ROAS below target? Check these five levels before you book new creatives
By
Josef Raasch
The reflex that costs money
The ROAS is below target, and the first impulse is almost always the same: new creatives. New hooks, new videos, a new creator. Sometimes that's true. But in the accounts we take over, it's rarely the cause. One metric demasks this reflex immediately: the click-through rate. An account can have an excellent click-through rate and still a poor ROAS. In that case, the creatives are working. The problem lies elsewhere.
We therefore work with a fixed sequence of five levels. Holistic before detail, marginal costs before average, every measure as a testable hypothesis. The order is chosen deliberately: the first levels are the ones most frequently overlooked and the ones that work the fastest.
Level 1: The Goal Mix
How much budget is running on campaigns whose goal is not purchase or lead? Traffic and engagement campaigns have often grown historically. They deliver cheap clicks, drive the account's click-through rate, and appear positive in the weekly report. They barely contribute to revenue. Every euro spent there depresses the blended ROAS, which is the ROAS across the entire account.
The hypothesis: shift budget from traffic goals to conversion campaigns and monitor the blended ROAS. Not the campaign ROAS individually, but the number your controlling team sees.
Level 2: Paused Winners
It sounds mundane, but it can be found in almost every account: campaigns with excellent ROAS that are on pause. Sometimes due to a budget freeze in the last quarter, sometimes because someone wanted to test, sometimes nobody remembers. Every day costs profitable revenue.
Check the campaign list of the last 90 days including paused units, sorted by ROAS. Reactivate paused winners that have no apparent reason for being paused and scale in a controlled manner, with a maximum of around 20 percent budget per day so that the learning phase does not restart.
Level 3: Creative Fatigue
Only now do the creatives come in, and here, too, the frequency comes first. If it is over 3.5 and the click-through rate has collapsed in the same period, the target audience has grown tired of the ad. Typical in retargeting, where small target audiences see the same ads over and over again. From a frequency of 5 upwards, it becomes expensive.
The solution is variety, not volume: conceptually new ads with a different hook or format, plus a frequency limit. Meta recognizes minor variations of the same motif as duplicates and they do not help.
Level 4: The Funnel after the Click
Now it gets interesting. Build the chain once completely:
Impressions to link clicks: the click-through rate, the job of the creative
Link clicks to landing page views: loading time and technical quality of the page
Landing page views to shopping cart: offer, product page, price, traffic quality
Shopping cart to checkout and purchase: trust, shipping costs, payment methods
The biggest drop shows you where the construction site is. If it breaks down between the landing page and the shopping cart, no new creative in the world will help. Then something is wrong with the offer or the page, or the ad promises something that the page does not deliver.
Level 5: Attribution
Finally, the question of whether the number is even correct. The ROAS depends heavily on the attribution window. Show several windows side-by-side in reporting: 1-day click, 7-day click, 1-day view, and the default setting. If the values are far apart, a large portion of your reported revenue consists of purchases that would have happened even without the ad. Establish one window as a binding valuation base, preferably the more conservative one, and evaluate all measures against it. Why this is the case is explained in detail in our article on how ROAS systematically lies.
A practical calculation example
A furniture online shop, 30 days, numbers rounded. The account ROAS was at 1.3, the target at 3.5. The click-through rate was around 5 percent, the CPM was about 5 euros. So the top-funnel was strong, creatives were working.
The funnel: 5.4 million impressions, 4.4 percent clicked, 37 percent of those reached the landing page, 0.8 percent of landing page views added to cart, 26 percent of those went to checkout, 92 percent purchased. The bottleneck was clearly between the landing page and the shopping cart.
In addition: around one third of the budget was in traffic campaigns with a single purchase, two paused campaigns with a ROAS of 12 and 5.5, two retargeting campaigns with a frequency over 4.5 and click-through rates under 0.7 percent. Depending on the attribution window, the ROAS was between 0.6 and 1.3.
The conclusion was clear: the leverage lay in budget allocation, reactivated winners, and the post-click path. New creatives would not have turned the result around.
The rule behind it
The algorithm optimizes for the cost of the next conversion, not for averages. Pausing segments, placements, or campaigns solely because of a higher average CPA often worsens the overall result. Therefore, every measure from the five levels is a hypothesis with an expected effect on the blended ROAS. Once you go through this diagnosis cleanly, you will know whether you really need new creatives or whether your account just needs to be cleaned up.
