Account-level ROAS is a summary, not a diagnosis. An account averaging 4x ROAS might contain campaigns at 8x subsidizing campaigns at 1.5x. Without looking at ROAS at the campaign, ad set, and audience levels, there's no way to know where money is working and where it's being wasted.
This guide walks through the process of segmenting ROAS across a paid ad account to find the specific placements, audiences, and campaigns that are dragging performance down -- and what to do about them.
Step 1: Calculate Your Break-Even ROAS Before You Start
Before you can judge whether a segment's ROAS is good or bad, you need a reference point. That reference point is break-even ROAS, not an industry benchmark.
Break-Even ROAS = 1 / Gross Margin
If your gross margin is 40%, break-even is 2.5. Any campaign segment running below 2.5x is losing money. Any segment running above it is contributing to profit.
Without this number, you're sorting campaigns by ROAS without knowing which are actually profitable. A campaign at 3x may look like an underperformer against a benchmark of 5x -- but if your break-even is 2.5, it's a profitable campaign. A campaign at 4.5x looks strong against benchmarks -- but if your break-even is 5x due to thin margins, it's losing money.
The free ROAS Calculator by EvvyTools calculates break-even ROAS from your gross margin. Get this number first.
Step 2: Export Campaign-Level ROAS Data
In Google Ads, pull a campaign performance report with ROAS as the primary metric. Filter for the past 30 days (or a longer period for low-volume accounts -- 90 days if monthly conversion volume is under 20 conversions per campaign).
In Meta for Business, use the Ads Manager breakdown tool to see ROAS by campaign. Ensure you're using a consistent attribution window across all campaigns -- comparing one campaign using 1-day click attribution to another using 28-day click gives misleading results.
Export to a spreadsheet, add your break-even ROAS as a reference column, and add a column calculating the gap between each campaign's ROAS and break-even (current_roas - break_even). Campaigns with negative gaps are unprofitable.
Step 3: Sort by ROAS and Identify the Distribution
Sort campaigns from lowest ROAS to highest. Look at the distribution:
- Are the low-ROAS campaigns concentrated in one product line, audience type, or campaign objective?
- Do the high-ROAS campaigns share common characteristics -- tighter audience targeting, specific ad formats, particular products?
- Is there a group of campaigns clustered just below break-even that could potentially be optimized above it?
The distribution tells you where to investigate further. Don't start making changes yet -- you're building a picture of the account structure.
Step 4: Drill Into Ad Set and Audience Level for Underperformers
For campaigns running below break-even, check whether the problem is concentrated in specific ad sets or spread across all of them.
In Google Ads, open the underperforming campaign and check ad group ROAS. Sort ad groups from lowest to highest. Often, one or two ad groups are pulling down an otherwise acceptable campaign.
In Meta, use the Campaign -> Ad Set -> Ad breakdown. Sort ad sets by ROAS. A campaign with a 2x average ROAS may contain three ad sets: one at 5x, one at 1.5x, and one at 0.5x. The 5x ad set is performing well; the other two are eroding the average.
The key question at the ad set level: is the ROAS problem caused by audience targeting (wrong people seeing the ad), creative (right people seeing an ineffective ad), or the product itself (a product that doesn't convert for anyone)?
Photo by u_wez38xk6qm on Pixabay
Step 5: Separate Audience Performance from Creative Performance
These are two different problems with two different fixes.
Audience problem: the ad is reaching people with low purchase intent or no connection to the product. Indicators include high impression volume with low CTR, or high CTR with low conversion rate. Fix: narrow targeting, use lookalike audiences, or exclude segments that click but don't convert.
Creative problem: the ad is reaching the right people but not compelling them to act. Indicators include reasonable CTR but poor conversion rate, with landing page traffic that doesn't convert. Fix: test new ad creative, improve landing page copy, or match the offer more tightly to what the creative promises.
Pausing the low-ROAS ad sets and watching whether campaign-level ROAS recovers is the practical test. If the remaining ad sets sustain their performance, the paused ones were the problem. If ROAS drops after pausing, the account may depend on volume from those sets in ways that aren't immediately obvious.
Step 6: Reallocate Budget Based on ROAS Findings
Once you've identified which segments are above and below break-even, reallocate budget toward the profitable ones.
The general approach:
- Reduce budget on campaigns consistently below break-even ROAS.
- Increase budget on campaigns consistently above break-even, starting with increments of 15-20% per week to avoid destabilizing the algorithm's learning phase.
- Pause ad sets that are dragging otherwise profitable campaigns below break-even.
One common error here: cutting budget on a campaign just because its ROAS is lower than another, without checking whether it's still above break-even. A campaign at 3x ROAS that's above break-even is profitable and worth running even if another campaign runs at 6x. The second campaign may be limited in how much additional budget it can profitably absorb.
More context on using ROAS for budget decisions is in this guide: How to Calculate ROAS and Make Your Ad Spend Actually Profitable.
Step 7: Reassess Monthly
ROAS by segment isn't a one-time audit. Campaign performance shifts with seasonality, auction competition, audience fatigue, and product changes. A campaign that was profitable in March may run below break-even in May for reasons unrelated to what you've optimized.
Setting a monthly cadence -- export, sort, compare to break-even, identify outliers -- keeps the account clean without requiring constant monitoring. Quarterly for accounts with stable, low-volume campaigns; monthly for higher-spend accounts where drift compounds faster.
For TikTok for Business campaigns, the same segmentation logic applies, though the platform's breakdowns are less granular than Google or Meta. Segment by campaign first, then by ad group where available.
The process is the same regardless of platform: calculate break-even, sort by ROAS, find the segments below it, and diagnose before cutting or reallocating budget.
Segmentation Thresholds: How Much Data Do You Need?
One risk in ROAS segmentation is making decisions on insufficient data. A campaign or ad set with five conversions over 30 days doesn't have enough statistical signal to conclude whether its 1.8x ROAS is a stable measurement or random variation.
A reasonable minimum before acting on segmented ROAS data: at least 20-30 conversions in the period being analyzed, and at least three weeks of data to account for weekly variation. Below these thresholds, ROAS at the ad set or audience level will fluctuate too much to be reliable. For low-volume campaigns, analyze at the campaign level rather than the ad set level, and use longer time windows (90 days instead of 30).
For high-volume accounts with hundreds of conversions per week, weekly segmentation analysis is feasible. For most small to mid-size advertisers with monthly conversion volumes of 20-100, monthly analysis at the campaign level and quarterly analysis at the ad set level is more realistic and less likely to trigger premature changes based on noise.
The free ROAS Calculator by EvvyTools helps translate segmentation findings into profitability assessments: once you've identified which campaign segments are worth investigating, enter their spend, revenue, and margin to determine whether they're above or below break-even. More detail on using ROAS for campaign evaluation is in the full guide: How to Calculate ROAS and Make Your Ad Spend Actually Profitable.
Top comments (0)