Dec 212010

Google Product Listing Ads (PLAs) Performance Data vs. Text Ads

Question: How successful are Product Listing Ads (PLAs) compared to normal AdWords Keyword Text ads and how much of the PLA traffic can be considered new by virtue of this recently debuted ad format?

Investigation: We looked at a client with both comprehensive AdWords keyword coverage as well as a major and stable PLA campaign. The AdWords account has over 50,000 active Google keywords (all receiving clicks on a regular basis), and has had an active PLA campaign on AdWords since November, 2010. For this analysis, we compiled data from Dec 1 – Dec 16. RKG has the ability to collect user search query data for each click, so we can generate reports showing what queries have led to clicks and orders in any program, including PLAs. To measure the extent of new traffic and orders stemming from PLAs, we must also study instances where AdWords Text ads were also present for the matching query, and decide whether or not PLAs were diverting traffic from those listings. Please note, the data below is based purely on queries delivering PLAs, and comparing those terms to positive keywords in the same AdWords account triggering Text ads. The PLA campaign represents 6.3% of the accounts overall AdWords traffic, so it is still a small channel when compared to the remaining Text ad campaigns.

To study the channel, we have classified PLA traffic into 4 groups:

1) - PLA Search Query does not match a Keyword in AdWords and Query has No Broad Match Text Ad clicks
2) - PLA Search Query does not match a Keyword in AdWords, but Query does have Broad Match Text Ad clicks
3) - PLA Search Query matches a Keyword in AdWords, but Query has No Text Ad clicks (inactive phrase, bid at low level, etc.
4) - PLA Search Query matches a Keyword in AdWords, and both PLA and AdWords Text ads Do have clicks

Groups 1 and 3 above can be considered PLA only traffic. These search queries did not generate any AdWords Keyword Text ad clicks, but did generate PLA clicks during the same date range. One might argue that group 3 above represents non-incremental traffic. That is to say, perhaps these ads captured traffic away from the text ads that also appeared (or may have appeared) alongside the PLA listings.

Groups 2 and 4 cannot necessarily be determined as incremental either, since we did receive AdWords Keyword Text ad clicks for the same search queries as PLAs. However, since both groups have clicks in the date range, each ad format appealed to consumers in their own regard.

Let’s combine groups 1 + 3 and 2 + 4 to measure any lifts or differences in performance for PLA groups. We’ll call groups 1+ 3 “PLA Only” and groups 2 + 4 “PLA & AdWords Ads”

The “PLA Only” group converts at a slightly higher rate than overall Text Ads, but at a slightly lower Average Order Value. Although it represents 37% of all PLA traffic, it is an ROI positive group, and does not compete with existing AdWords Keyword Text ads.

The “PLA & AdWords Ads” set converts at roughly the same rate and order value as queries matching AdWords Keyword Text ads. Text ad do account for about 2/3rd of the group’s clicks, however, so they still control the bulk of traffic for these matching search queries. The final question then is just how competitive are these two ad formats within this group? PLAs play in the same sandbox as Text ads, so we would expect some cannibalization, right? While it is very difficult to know for certain, we do have some data points that suggest that PLA traffic does not lead to lower AdWords Keyword Text ad traffic. First, Keyword Text ad clicks did not decrease once PLAs entered the ad landscape. For the same terms, AdWords Text ad click volume is still healthy, even amongst terms that were served right alongside of PLAs. Second, Keyword Text ad CTR remained flat when compared to CTR before PLAs were introduced. On average, PLA CTR was ~ 7% higher than the equivalent Text Ad during the same date range. (Perhaps a more appealing ad format for some queries?)

Conclusion: About 1/3rd of PLA traffic can be considered exclusive to PLAs, where no Keyword Text ads received clicks. This is very promising given that this is a well established AdWords campaign. While it is very difficult to determine if the other 2/3rd of the traffic is incremental, we haven’t found any evidence to suggest that it is overly cannibalistic. In fact, the data suggests there is some incremental traffic, and that might indicate the new ad format appeals to a distinct audience of consumers. As a final note, I would like to mention that while each new advertising channel presents new traffic opportunities, it is still extremely important to gauge the program’s ROI. While we have found it an efficient and effective space, your mileage may vary. It’s also worth noting that there are targeting options within Google PLA Walkthrough AdWords and the Google Merchant Center to make these PLAs more profitable.


13 Responses to "Google Product Listing Ads (PLAs) Performance Data vs. Text Ads"
Very interesting data, Matt. Thank you for sharing. Two questions: (1) Had this advertiser been running PLA's prior to the AdWords integration? (2) Are you utilizing the advanced feed-based targeting features here? Best, nicholas
Hi Nicholas, 1) No, this advertiser did not participate in PLAs when it was only available via the Google Affiliate Network. 2) Yes, we are submitting (and recommend using) the Google Merchant Center AdWords specific columns to classify and exclude certain items from Product Ads. In case others are interested, those details can be found in the link within the last sentence of the post, named "Google Merchant Center".
Thanks for clarifying! The apparent lack of cannibalization is encouraging to note, though I am wondering if it only appears that way because advertisers were already occupying those spots before your client's PLA campaigns launched. That point may be moot, as they are here (seemingly to stay) and generating results for advertisers.
Eddie says:
Thanks for the info... There is a lot to chew here. I appreciate your clarifications in the comments!
I think Google may very well be tapping into a different type of shopper with PLAs - a more internet savvy, price conscience shopper - and agree that there isn't much cannibalization. At a couple months into PLAs I am seeing that a higher percentage of PLA orders are new-to-file than any other channel (Google competitive search being the next biggest percentage). The PLA order volume is pretty low right now, but I think this is an encouraging start.


Check out what others are saying...
[...] This post was mentioned on Twitter by Rimm-Kaufman Group and others. Rimm-Kaufman Group said: rkgblog: Google Product Listing Ads (PLAs) Performance Data vs. Text Ads: Question: How successful are Product L... [...]
[...] Google Product Listing Ads (PLAs) Performance Data vs. Text Ads, Rimm Kaufman [...]
[...] difference today is that there is a bunch of new noise. Google now offers Sitelinks, Product Listing Ads, Product Extensions, Seller Ratings, and more! The number of controls has also increased [...]
[...] Mierzejewski did some research on our end to try to answer three key [...]
[...] Mierzejewski did some research on our end to try to answer three key [...]
[...] Mierzejewski did some research on our end to try to answer three key [...]
[...] has managed PLAs for dozens of clients and we were an early adopter of the product back in November 2010. PLAs have been a very strong product addition for [...]
[...] several months of preparation leading to the change and PLAs being a part of AdWords since 2010, confusion still looms about how to make the most of these ad [...]

Leave A Comment