Jani KahukiwaDiscover how to use AI to spike up your Google Shopping Ads until they outrank your biggest competitors.
Glen Wilson from Solutions 8 reveals how the new AI tool FeedGen can significantly improve competitive visibility and ad performance for client campaigns. With the help of Regina Bellows, we explore the benefits of implementing FeedGen into your Google Ads strategy for immediate results. Ready to find out more?
Table of Contents
What Is FeedGen?
FeedGen utilizes generative AI to optimize titles, descriptions, and attributes of products in shopping feeds, ultimately improving visibility in the auction and increasing exposure.
It’s a free tool we’re currently integrating here at Solutions 8, and it’s the same tool Google uses for its Shopping ads! It is accessed by GitHub.
While it’s not the most intuitive tool, performance is impressive and results are undeniable. Once it is configured, the benefits are numerous and we will outline its many advantages here in this blog.
FeedGen optimizes Google shopping titles and Google descriptions along with additional information related to each product. This in turn gives you more leverage in the real-time auction that happens every time a user enters a search term and helps you get more exposure.
Go Data Feed and DataFeedWatch are similar tools for optimization, but not all customers have the budget to use these. There is still a need to formulate rules for these tools.
What’s the cost of rewriting your product titles and descriptions?
The cost of rewriting 1,000 products is approximately $2.50. You pay through a Google Cloud project. A $300 credit is automatically offered, making it a great incentive to spend on your FeedGen runs.
We cannot think of a more cost-effective way to optimize titles. If a business has thousands of products, FeedGen is just the tool to manage a huge range of SKUs. Without implementing FeedGen, Google Ads tends to do this.
FeedGen uses Gemini Pro Vision, a foundation model that can understand, classify, summarize, and create content from images and videos. Its ability to recognize visual images at lightning speed gives an edge to any business aiming to put as much information in the system as possible.
How It Works
FeedGen uses highly developed LLMs (large language models) to improve product titles. It can generate comprehensive descriptions and fill in missing attributes in product feeds.
This increases the quality of the feeds and helps detect and resolve quality issues. Google Ads likes products that have a lot of information. It favors those products and as a result, gives your client a competitive edge.
Remember, this is a Google Ads tool, and using it will help you in the Google Ads auctions!
With automatic links to your website, existing titles, descriptions, attributes, product IDs, and images, FeedGen will tell you what is needed to maximize performance.
Google’s best practices say that we should put the brand name first followed by gender, product type, color, size, and any other attributes of your product. Google views this as a good flow of a title if your brand is well-known.
A lesser well-known brand would place the brand name at the end.
Many ad agencies optimize campaigns by giving the algorithm what it wants. FeedGen is unique because it hands you the right words in perfect order to satisfy the feed monster.
FeedGen is a script-based application that runs as an HTML sidebar in Google Sheets. You initialize the template and run it through FeedGen.
When you’re happy with the created Output Feed, you make a supplemental feed with optimized titles and descriptions. This becomes the default, which attaches itself to the main feed.
Let’s initialize your template!
FeedGen Template
The Google Sheets spreadsheet template is where all the magic happens.
The prompts outline the best practices for the different types of products you are selling. It suggests how to set up the title structure.
Character limit plays an important role in the Shopping ads we see. Most people don’t use the 150-character limit in the title tag. They usually use 70. And most people don’t use the 5,000-character limit available for descriptions.
How many keywords can you fit into 5,000 characters? What are the advantages?
All data from the Input Feed will be provided as part of the prompt. This is all the vital information related to product inventory that you hand over to Google.
The configuration platform holds the Input Feed that requires optimization. Specific configuration values then control how content is generated.
Setting up FeedGen is complex; as mentioned earlier, it’s not intuitive. But, once you have it up and running, it’s a real time-saver.
However, using FeedGen can be challenging as it requires the creation of a Google Cloud Project, Vertex API activation, and a solid understanding of Vertex AI for machine learning operations. There are clear instructions to follow but you will find within some steps that there are another 30 steps to take. This can be challenging and confusing. This is where Solutions 8 comes in handy!
The spreadsheet is used for both (optional) human validation and setting up a supplemental feed in Google Merchant Center (GMC). We highly recommend manually reviewing and verifying the generated titles and descriptions.
Success With FeedGen
Success can come in an instant. Just look at our GMC- data below. We turned on the supplemental feed on the 14th of March. This graph demonstrates the competitive visibility within the auction.
It shows the competitors’ relative overlap, high-position ads, and organic ratio.
Compare our brand to its biggest competitors in the graph above.
From March 14 onward, there was a steady performance of our client’s brand, evidently outranking its competition thanks to its new and improved titles and descriptions.
It was always difficult to scale and maintain MER for this specific client. We want the budget to spend more but also ensure a good return on investment. We could easily increase the spend but revenue would inevitably drop.
When the product feed is fully optimized the way Google likes, it gives a better score, making winning the bid in the ad auction easier. Your business stands out against your competition simply because Google likes your ads and your title tags and descriptions.
One more convincing argument for the success and efficiency of FeedGen became clear when we did a quick check on the converting keywords before we turned it on.
There were a lot of non-branded search terms in the insights, specifically for a PMax campaign. We excluded the brand and bid on new customers only.
When we turned on FeedGen, the results showed a jump of 50% in non-branded keywords simply because Google had more information from the titles and descriptions.
Checking Spend, Revenue, And MER
Improving ad spend proficiency was the most obvious benefit of implementing FeedGen. An increase in spend by 100% while maintaining revenue at 83% is unbelievable. MER (media efficiency ratio) dropped by 10% for this client.
Within a month, we have doubled our spend while doubling our revenue.
Please note no increase was made to the budget on March 14.
The daily budget was never spent entirely until all the campaigns recognized they were more prominent in the auctions. That’s when the budget began to spend in full.
Utilizing FeedGen For Solutions 8
Solutions 8 uses FeedGen to stay ahead of the competition. It’s the most obvious solution to optimization, saving time and money for our clients.
Our clients are all being offered this tool to optimize their campaigns. In-house specialists at Solutions 8 work day and night to ensure all client feeds are optimized to best practice. That is the key to success for now.
We intend to roll it out as standard practice within Solutions 8 shortly. If performance is anything like the above example, why wouldn’t we, as a digital marketing agency, take advantage of this tool?
Thanks to the dedication of our team, hopping onto this so early in the game will give our agency anywhere up to a year ahead of the competition.
For more information, why don’t you check out our complete digital marketing services page?
