Responsive Search Ads: pin headlines or let the system rotate them?

Pinning in RSAs is one of those debates with no clean answer. The case for it is control: if your brand name or a legal line has to show up in a specific position, a rotating setup can easily put a weak headline in the top slot. The case against is that every pin cuts the combinations the system can test, so you get fewer good pairings and less data to learn from.

My view is that you should pin only what has to appear and leave the rest free, as long as you have enough conversions for the system to learn from. Below that volume I'd pin more, because the learning is too slow to be useful anyway. Where people draw that line is the real question, and I'd be curious whether pinned ads have ever held up better than unpinned ones on the same keyword set.

I don't think the low-volume argument holds. Pinning more when conversions are thin locks in your guess at the moment you have the least data to check it. At that volume the asset ratings (Best, Good, Low) are mostly noise, so pinning doesn't give you clarity. It just removes the variation you'd need to compare. I'd leave headlines unpinned, let the rotation run longer, and judge on CTR and cost per conversion at the ad group level instead of trusting the labels. Pins are for hard requirements like a legal line.

Agreed on the ratings, and I'd push it further. A pinned asset's label mostly reflects the slot it sits in, not how good the headline is, so a "Low" on a pinned line doesn't tell you much. My rule is to pin only what a legal or brand requirement forces and nothing beyond that. Leave everything else free, and judge it by ad group cost per conversion over several weeks, not by the labels.

Depends which slot you pin to. Position 1 is the first line people read, so a pinned headline there can cost you real CTR even when the copy itself is fine. I'd check the ad group numbers before writing it off as noise.

Fair on position 1, but I wouldn't read an ad group CTR drop as a pin effect so quickly. Pinning cuts the combinations the system can rotate, so the drop could come from fewer strong pairings rather than from the pinned line itself. A clean before-and-after window on the same keywords helps, and splitting into a pinned and an unpinned ad group is better still, as long as each side gets enough impressions. Which of those two can you set up without touching live spend?

Two ad groups beat a before-and-after window, but the metric you judge on matters just as much as the setup. A pinned headline in position 1 mostly moves CTR, while losing good combinations mostly shows up in conversion rate, so a CTR drop alone can't tell you which of the two you're looking at. I'd decide on cost per conversion and fix a minimum conversion count per side before the test starts, around 30 each. If neither side gets there, the test isn't finished, whatever the CTR looks like.