As an advanced Google Shopping user, you know performance varies considerably depending on the product and search query.
However, most people just hand over the management of search queries to Google and use its black box bidding.
If you want the best results and want to take back control from Google, you need to get the structure right to allow you to bid based on each product and on the search query – this is where query sculpting comes in.
What is query sculpting?
Query sculpting is an advanced Google Shopping tactic that gives you back a lot of the control that Google has taken from Google Shopping users.
It is, in effect, a way of structuring your Google Shopping campaigns so that you can bid based on the type of search query as well as the individual product.
The diagram below shows the performance based on search query for a product.
You will see that generic terms perform the worst and branded terms have the best performance. Query sculpting uses negative keywords to filter these terms into different campaigns where you can bid to performance.
You effectively create two duplicate campaigns and add the branded terms, in this case Adidas, as a negative keyword in one campaign.
This campaign then only picks up non-branded terms that perform worse, so you can set lower bids.
The other campaign picks up the branded terms and should have higher bids, because the performance is better.
Three dimensions of Google Shopping
The most simplistic way of explaining how to structure your account to filter queries effectively to be able to bid on individual products comes from Viljem Pitako, who talks about the three dimensions of Google Shopping.
Source: Smarter Ecommerce
Dimension 1: Product SKUs
To give you the most control, you need to be able to set bids for each product. The way to do this is to use what is called a single product ad group (SPAG) structure, where each product has its own ad group, bid and negative keyword list.
Setting up this structure can be time-consuming, so we suggest you turn to automation to do this. At DemandMore, we use a tool called Optmyzr to create these automatically at scale.
Because campaigns have a limit of 20k ad groups, you will want to create different campaigns for different product groups, which is where having a well-optimised feed comes in.
Dimension 2: Devices and audiences
The second dimension that you can control is bid modifiers for each device, audience, time of day and day of the week.
You can set the bids manually, but at DemandMore, we have opted to automate this process using Google Ads scripts, to save time and to provide more regular optimisation than humans.
We suggest using this script to set bid modifiers across devices and audiences and this script to change bids 24 times per day, making your campaigns more responsive than off-the-shelf bid management platforms such as Kenshoo and Marin.
Dimension 3: Query sculpting with negatives
Query sculpting is the final frontier in taking back manual control of your Google Shopping campaigns. Using the priorities within Google Ads and a set of negative keywords, you can filter certain types of queries into certain campaigns based on how likely they are to convert.
There are two main ways that you can sculpt search queries: you can do it manually using a method that Martin Roettgerding devised, or you can do it programmatically using Google Ads scripts to filter queries.
Manual query sculpting
Query sculpting as a concept was created by Martin Roettgerding, and over time, different people have refined the strategy and put their own personal spin on it.
Martin’s model is quite simple and uses negative keywords and different priority levels to filter different types of search queries into different campaigns. It works by instead of using a single shopping campaign, creating three with different intent levels.
- Product Campaign: This is focused on picking up search terms that contain the product name within it. Going back to the Adidas shoes example from earlier, this campaign would pick up terms such as “Adidas swift run shoes”.
- Brand Campaign: This is focused on picking up search terms that contain the brand terms. So, in the Adidas shoes example, this would be any terms that contain the word “Adidas”.
- Rest Campaign: This is a catch-all campaign that picks up any terms that are not specific to the product or the brand. An example here would be search queries like “Men’s running shoes”.
The diagram above shows how the model works with the three different campaigns. One thing to note is how you should use priorities and bids based on the search query.
You will expect your product campaigns to perform the best, as the terms are specific to that exact product, so you should bid the highest. You should set high bids for the brand, as these will also perform well, and then low bids for the ‘rest’ terms that are likely to perform the worst.
Priorities should also be used to filter the queries down into the right campaigns, with the products campaigns getting the lowest priority and ‘rest’ getting the highest priority.
This is a basic model that works effectively to bid based on search queries, and we have seen some success with this model at DemandMore when running clients’ Google Shopping campaigns.
For a more advanced approach, we suggest programmatic query sculpting.
Programmatic query sculpting
Being an agency focused around automation, we couldn’t help but add this method of using Google Ads scripts to filter search queries.
This method allows you to effectively create exact match Google Shopping campaigns.
You take a list of all of the exact match keywords that you want your Google Shopping ads to appear for and add them as negatives to one of your Google Shopping campaigns.
This Google Shopping campaign then effectively only picks up terms that are not on your list, and therefore you expect not to perform as well, so you downweight the bids. We call this the broad match campaign.
You should then have a second duplicate campaign, called the exact match campaign, that picks up the exact keywords that you want to appear for.
Source: Search Engine Land
The only snag is that it can also pick up keywords that are not exactly on your list, which is where using Google Ads scripts comes in.
You first need to set up a Google Sheet with the list of keywords that you only want to appear for using these instructions here.
Once you’ve done that, you can set up the script to run daily and it does the following: reviews each search term that your Google Product ads have appeared for; checks to see if it is on the list within Google Sheets; and if it is not, it is added as a negative keyword.
Over time, you will find that your exact match campaigns only appear for search terms that are on your keyword list.
You can use this as a stand-alone strategy or on top of the other strategies that we have discussed in the article.
Search query sculpting is a long-term effective strategy for gaining back some of the control that advertisers have lost through Google’s black box bidding and automatic query matching algorithms.
For the best results, you should look to take control of the three different elements of Google Shopping within Google Ads.
You should always look to use a single product ad group structure. Using automation makes your life a lot easier here, and don’t forget to break campaigns down by category or brand, so you don’t exceed the 20k ad groups per campaign rule.
Using automated bidding through Google Ads scripts allows you to set bid modifiers for both audiences and devices, giving you additional control over when your Google Shopping ads show.
Finally, you should focus on using query sculpting, and for the most advanced users, programmatic query sculpting to effectively bid based on the intent of the search term the user has entered.