In a recent study by Adobe, Product Listing Ads (PLA) accounted for over 14.7%* of total paid search spend in December 2012. To put it in perspective, that is more spend than the entire Yahoo-Bing network (13.5%). The percent of PLA spend has steadily increased as more and more advertisers focused their efforts on this channel. At DataPop, we’ve seen PLA’s account for as much as 30-50% of total paid search spend for retail advertisers.
One area marketers have yet to figure out is the interplay between PLA’s vs. Paid Search. How can you make both ad products compliment each other without cannibalizing performance in more efficient areas?
Same SERP, Different Performance
When looking at topline metrics, the largest deltas we see from PLA’s and Paid Search are in CTR and Average Order Value (AOV). On PLA’s, CTR is 34% higher and AOV’s are 12% lower. There are a couple factors at play here.
CTR is higher driven by the contextual nature of PLA’s. It includes all of the relevant information a consumer is looking for: Picture, Price, Brand, Models. This gives consumers better ability to choose the specific product they are looking for, at the right price.
Price transparency leads directly into the next stat: Lower AOV per an item. When consumers are given a list of products to choose from, all else equal, they tend to choose the product with the lower price point or stronger promotion.
As a savvy marketer, you need to understand these stats in order to take full advantage of PLA’s. From a CTR perspective, you need to have a clean and optimized product feed in order to create titles, descriptions, and promotions that appeal to your consumer base. From a competitive standpoint, you need to understand that price transparency plays a key role in how well you will compete or not within this space. You need to make sure the right products are showing up for certain search queries. This is done by having a granular structure, tiering your bids, and appropriate use of negatives.
Base Your Optimization On Data
It is important that your Paid Search and PLA programs are in perfect unison with each other. Since they both play in the same space, there needs to be proper allocation of bids and budgets being set. You don’t want one ad product to cannibalize the other if it is less efficient.
This has to be done on a query by query basis since PLA’s are driven by product feeds and Paid Search is driven by KW’s. This can be a cumbersome project but there are a few ways you can address it in an efficient manner.
The 80/20 Rule
You’ve all heard this before. 80% of your traffic comes from 20% of your queries. Same thing applies for PLA’s and Paid Search. Focus on the 20% of queries that drive the majority of your volume. Compare your goal KPI’s across individual queries and how they perform on Paid Search v.s. PLA’s. If one performs significantly higher than the other, there should be higher bids and budgets placed on it.
If you want to set a standard default for your entire PLA and Paid Search campaign, you can take a look at top line performance for the entire program. The delta between the two programs on your goal KPI will give you guidance on how you should tier your bids and allocate your budgets. For advertisers looking for more traffic, PLA’s helps you drive more volume and sales for any given query (Higher CTR’s and Conversion Rates). For advertisers looking for more efficiency, although conversion rates are higher on PLA’s, AOV’s tend to be lower. There is a counter balance of these two KPI’s on ROAS. A lift in conversion rates will help offset any decrease in AOV’s. If you notice both conversion rates lower and AOV’s lower, you have to ask yourself if the increase in traffic is worth the decrease in ROAS.
Keep in mind, in order for you to tier your bids appropriately within PLA’s, you MUST have a granular structure (since bidding is done at the product target level). This doesn’t mean that all of your bids have to be at the SKU level, but it does mean that each ad group should have products in them that have a similar margin and performance.