Despite years of sweat, blood, and heavy calculations, even the most seasoned Search Marketing / Words Go Here vets can sometimes make a mistake (shh!), or run into a bewilderingly difficult anomaly.

Several months ago such a unpleasant scenario befell me, and I thought it best to share with the WGH Community…

I took on a legacy PPC account that was in some, shall we say, disrepair.  In an effort to right the sinking and burning ship I began restructuring the layout of things, following segmentation best practices — similar keywords into separate ad group buckets, ad copy reflecting keywords, etc.  Pretty basic by all standards.  One of the main portions of restructuring that needed to happen was in the branding campaign.  Over time there had accumulated a number of branded keywords with variants and stems that could be categorized, but had been instead left within one ad group.  Naturally, I separated the one ad group into a more reasonable three groups, and added more keyword rich ad copy.  And as this was a different structure on an old account, I made sure to up the CPC bids a bit just to give Google some extra love while getting the campaign over that expected initial re-evaluation hump.

Campaign relaunch.  Conversions plummet.  Freak out.  Panic.  Governments collapsing.  Babies sold on the black market.

The company had (and has) a very strong brand presence within their field, so a sudden and inexplicable drop in their brand conversions was making a lot of people antsy.  Understandable.  Naturally the only culprit in the mix was the restructuring I had done.  But what had I done?!  What, having followed best practices, could have caused such a tanking?

The branded keyword Quality Scores of 8 to 10 had all pretty much dropped to 4 and 6!  I reached out to a few friends at top agencies, as well as Google, to see if they had thoughts.  They were just as baffled.

In the end, I figured “at least I can drop the bids back down so we’re not burning through branded dollars even more than usual.”  I put the bids back to or under their previous ranges, and in less than 24 hours the Quality Scores were right back where they’d been originally, in the 8 to 10 range.  Confused?  Yup, me too.

After running it around again with some peers, the best we can all figure is that the Google algorithm watches certain tolerances in % change for CPC’s on an account and freaks out if those tolerances are exceeded.  A better way of putting it is, when the Google algo sees something that seems way out of whack with expectations — which branded terms can set much easier than some of the larger more general terms — then it will suspect foul play / spamming and ding it.

I know, I know… it’s crazy.  It doesn’t make sense.  Or does it?  You establish a precedence of bidding on your branded terms, and suddenly you change the long-standing structure of that segment of your account, while upping the bidding element, and suddenly “you might be a spammer!”  Guilty until proven innocent?  After much thought, that’s the best our internal and external discussions could make of the situations.

If you’ve run into a similar situation, please share it with us.  This sort of thing is a “ghost in the machine” situation, but if we all have an occasional situation we can contribute to the pile, we’ll all be more ready to react when something peculiar occurs.