What to check before you touch anything
- Campaign structure. Is the volume and makeup of keywords/targeting comparable — number of keywords, match types, number of ad groups, audience signal coverage (for Performance Max)? A campaign with a narrow structure physically can't deliver the same volume as one with broad coverage, even with an identical target — this isn't an efficiency question, it's a question of the structure's scale.
- Traffic type. The same tROAS can apply to fundamentally different types of traffic — brand versus non-brand, search versus Shopping/Performance Max, cold audiences versus remarketing. Each type has a different "natural" conversion rate and cost per click, so an identical target naturally produces different volume simply by nature, not because of a setup difference.
- Demand capacity. How large is the actual market/niche each campaign operates in — the real number of people searching for the relevant products/services? A campaign in an objectively small niche physically can't deliver the same volume as one in a large category, regardless of target or settings.
- Competition. How does the competitive environment differ between the niches/segments each campaign serves (Auction Insights: overlap rate, number of auction participants, outranking share)? The same tROAS can be easy to hit in a low-competition category and nearly unreachable in a high-competition one — and that shows up in volume, not in whether ROAS is being hit at all.
- Does the budget differ between the campaigns? If one campaign's budget is lower, it may be "Limited by budget" regardless of the fact that the target is the same.
- Is one of the campaigns still in "Learning" status, or did it recently undergo a change? Comparing a mature, stable campaign against a recently changed one isn't a fair comparison.
Possible approaches
- On structure. If the difference comes down to different keyword/targeting volume, comparing "campaign vs. campaign" on the same tROAS isn't quite fair in the first place — it's more useful to compare their reach potential separately from the target, and then judge, adjusted for that, whether one of them is underusing its real potential.
- On traffic type. If the campaigns represent different traffic types (brand vs. non-brand, search vs. Shopping), it's usually worth abandoning the idea of one shared tROAS for both — better to set the target separately for each traffic type, based on its own achievable economics, rather than unifying the target for the sake of management simplicity.
- On demand capacity. If one segment is capped by an objectively small market, that's not something settings can fix — treat the current volume as the ceiling for that niche, rather than treating the difference from a bigger campaign as a problem to be solved.
- On competition. If the gap is explained by different competitive intensity between niches, it's worth accepting that "the same tROAS" doesn't guarantee the same achievable volume, and setting the target separately for each niche based on its own auction (via Auction Insights), rather than a single unified figure across the account.
- If one campaign is budget-constrained, it's worth equalizing budget conditions (or explicitly accounting for the difference in your comparison) before drawing conclusions about the target's efficiency on its own.
- If the gap remains unexplained after all these checks, you can test equalizing conditions (temporarily matching budget, comparable keyword coverage) in a controlled experiment, to isolate whether the target itself really behaves differently in these two environments.