---|---|
| $100 | $400 | 4.0x | - |
| $200 | $600 | 3.0x | 2.0x (the extra $100 brought in $200) |
Illustrative numbers, not a client account.
The account got worse on average, but the extra $100 still returned $200. Whether that's a good trade depends entirely on your margin, which is why you need your break-even ROAS before you scale. We show how to work it out in what is a good ROAS for Facebook ads.
2. The bigger budget reset the learning phase
Every ad set goes through a learning phase while Meta works out who converts. Meta's guidance is that an ad set generally needs about 50 optimization events in a week to leave it. A significant edit can send it back to the start, and a large budget change counts as one. While an ad set is learning, delivery is less stable and cost per result usually runs higher.
Meta doesn't publish an exact threshold for how big a change has to be. That's why experienced buyers raise budgets in steps, commonly around 20% at a time, and wait a few days between them rather than doubling overnight.
3. The sales haven't been counted yet
Meta's standard attribution setting credits a sale to an ad if the buyer clicked it in the previous seven days or viewed it in the previous day. Ads Manager then reports that sale on the day of the ad, not the day of the purchase. Someone who clicked on Monday and bought on Friday adds to Monday's numbers, on Friday.
So the most recent days in any report are always missing sales that haven't happened yet. The spend from a budget increase shows up at once; much of the revenue it drives arrives over the following week. Judge the change after two days and you're comparing complete spend against incomplete sales. Wait at least seven days after the last change before deciding whether it worked.
4. You ran out of new people
A local business advertising within a 10-mile radius, or a niche product with a narrow audience, can only reach so many people. Double the budget and Meta doesn't find twice as many buyers. It shows your ads to the same people more often. Frequency climbs, click-through rate falls, and cost per result follows.
Check frequency before and after the increase. If it jumped and results fell, the audience is saturated. More budget won't fix that. A wider audience, a new offer or new creative will.
5. The winning ad was carrying everything
Most accounts lean on one or two ads that do the heavy lifting. At higher spend those ads wear out faster, because they're shown more. If nothing new is ready to take over, the whole account drops at once.
This is the most fixable of the five, and the one owners skip. New creative, tested continuously, is what lets an account absorb more budget. Even small accounts need it: on one climbing gym we manage, we tested 118 creatives across 34 campaigns in ten months, on daily budgets between £5 and £15.
How to scale without the drop
Know your break-even ROAS first
Divide 1 by your gross margin. At a 40% margin, break-even is 2.5x. Anything above it on the added spend is profitable growth, even if your average ROAS falls.
Raise in steps, then wait
Increase by around 20% and hold for at least seven days before the next step, so the learning phase and the attribution window both settle.
Measure the return on the added spend
Compare total revenue and total spend for the week before and the week after. Divide the change in revenue by the change in spend. That number, not the average, tells you whether to keep going.
Add creative before you add budget
Have new ads live and learning before you scale, so fatigue on the current winner doesn't take the account down with it.
Watch the checkout-to-purchase ratio
For online stores: if a bigger share of people start checkout but don't finish after you scale, the new traffic is less ready to buy. That points to the audience or the offer, not the budget.
Set a floor when you need one
If you can't afford to go below a certain return, a ROAS goal or cost-per-result goal tells Meta to hold the line. You'll usually spend less, but what you spend stays inside your margins.
When a lower ROAS is the right call
A falling ROAS isn't automatically a problem. If the added spend still returns more than your break-even, you're making more total profit, just at a lower rate. And for a business that sells memberships or repeat orders, a first purchase can sit near break-even and still pay back over the following months. The question isn't whether ROAS went down. It's whether profit went up.
The reverse is also true: more budget is not always the answer. When we took over a two-site climbing gym's account, its sales campaign had been returning £0.82 for every £1 spent. It now returns £2.99, on a third less spend. The fix there was structure and creative, not budget.
Questions people ask
How much should I increase my Facebook ad budget at once?
There's no official number. Steps of around 20%, held for at least a week, keep most ad sets out of a fresh learning phase and give the attribution window time to fill in. If you need to grow faster, add a new ad set or campaign with fresh creative rather than doubling the existing one.
How long should I wait before judging a budget increase?
At least seven days after the last change. Meta credits a sale up to seven days after a click, back to the day of the ad, so the most recent days in any report always look worse than they'll end up.
Is it better to duplicate a winning ad set or raise its budget?
Raising the budget in steps keeps what the ad set has already learned. A duplicate starts a new learning phase and splits delivery between two ad sets chasing the same people. Duplicate when you're changing something else as well, such as the audience or the creative.
Does this apply to Advantage+ and campaign budget campaigns?
Yes. Campaign-level budgets and Advantage+ campaigns move spend between ad sets and ads for you, but they buy from the same auction under the same limits. Extra budget still reaches less-ready buyers, the same attribution lag applies, and creative still wears out.
What if ROAS drops when I haven't changed the budget?
Then it's usually creative fatigue, a tracking fault or a seasonal shift. Our weekly Facebook ads health check covers all three in about 20 minutes.
When is it worth getting help with scaling?
When the account falls apart every time you scale and you've ruled out the five causes above. The fix is usually creative volume or tracking, which is most of what we do. We're not the right fit if your budget can't cover a real ad spend alongside a management fee; at that stage, the steps above applied patiently will get you further than an agency would.
If your account falls apart every time you try to scale, book a free strategy call. We'll look at the account with you and tell you which of the five you're dealing with.
This post replaces one we first published in December 2020. The auction has changed a lot since then. The question hasn't.