Over the past several years, Google’s Product Listing Ads (PLAs) have become a prominent part of many retailers' paid search programs. Since the full roll-out a little over three years ago, we’ve seen an evolution that included countless ad format tests, several feature additions, and a 2012 transition that moved Google Product Search to a commercial model built on Product Listing Ads. Fast forward to last quarter, when PLAs drove 42% of non-brand Google paid search clicks for RKG clients, and it’s clear that this program will be an important component of the paid search landscape for the foreseeable future.
PLAs are unique in that they don’t entirely exhibit the same qualities as either CSEs or paid search. We’ve seen them referred to as each, and in their current state there are merits to both arguments.
PLAs as a Comparison Shopping Engine
There are several characteristics of Google Shopping ads that look very much like a comparison shopping engine. They are feed-driven and the ads, with product titles, images, and pricing, contain a lot of the same information as a CSE ad unit.
Google PLA Units:
CSE Ad Units:
On the Shopping page specifically, users can sort and filter products and, essentially, comparison shop. On the management side, CSEs and PLAs are both partially at the mercy of an engine’s algorithm to determine what products will show since there is no bidding on specific keywords. Relevancy and bid determine whether your product is displayed.
PLAs as Paid Search
The argument for Google Shopping as paid search also has its merits. Like text ads, PLAs are part of a live auction and both can appear in the same ad space on the SERP. The searcher’s intent is also similar and less commercial than on CSE’s, since by definition someone on a Comparison Shopping Engine has indicated that they are in fact shopping, regardless of their query. In contrast, a user on the Google SERP is not necessarily there for commercial search.
The SERP does not allow sorting or filtering of products, and the selection shown is limited to the number of placements Google has served you(unless, of course, you choose to click through to the Shopping page). This is similar to text ads in that if an advertiser's bid does not place their product ad on the first results page, their ad will not be seen. CSEs, on the other hand, offer sorting options which most users utilize to sort the lowest price items first, such that bidding is more or less on/off and an advertiser's product can get seen as long as it is price competitive.
On the management side, PLAs also look quite different from a traditional CSE. There is no arbitrarily set rate card, with competition and demand dictating what you need to bid to show up. If a product target isn’t meeting your efficiency goals at its current bid, you can set a bid that makes sense based on your sales per click. For CSEs on a strict rate card model this is not an option – your product must either find efficiency at the rate card minimum or be removed from your feed entirely.
PLAs are also similar to paid search in regards to how advertisers' are able to refine traffic through keyword negatives. By analyzing search query reports for PLAs and adding negatives for poor performing queries, advertisers can reduce wasted spend and bid more effectively for relevant queries, much like with text ads. No such functionality exists within CSEs.
In terms of how quickly bid or product changes are processed, updates for Google PLAs are processed within the hour, similar to how quickly text ads are updated in paid search, while CSEs can take up to 24 hours to update.
These factors are what lead our team at RKG to treat PLAs more like non-brand paid search than a comparison shopping engine.
Managing PLAs Like Paid Search
We place high importance on bidding for PLAs because results are heavily impacted by your maximum CPC. Bidding options on CSEs can be quite limited, with some engines allowing only high-level category bids broken into 3 price bucket levels. PLAs, however, offer a wide variety of bidding options.
In a realm where keywords don’t exist, bidding smartly relies substantially on building a robust set of targets that range from All Products to ID level and cover everything in between:
By creating relevant clusters of products in multiple layers, we are able to apply different bids to segments of products that perform in different ways. It is important to consider the differences, for example, between full-priced women’s Seiko watches with $10 shipping and those that are currently 25% off and ship for free.
Even if these groups don’t end up needing different bids, it is key to think about the potential performance differences and investigate the data accordingly. We couple detailed bidding strategies with strong focus on feed optimization to build a solid campaign for PLA success.
Regardless of whether you think about PLAs as a CSE or paid search, the most important thing is to make sure you’re thinking about them and making efforts to build a strong program, from feed to structure to bidding. Check out the RKG Digital Marketing Report in April to see how trends looked for Q1!