by
David Kirkpatrick, Senior Reporter
CHALLENGE
At BabyEarth, a baby products retailer with an e-commerce website, paid search is an important part of the marketing mix.
When Google Shopping existed as natural search option, BabyEarth was active in optimizing organic search for its products. Once Google Shopping became "pay for play" and allowed merchants to improve results with product listing ads (PLAs), BabyEarth developed a strategy to take advantage of that marketing opportunity.
"As Google made those listings [PLAs] more prevalent — put them on the middle of the [search engine results] page — and then also added PLAs within the normal PPC listings and the text ads," explained Steve Steinberg, Chief Marketing Officer, BabyEarth. "It became more significant to try and capture more of the real estate on the page to make sure that you appear everywhere that you possibly can."
He continued, "That was the impetus for us to focus in that area."
This case study goes into the details of that focus to uncover the tactics BabyEarth used in its Google Product Listing Ads. The team’s strategy increased PLA targets from under 100 to more than 10,000 within the first month and generated an average monthly increase in PLA revenue of 129%.
CAMPAIGN
BabyEarth’s paid search also involves pay-per-click ads through Google Ads and the Yahoo/Bing Network.
Steinberg explained a PLA differs from a PPC ad in that it is primarily focused on an image, so an image result for a product or product category search will pull up a
PLA listing. Clicking on that listing takes the user to a
product page on BabyEarth's website.
Step #1. Focus on bidding and test to optimize the product images
Before executing the PLA strategy, Steinberg said it was key to ensure all of BabyEarth’s products were "in the feed" to Google. From there, the first order of business was to focus on bidding.
"We focused our bidding primarily based on price tiers so that we wouldn’t bid too much on lower-priced products, and too little on higher-priced products," he explained.
Steinberg added the team optimized and expanded the program by delivering PLAs to specific categories, subcategories and brands.
He said, "For example, if someone searches for strollers, a very generic term, we want to make sure to show them the particular stroller SKU image that makes the most sense."
Steinberg explained without testing to find the product targets that deliver the highest clickthrough and conversion rates, Google will arbitrarily choose the image displayed with the PLA. This chosen image might not produce the highest ROI.
The entire bidding process was manually adjusted on a daily basis based on impressions and performance. Bid performance was judged primarily on the "return on ad spend" metric.
Step #2. Expand the product targets
Steinberg said an important part of the overall PLA strategy was submitting all of BabyEarth's products to Google for listing, but taking that concept deeper involved creating groupings of product targets.
"Product targets are more of a group," he stated. "It's not only groupings of products so that certain SKUs will show up within a category, like strollers. It's also if a particular product has five colors, for example, and you want Google to show the red one because the red one is the most popular and the one that people are most likely to click on."
He explained a product target would be grouping those five colors in the same product target and then telling Google the red product image should appear in PLAs.
BabyEarth dramatically increased product targets from 100 to more than 10,000 through combinations at the product level.
The original product targets were limited by being centered on broad product categories. The team used a technology solution provided by its paid search agency to expand the range to SKU-based product targets. This expansion meant BabyEarth had one product target for each product in its PLA feed.
Steinberg said this expansion of product targets also ensured the ads delivered the right "color, pattern or style" of the correct product, including at the category and brand level.
Essentially, grouping the product targets allows BabyEarth to determine the basic PLA image in a less exact search, and also to have the PLA display the exact product searched.
For example, if someone is searching for a particular product, but doesn’t specify a particular color, BabyEarth determines the color in the product image displayed. If the search includes a specific color that BabyEarth stocks, the ad will show that exact product in the color specified in the search.
Beyond just expanding the product targets, the team also created aggregate targets by combining multiple related products into a single target. The goal of aggregate targets was to catch more impressions and also fill gaps in coverage.
These aggregate targets included:
- Brands
- Brands and Categories together
- Subcategories
- Parent IDs (for products with different colors or sizes with multiples SKUs)
Step #3. Perform keyword research and analyze search queries
Google uses the titles and descriptions within the PLAs to find searched-for keywords, although only the title shows up in search results.
"When you are using PLAs, the title and the description of your product that you feed to Google is where they pull out all of the search phrases that are applicable to the product," Steinberg explained.
He said keyword research on phrases to target, where traffic was going, and comparing that research to the competition provided the team with material to test for ad placement.
The idea behind the keyword research and testing was to make sure each product was properly segmented to the proper category.
Steinberg said BabyEarth used information such as product model numbers to achieve this segmentation.
The entire keyword segmentation process is similar to pay-per-click advertising, but the strategy does differ in PLAs.
Steinberg said, "When you're doing normal PPC text ads, you can just pick the key phrase. You can say, 'exact match,' 'broad' or whatever."
He continued, "[For PLAs,] you have to work that into the title and the description in such a way that Google is then most likely to present you on that particular search."
He added keywords are more heavily weighted in the title.
Analyze search queries
Part of the team's keyword research was taking in all the data around the PLA effort:
- Where the products show up in particular searches
- Number of impressions on a particular search
- Where the customer clicked
- Whether non-targeted phrases are getting clicks
Steinberg stated all of these data points are taken into the process of negating some key phrases, and also aggressively bid on other key phrases.
He said, "We might bid more aggressively because we realize that a particular phrase is truly a phrase where the customer intent is to purchase."
Step #4. Continue to optimize the entire process
"Optimization is important in this process now that Google Shopping is a paid platform — ensuring that you have the right product show up at the right time by analyzing all of the different segments that you’re trying to target," he explained.
Providing an example of BabyEarth's testing and optimization of its PLA program, Steinberg said the team would test different products showing up with more generic search terms, or testing the performance of the color of the product that shows up in the ad.
For example, the team would conduct an A/B split test on either a red version or blue version of the same product with the same set of search terms to determine which performed better.
Steinberg said for tests, both clickthrough rate and conversion to sale were tracked.
He said, "Just because you increase the clickthrough rate doesn't make that a positive ROI. You need to convert that sale as well.
RESULTS
Steinberg said with the PLA program the team looked at revenue, but also had a "big focus" on margin.
"We are trying to ensure that we're really looking at the overall profitability of the sales that we garner from PLAs," he said.
Steinberg continued, "If we wanted to be aggressive and care less about earning money, or breaking even on the first sale, we could have increase revenue even further. But, we try to ensure that [with] the products that we’re targeting, we're getting a positive ROI even on that first sale."
Some of the metrics from the PLA program include:
- Average monthly income from PLAs has increased 129%
- Year-over-year revenue growth from PLAs in August 2012 of 193%
- Year-over-year revenue growth from PLAs in September 2012 of 201%
- Increased PLA targets from under 100 to over 10,000 to promote entire product catalog
- Year-over-year clickthrough increase in August 2012 of 175%
- Year-over-year clickthrough increase in September 2012 of 91%
- Year-over-year conversion increase in August 2012 of 38%
- Year-over-year conversion increase in September 2012 of 13%
Steinberg said his main takeaway from the program has been trying to react and get ahead of the competition by utilizing testing and optimization to ensure BabyEarth is displaying the "right products at the right time" in PLAs.
He added with the switch from free listings to paid product search on Google, BabyEarth sees fewer competitors within this branch of search engine marketing.
For weekly email marketing case studies and how-to articles delivered straight to your inbox, sign up for the free MarketingSherpa Consumer Marketing Newsletter.Creative Samples
- Google Product Listing Ad
- Result of clicking on an ad
Sources
BabyEarthAdlucent — BabyEarth's paid search agency
Related Resources
Google as a Grocery Store: Use SEO and search engine marketing in tandem to boost lead generationPPC Advertising: 5 winning display ad tactics that increased paying customers by 2,900% and dropped cost-per-lead 37%Holiday Marketing: 8 trending tactics and tips for 2012New Chart: Search Engine Marketers Rank SEM ChallengesSearch Marketing: 46% more conversions from PPC ad test
