Let’s break down why you shouldn’t duplicate the same product across different Performance Max campaigns, especially when audience signals in them roughly overlap.
Breaking Down a Real-World Example
Let’s take a specific product — it’s very telling. This item was present in as many as four campaigns at different times, now it’s down to two. At first glance, if you look at the aggregate numbers — it seems not too bad.
How many calls and sales will I get by ordering contextual advertising from you?
Let’s peek into one of the campaigns where this product is currently running. Over the last seven days — 21 conversions, ROAS 4.85. Seems like a great result, makes you want to celebrate.
But extend the period to 30 days — and the picture is less rosy. Somewhere the ROAS was acceptable, but elsewhere it dropped to frankly weak values. There’s no stability here whatsoever.
Now let’s switch to the first campaign, where this same product was top-performing by conversions. And here’s what we see: exactly at the moment when the item started “taking off” in the second campaign — it simply faded in the first one.
So this isn’t a situation where conversions from one campaign add to conversions from another, and you get double the profit. It works differently. When the target ROAS was raised in the main campaign — the product there “deflated.” But in the second campaign, where the ROAS remained lower, it started growing. Essentially, PMax simply shifted the volume from one campaign to another, rather than multiplying it.
It turns out that duplicating products in PMax accounts doesn’t bring additional sales — campaigns start competing with each other and pulling traffic, like two sellers at the same market elbowing each other for the same customer.
Second Example: Product in Three Campaigns Simultaneously
Let’s take another product for clarity. This item was active in three PMax campaigns at once.
How many calls and sales will I get by ordering contextual advertising from you?
Looking at the third campaign. An interesting point: up to a certain time, this product wasn’t getting any impressions at all — zero, silence. It only “came to life” when the top campaigns started raising their target ROAS. In fact, the result in this campaign is mediocre — ROAS 5.25, only two conversions. I would pause this item here, because the numbers are clearly below expectations.
Moving to the second campaign, where this product was also running. What’s here? Over the last seven days, clicks dropped. There was a short burst of activity — right when other campaigns were ramping up their ROAS. But the result — 9 conversions with ROAS 2.56. Definitely pause this one.
And now let’s peek into the top campaign, which previously generated the most conversions for this product. As soon as the client raised the target ROAS to 400% — the product simply stopped converting. The overall campaign ROAS looks fine, but this specific item “leaked” into other campaigns where conditions were softer.
Third Example: Coincidence or Pattern?
There’s a good saying: once is chance, twice is coincidence, and three times is already a pattern. So let’s break down one more product.
Let’s take one of the campaigns where it’s present. Here the ROAS dropped — 2.09 instead of the expected 4.78. Although for this specific item, the result is still tolerable. There was literally one day when metrics spiked, otherwise — steady, average performance. For now, this campaign can be considered more or less effective for this product.
But in the top campaign — same story again. After raising the target ROAS, the product item deflated. In this case, the logical decision is to exclude the product from the top campaign and leave it where it’s still delivering results.
Why You Shouldn’t Duplicate Products Across Different Performance Max Campaigns?
All three examples show the same picture. When the same product overlaps across several PMax campaigns with similar audience signals — chaos begins. You raise the ROAS in one campaign — the product fades and “flows” into another where conditions are softer. But there it often performs worse, because the campaign isn’t adapted to it.
If you do allow product overlap in Performance Max, you need to build clear impression priority. And with identical targeting without regular product-level analysis, this is practically impossible to control. This is exactly where the main difficulty of working with such projects lies.
