Did Google's 'Not Provided' Decision Slow Down Brand Bidders?

July 1, 2015

By now, nearly everyone in the paid search industry has heard about Google’s decision to stop passing query data in its referring ad URLs. After an initial panic, many of us were brought back to our senses by some well-reasoned posts from industry experts. Among them was WordStream’s own Larry Kim, who explained very clearly that search query data isn’t dead. Advertisers will still be able to pull this information from the Search Terms Report and through third-party platforms that use the AdWords API.

Figuring Out Who Was Impacted by Google’s Policy Change

So if the impact is pretty minimal for most advertisers, where are the effects being felt? After asking ourselves this question, we developed some criteria to evaluate whether an advertiser’s campaigns would be significantly hampered by the recent changes. Our list quickly narrowed as we hypothesized. After all, why would an advertiser specifically need the search query data to be held in the referring ad URL? Wouldn’t a tool like the Search Terms Report or a third-party platform such as WordStream meet their needs?

Below are the basic indicators that we ultimately identified. To feel a major effect from this decision, an advertiser would have to be:

  1. Targeting a wide range of keywords in each campaign - most likely by broad matching and using Dynamic Keyword Insertion to increase the relevance of their ads.
  2. Spinning content on their landing page - by passing the query string from the referring ad URL.
  3. Running many different campaigns - otherwise, it would be pretty easy to modify the match types within a single campaign.

In other words, exactly the kind of behavior you might expect from brand bidders.

Wait, What Is Brand Bidding?

Before jumping into this post, I should probably clarify my terms. “Brand bidding” may sound like I’m talking about an advertiser placing PPC ads on keywords related to its trademark (for example, Nike bidding on search terms like “Nike” or “Nike shoes” on Google). However, while this is a valuable topic and has even been covered right here at WordStream, it’s not what I’m referring to.

I’m using “brand bidding” to describe situations where an advertiser other than the brand is placing ads on that brand’s trademarked terms. For example, if Puma were placing ads when people searched for “Nike,” I’d say that Puma is brand bidding. Similarly, if a coupon site places ads on “Nike discount,” the coupon site is brand bidding.

Here’s a more visual example of brand bidders targeting AllState. I’ve marked AllState’s own ad in green, and the two brand bidders in red:

what is brand bidding

In this case, the two brand bidders are actually marketing partners of AllState. They may be allowed to do this, or they may not. I don’t personally know for sure. Sometimes brands will allow select partners to bid on their brand so that they can block out competitors on the SERP. However, in many other cases, brand bidders can harm your own PPC campaigns. They can divert traffic to their own sites and even increase your Cost-per-Click, ultimately making you pay more for fewer visits.

I could certainly go into more detail, but I think that should be plenty for our purposes here. Just remember that brand bidders are the other advertisers who show up when people search for a particular brand.

So, What Does Brand Bidding Have to Do with ‘Not Provided’?

Okay, we’ve detailed some criteria above. We know a little bit about what type of advertiser we’re looking at: someone who’s running a bunch of broad campaigns that pull query strings onto landing pages and probably use DKI. What does this have to do with brand bidding?

Well, over time BrandVerity has become quite familiar with certain advertisers who seem to meet our criteria above. In particular, we typically find Comparison Shopping Engines, or CSEs, running ads that appear to pass along query strings to their respective landing pages. These CSEs’ ads reproduce searches on their own sites, taking the user’s original search terms and placing them into the search box on the comparison shopping engine’s own landing page. Let’s walk through a quick example to illustrate the setup.

comparison shopping engine

The sequence above goes like this:

  1. User searches for “adidas shoes” on Google
  2. PriceGrabber places an ad featuring “Adidas Shoes” in the title
  3. PriceGrabber’s landing page features a search for “adidas shoes”

We see the same string move from place to place, ultimately ending up on PriceGrabber’s landing page. So, how did it get there? Our suspicion was that PriceGrabber and other CSEs were taking advantage of Google’s search query parameter and passing it through to their landing pages. This would explain the prevalence of CSE ads that we find recasting searches on their own sites. Plus, it would also meet our original criteria for advertisers that would be heavily impacted by Google’s changes. By making a before/after comparison of comparison shopping engines’ ads, we could start to see whether their campaigns were truly affected.

Were CSEs Actually Passing Query Strings?

Once we had our hypothesis it was time to start testing. To determine the impact of Google’s change, first we’d need to know what the CSEs had been doing in the past. Fortunately, some of our monitoring turned up a relatively continuous stream of CSE ads dating back to December of 2013. Within that data set, we focused on a subset of 12 popular CSEs.

So, what did we find? Let’s start by revisiting the PriceGrabber example from above. That particular ad was placed on March 30, 2014 (before the new restrictions came into play. You can see that the query string was still present in the referring ad URL:

Here we have an ad that could benefit from passing along the query string (higher landing page relevance) and had the chance to do so. That’s motive and opportunity—but what about the means? It sure looks like it’s passing along the query string, but we can’t confirm anything quite yet. The missing evidence can be found in a number of ways, but one key indicator here would be match type. If the advertiser were using exact match, there would be no reason to pull the search query from the referring URL. The advertiser would already know it! Depending on how the campaign was set up, you could either A) “hard-code” a specific Ad Group to lead to a landing page that searches for “adidas shoes”, or B) use the ValueTrack keyword parameter to pull “adidas shoes” onto the landing page.

If we look at the PriceGrabber example a bit closer, here’s what we see in the landing page URL:

brand bidders

There we have it: exact match. With that setup, PriceGrabber can easily pull in the search query through the {keyword} ValueTrack parameter. We actually see them doing exactly that in the string above. So, in this particular example, the trail ends here. The CSE seems to be pulling the query string through alternative means.

What If CSEs Had Already Changed Their Tactics?

Of course, it’s worth remembering that our example above occurred only a week and a half before Google rolled out the update. Perhaps some of their more prominent advertisers had already been notified of the changes. Let’s look further back and see if we can find anything different. Here’s another example from PriceGrabber, this time targeting the merchant Genius Chargers:

Brand Bidding Examples

Again, things start to look like they’re lining up. The comparison shopping engine plants branded keyword in the headline, then copies the search over to their own site. Seems as though the ad might be passing the query string through. But I didn’t have to look much longer to find evidence to the contrary. Fortunately, since I started looking at an even larger time window, I had a larger pool of ads to analyze. I was quickly able to see that this particular ad had actually been triggered by multiple keywords. Here’s a BrandVerity screenshot showing exactly which keywords it appeared on:

Brand Bidding for Keywords

Each of those 6 keywords on the left resulted in the same PriceGrabber ad. The same copy. The same landing page. The same search query on that landing page. From that evidence, we can see that PriceGrabber could not have been passing query strings in this case either. Otherwise, the search on the landing page would have varied instead of consistently showing “Genius Chargers” in the search box. Furthermore, after looking through the landing page URLs more closely, it seems that PriceGrabber was simply running the same static ad and broad matching it on an array of searches.

What about Other Advertisers?

I’ve deliberately focused on a specific advertiser to make this piece as clear as possible. But you may be wondering about the others. After all, one advertiser’s results won’t necessarily speak for all the others. So, what did we notice across our set of data?

After analyzing ads from a dozen different CSEs, we were able to find evidence from each of them to suggest that they were not passing search queries in their ads. We looked for signals similar to the ones we discussed in the PriceGrabber examples above, focusing on ads placed before Google’s policy change. While we couldn’t find a counterexample for every last combination of advertiser and keyword, we believe that the counterexamples we did find are strong indicators of how a given advertiser’s AdWords account is set up. In other words, even if we only found evidence to prove that Sale-Fire was not passing query strings on Adidas keywords, it’s still a good bet that Sale-Fire wasn’t passing query strings on Nike keywords either.

We also made sure to look into some advertisers other than CSEs. In particular, we wanted to check on search arbitragers (for a quick background, check out this post from Flying Cork Media). Interestingly, although we selected some of the most notorious arbitragers in paid search, we found that they were not passing search query data either.

The Takeaway for Brands

Based on our findings here, it seems that brand bidding will stay relatively constant. I hate to be the bearer of this somewhat bad news, but there’s no compelling evidence to suggest that Google’s recent policy change will have any effect to deter brand bidders’ campaigns. For now, it looks like everything will be status quo.

However, please remember that our study was limited to a specific set of brands, keywords, and advertisers. It’s very possible that other CSEs were doing things differently—or that these same CSEs were advertising differently on different sets of keywords. Over the next few months, it would certainly be interesting if any of you experience CPC drops on your branded keywords—or anything else that might indicate a decrease in brand bidding. If you happen to notice any new trends, I’d love to hear about them!

This post comes from Sam Engel of BrandVerity. Sam is BrandVerity’s Marketing Manager, and enjoys exploring the trademark and brand issues that come up in paid search.

Sign up to get our top tips and tricks weekly!