Pinning headlines in Responsive Search Ads: how many pins is too many?

Pinning gets recommended mostly for keeping a brand name in slot one or a compliance line in place. Google's own docs warn that pinning limits how the system can combine assets, and people argue about whether it hurts performance. I think pinning one or two headlines where the wording really can't change makes sense. Pin all three headline slots and you've basically got a manual ad with far fewer combinations to test.

How much that costs depends on volume. A small account rarely gathers enough data to find the good combinations anyway, so the loss is mostly theoretical. A high-volume account gives up real testing, so I'd pin only what has to be there and let the rest rotate. Has anyone compared performance before and after removing a pin?

Agreed on the one or two rule. For a plumber or clinic, I'd pin whichever line a caller can't afford to get wrong, like the service area or the "24/7 emergency" line, and let the rest rotate. The thing that trips people up is the Ad strength rating. Pinning all three headlines usually drags it down and it looks alarming in the UI, but it doesn't predict clicks or calls. Judge a pin by conversions over a few weeks, not by the color on that meter.

Judging a pin by conversions over a few weeks only works if the account gets dozens of conversions a month. A plumber or a small clinic might see three or four, and a few weeks of that tells you almost nothing. A pin that looks like it's costing you 10% could easily be noise. For accounts that small, I'd go by the asset performance labels and how often each headline actually shows, and give the pin a longer window before making a call. Agreed that the Ad strength meter isn't a verdict, though. It's a hint about variety.

For small accounts I'd keep it simple: only pin a headline if the wording is non-negotiable, like a legal line or a service area. Everything else, unpin. At three or four conversions a month no test is ever going to prove the pin helps, so the call has to come from what the ad needs to say, not from the data.

I don't fully buy "the call has to come from what the ad needs to say, not from the data." You're right that three or four conversions won't prove a pin helps. But the asset report tells you something well before conversions show up. If a pinned headline is soaking up most of the impressions, the other two are barely getting a fair run, and you can spot that without waiting a month.

Gut feel also leaves no record. So when you pin something, write down the date and the reason right then. When a client asks why cost per lead moved, you'll have an actual answer instead of a guess.

I'd push back on reading impression share as a warning sign for pins. A pinned headline holds its slot by design, so it takes most impressions anyway, and the unpinned ones look weak for that same reason. Log the pin, sure, but judge it on whether leads hold up.

Agreed, impression share mostly tells you about the pin, not the copy. The combinations report is more useful here because it shows which headline pairs actually served together. If each unpinned headline shows up in a decent spread of combinations alongside the pin, the rotation is working. If one of them barely appears next to it, the pairing is the problem, and you can fix that by rewriting the unpinned line instead of dropping the pin.

The combinations report has the same volume problem as conversions. In a small account a headline might show up in two or three combinations all month, so a thin row doesn't prove much. A decent rule of thumb is one pin per ad, only for wording that genuinely can't change, plus a review date about 30 days out. If the unpinned headlines still barely appear, rewrite them before adding a second pin. Each extra pin shrinks the combination space faster than it adds control, and at that point you're basically running a manual ad with fewer tests.