Google Smart Shopping Campaign has been getting popular because it makes marketer’s work easier than conventional campaign management. For example, many marketers use Shopify which has a feature to kickstart Smart Shopping Campaign really easy like this.
With Google Smart Shopping Campaigns, your product data feed and assets are leveraged with Google’s machine learning to show various ads across ad networks. Setup involves integrating with Google Merchant Center, setting a budget, prep assets, and tell Google the country of sale. Google will pull information from your product feed and test different combinations of the image and ad copy you provide. It then shows the most relevant ads across Google networks (Google Search Network, Display Network, YouTube, and Gmail).
Following is our paid media use case with Google Smart Shopping Campaign where we helped our client quadrupled their sales and ROAS. Shopping Campaign has the potential for many eCommerce sites to boost their sales.
In this blog post, we would like to share some interesting execution from our data management deployment.
There are two important points in the Smart Shopping Campaign that effectively make use of machine learning:
Once it’s set up, you can just set and forget Smart Shopping Campaign. However, if you want more out of it and drive better results, you need to constantly improve the following.
You can use any kind of product data as a product feed as long as the data meets the requirement of Google Merchant Center.
That is exactly why a lot of eCommerce platforms like Shopify have a feature to create product data feed.
However, in most cases, these native data feeds are suboptimal because of these reasons:
Fewer impressions, clicks, and conversions mean lost opportunity for optimizing an ad because machine learning and it’s model will improve as the data point and sample increase.
We have faced such challenges, but we had to learn to overcome those constraints to optimize our client’s ads. So here is how we approached the problem and situation.
Before we start, let’s quickly review what stage of optimization you are currently on:
Stage 1: Fulfill the minimum requirements of Google Merchant Center
Stage 2: Frequent and automated feed refresh
Stage 3: Include optional fields to the product data feed
Stage 4: Creative and Product Category optimization
Stage 5: Tailored and Systemized approach to managing feeds
If you are not on Stage 4, we suggest you review and complete Stage 1, 2, and 3.
Sometimes it might be difficult to know if you have satisfied Stage 3 because you just don’t have the optional data points to add (For example, cost_of_goods_sold, unit_pricing_measure…etc.),
Here is a reference that is helpful that lists the specification for the product data feed.
If you have the basic fields required to export from your product database into your product data feed, you can proceed to Stage 4.
Most of our clients and use cases start in Stage 3 because they have satisfied Stage 1 and 2. Most marketers stop optimizing the feed for Smart Shopping Campaign at Stage 3. That is unfortunate and a lost opportunity for many marketers.
In the case of typical data feed ads, you can get the best practices from the data feed platform (feed management tools).
Here are some basic points you could check and improve your data feed.
From Google Shopping Campaign perspective, these points help to increase impressions:
Sometimes improving these points leads to 2x impressions so it is worth a try.
In the success case, we also did this using a data feed management tool.
Data feed management tool helped us a lot, we could probably attribute half of the performance lift to the feed management tool vendor. Because once we enhanced the upper funnel, the next step would be the middle of the funnel.
If your Smart Shopping Campaign performs well, you can stop tinkering with the product feed and get ready for Step 5.
Some enterprise-grade feed management tools like Feedonomics, you can improve your product data feed even further.
Here are some features Feedonomics offers that you can take away knowing these ideas could further optimize the feed and ad performance.
Now, with these enhancements, we have gotten more impressions and clicks. The next step is to prepare and properly send user behavioral data to Google Ads with tags.
This is where a typical tag management issue happens. Everyone wants to start a Dynamic Remarketing Campaign or Smart Shopping Campaign. But as soon as they start the campaign, they soon realize the campaign underperforms. Some of those disappointing measures are bounce rate, pages per session, etc.
Few reasons why Dynamic Tag goes wrong:
Our client went through these situations prior to Principle fixing the miss-match of SKU and Parent ID that wasn’t firing according to user behavior timing.
Key Thing to know: in case of an eCommerce website, you can send either SKU or Parent ID as “id”.
For one of our clients, when we started dynamic ads, the bounce rate of the campaign for specific SKUs seemed much higher than expected.
So our consultant dug deep into our client’s data feed and found that it included many SKUs that have no inventory in their warehouse. Since these products are sold out and these products are hard to refill inventories, such SKUs showed up in product data feed as out of stock products like below.
The product detail page was sending Parent ID (“ABCD”), meaning a customer who wants a specific SKU without inventory, Google Ads knows that this particular customer is interested in “ABCD”, hence explains why Google shows a banner of given products.
Our consultant analyzed the bounce rate for each product SKU and noticed SKU that has no inventory tends to have a high bounce rate, which was aligned to the consultant’s hypothesis.
Given the situation, here is what was happening for this customer:
Since we want to avoid this kind of situation, we suggested adding a tag that sends SKU (“456”) after steps 3 (at the point that the customer selected SKU). Here are some more tips:
If your product detail page has just one SKU, it should be fine as the case above.
When you have multiple SKUs in a product detail page, then:
As you can see, Dynamic Remarketing tag is actually quite technical and sometimes changes take longer than you’d expect. If you are going to invest in Google Shopping Campaign, we recommend thinking about these three aspects.
Senior Data Engineer
Google Marketing Platform expert with experience working with enterprises deploying marketing analytics capabilities. Guy is focused on delivering marketing technology and supports companies drive their marketing execution through data.