What are product feeds?

A product feed is a form of structured data that contains detailed and up-to-date information about your products. Product data feeds provide platforms like Google and Facebook (or anywhere else that you are serving shopping ads) with all of the information they need to generate those product listings. 

Your product feed defines your product ads, such as those screenshotted below.

product feed screenshot from Google

Product feeds can be in XML, CSV, TXT form, for example.

You would then submit a product feed to a platform such as Google Merchant Center, or Facebook Commerce Manager. This connects with your ad account in order to turn your product feed into ads.

Product feeds for Google Ads

You set up a Google Ads product feed within Google Merchant Center.

Google Merchant Center requires product data to be in a specific format in order to process it correctly. Google provides a full list of the product data attributes and values that you can provide here

For the human user, the Google Sheets product feed template is very useful for understanding how product feeds are structured. 

But, there are a variety of alternative ways to submit your product feed to Merchant Center. You can

  1. Submit a feed manually through Google Sheets
  2. Scheduled fetch - host a file on your site and point Merchant Center to it
  3. Automatically upload a product feed into Merchant Center
  4. Set up Google Merchant Center’s Content API and sync it with your product feed.

Why are product feeds important?

Your Google product feed defines a large proportion of how your product ads perform. It sets up not only how your ad looks, but also who your ad is deemed as being relevant for.

For example, Google uses your product feed to decide what keywords each product might be relevant for. In other words: Your product feed defines your keywords. 

Attributes such as “gender” and “age” define what demographics your products are suitable for.

Google also uses the product feed to populate your ad. From images and ad copy to shipping rates, nearly everything is defined by your product feed.

Let’s take another look at those shopping listings, but with all of the elements labelled. In green are those defined by the product feed.

product feed - shopping listing with all elements labelled such as product title, price, sale price, product image and brand name.

Your product feed also automates your ads, taking a product ad down if, for example, you have run out of stock. 

Finally, your product feed also tells Google what products are similar to others, better enabling dynamic remarketing.

We asked PPC marketers around the world to mark from 0 to 10 how important product feeds were to the performance of their shopping ads. They gave it an average of 7.8 out of 10.

Product feeds are incredibly important to the performance of your Shopping Ads on Google. 

Optimizing your product feed

Despite knowing how important they were to performance, the same group of marketers said that they only spent 22% of their time on Google product feed optimization.

The greatest barrier to optimizing product feeds, they said, was a lack of understanding of what actual changes they could make to improve product feed quality. So, here we focus on giving you actionable product feed optimization tips.

1. Test your core attributes

As with most things in Google Ads, the key to product feed optimization is to set up tests to measure their impact.

Optimizations that work for some ecommerce businesses, won’t work for others. Some changes might be a lot more impactful within certain sectors than they are for others. 

The first tip is therefore that you should set up A/B tests for all of the following tips.

For example, like the test below that details changing the wording to highlight exclusivity, changing the image, changing the brand description, and highlighting the gender it’s targeted at: 

product feed optimisation A/B test that details changing the wording to highlight exclusivity, changing the image, changing the brand description, and highlighting the gender it’s targeted at

Always prioritize the tests to identify the most important to run first.

How to test product feeds

Just as you would test PPC ad copy, you should be looking to test key attributes of your product feed. Examples of this include your product titles and images.

You can set these up through one of two options:

1. Testing with feed optimization tools

If you have the budget, feed optimization tools are great ways to both manage your product feeds, and to easily run Google product feed tests.

Examples of these tools include Productsup and feedoptimise.

These tools typically allow you to set up product feed tests quickly and easily but check with your provider first.

2. Testing using feed rules within Google Merchant Center

If you don’t have the budget or the need for a feed optimization tool, you can also set up tests within Google Merchant Center. But, it is a little more complicated.

The best way to set up tests is through Feed Rules. We’ll show you how to set these up.

You first select the attribute you want to make edits to. In this case, we’ll run a test on the title attribute.

Screenshot of changing the title attribute in Google Merchant Center with feed rules.

We then define our test group in the conditions section. This determines where our test changes will be applied. We want a roughly random split of the product feed in two. 

Defining the test group in the conditions section of Google Merchant center.

Here, we’ve specified a bit of regex to tell Google Merchant Center to only apply this test on products whose ID ends with 1,3,5,7 or 9. In other words, those ending in odd numbers. Products with IDs ending in even numbers will therefore be our control group

In the next session, we define what changes to actually make in our test group. 

define the change to make to the test group

Here we are telling Google to prepend each product title with the product’s color and then the brand name. 

Finally, we append each product title with “ for “, followed by the suitable gender for each product. So, if our product was a t-shirt, it might now be called “Blue Ayima T-Shirt for Women”.

We let the test run for a while, then we assess how the odd-numbered products (our ‘Test’ group) compare to even-numbered products (our ‘Control’ group)

2. Give Google as much information as you can

The more information you give Google in your product feed, the better Google can match you to relevant users and queries.

For every optional attribute that you assign- such as the color, suitable ages, and genders for a product - the more likely Google is to place it in front of relevant queries. 

For example, tell Google explicitly that you sell a dress whose color is ‘dark red’ to match with users searching for ‘dark red dresses’. Don’t just assume that Google will guess the color from the product title.

Adding attributes like item group IDs to your product feed will also help Google to understand which of your products are similar. This means that, among other things, they can serve closer variants in your remarketing.

Adding attributes like item group IDs to your product feed will also help Google to understand which of your products are similar.

3. Give the customer as much information as you can

This is because the user can see this information before they click your Shopping Ad, and therefore before you get charged. This means that users who click are more likely to convert.

There is some key information to provide in your product feed for this. 

Highlighting your returns policy is particularly crucial in fashion shopping feeds. 

Delivery rates are important for all industries: You can understand the frustrations of getting close to checkout on a product and suddenly getting hit with a £10 delivery charge. You don’t want to be paying for the click before that realization.

Furthermore, providing product-specific parameters like material, gender, age, and color will only reduce wasted ad spend.

This also extends to providing product descriptions that are as descriptive as possible. SEO-friendly product descriptions will generally be effective for product feeds, so putting in this work kills two birds with one stone. 

For example, the ad in the bottom left ranked for the term ‘dark red velvet dress’. They have clearly specified the colour (‘wine’ being semantically similar to ‘dark red’). And, within the description, they have specified ‘luxe velvet material’.

red velvet dress product feed listing ad and then website description of the same dress matching up

And, always be truthful with your descriptions. Misrepresenting your products in your ad only means that you are paying for clicks that aren’t relevant.

4. Add key product information into your titles

As well as providing this information in attributes, you can go a step further and include the most important values in your product titles.

Once you have them built out, you can use feed rules to dynamically insert values such as product color into the title. 

The product title is the most important attribute in Google Product Feeds for specifying keyword targets.

Of the top 10 products for the search “dark red velvet dress” at the time of writing, 9 had the word ‘velvet’ in the product title. The only one that had not specified this value within both the description and the color parameter. 

The most important variable will differ depending on the product. So, you will need to identify and test what these are.

5. Use n-grams to harvest profitable keywords

The best way of identifying these is through search query reports.

Because they dynamically match to keywords, shopping ads drive a huge number of different search queries. A lot of these are ‘long-tail queries’: Very low volume and highly specific. 

Performance is hidden among these. Particular styles might be hugely popular, but their true volume will be made up of potentially thousands of slightly different search queries.

We released our free ngrams scripts for google ads for this reason. This lets you analyze the performance of phrases within millions of search queries. 

You can use these to learn what attributes to add to your feed. If you see a lot of colors in your top phrases, add your color parameters using feed rules. Test the impact of these changes using tip 1.

You can also use these to negative out poor-performing phrases, such as the impact of bidding on competitor keywords in your shopping ads.

For example, below we see a lamp shade appearing among a sea of sunglasses for the term ‘white shades’. Analyzing the n-grams would clearly highlight the performance of this term.

a lamp shade appearing among a sea of sunglasses for the term ‘white shades’. Analysing the n-grams would clearly highlight the performance of this term.

6. Increase the frequency of product syncs

The more frequent your product feed updates, the more up-to-date the product information your customer sees on Google.

Increasing the frequency of your product feed uploads will reduce the amount of time that ads will be live for out-of-stock products. This will be particularly important at high volume periods such as Black Friday, where being live for an extra day can mean millions in lost revenue.

It also improves the user experience: A user might have clicked on your product ad because it said that the item was on sale, and therefore the lowest price. They will be disappointed to learn that it is actually full price. When they leave the site to find the second cheapest, you have again paid for a click that you don’t want.

Multiply this by a factor of thousands per day, and it gets expensive very quickly.

Making your feed always match what is on your website will also reduce the risk of product feed disapprovals.

There are four options that will impact how quickly your product ads will update.

Upload typeHow often does your product information update on Google Shopping?
Manual uploadWhenever you make a manual change to your product feed
Scheduled FetchMaximum frequency is Daily
Automated uploadUp to 4 times per day
Content APIAlmost immediately- you can specify how often you make calls. You might even increase these at key sales periods. Google has a detailed guide here

The option that is most appropriate to your business will likely depend on your ad spends and developer resources. This tip will be more impactful the more money you spend on Google Shopping.

7. Test different product groupings

Another major optimization you can make to your product feed is to experiment with your product groups.

As far as you can, look to group products based on how similarly they perform. 

For some sites, this might be the product category. 

For others, it could be sales performance. We’ve even tested grouping products based on price band. You can see below the close correlation between price and ROAS. On the right-hand side, you can also see how we used feed rules to implement this.

grouping products based on price band. Here is close correlation between price and ROAS. On the right-hand side, you can also see how we used feed rules to implement this.

By grouping by performance, we were able to get Google to push the highest price bands much more. 

You might also choose to group products based on their profit margin. This means that you can set different POAS targets and bid to profit. We’ve written extensively about the benefits of POAS marketing.

8. Sync your product feed with your tags to remarket dynamically

To improve the performance of your shopping ads on Google, you need to set up dynamic remarketing. 

This seems an overly simple tip, but it is shocking how often this is done incorrectly.

To do this, you have to include the product ID within your Google Ads remarketing tag from each product page. 

There’s an example of this below, where our ecomm_prodid attribute lets Google identify which exact product in our feed the user is visiting.

dynamic remarketing tags on a product page


Without this, Google cannot see what products people have browsed. And so, Google cannot serve them an ad for this product when they leave your site. Or to show them similar products.

This is therefore particularly important if you are running ads on display.

People seem to think that Google can do this automatically. It can not.

9. Make your product images fit in, then make them stand out

This tip comes in two parts.

Take a look at the shopping page on Google for the terms that you are going after, and firstly note what kind of image specs Google Ads is surfacing.

For example, for the ‘dresses’ search below, Google is quite clearly surfacing a more portrait view. On the left-hand side, where their product image doesn’t match this spec, the ad takes up much less real estate.

Image that shows two dresses. Image specs for Google ads on shopping page with portrait view is preferable over square image.

Secondly, look at what sort of context the products are showing within. Is it the product alone? Is it a lifestyle shot? Are there any non-white backgrounds?

In the image above, for example, the right-hand product proves that we can have eye-catching backgrounds. 

We can use this to our advantage.

Consider the images below:

All of the products in this image have white backgrounds

All of these have white backgrounds, and at first glance, there is nothing particularly eye-catching.

If we drop a bold, and different style background into our product ads, it’s really going to stand out in this range. That will get people to click, and that in turn will improve the perceived quality of our product.

As with all of these tips, test different variations within your images, Google will reward some images and punish others.

The only way to know what will improve your performance is to test. We hope we have given you some ideas that you can take back and improve the performance of your Google Shopping Ads.

We’re passionate about Google Ads for ecommerce. If you’re interested in learning more, our experts are on hand. We can provide a more in-depth personal analysis or a bespoke strategy. Get in touch with the team here, we love to hear from you.

Learn about our free, personalised insights club

More Paid Media

July 2022 Digital Roundup

03 August 2022

PPC Budget Calculator - Our Free Tool to Estimate Cost

12 July 2022

Tips to optimise your Google Product Feeds

06 June 2022