Our Feed Technology

Taking your products to more users

Optimising your product feeds is the #1 suggestion by Google to increase the amount of impressions your product ads deliver and increase quality scores to lower CPCs.

Did you know that every product in your Merchant Centre has a Quality Score assigned to it? It’s a number between 1 and 10 and it has a serious impact on the amount you pay to buy traffic. For example if you have a quality score of 10/10 then you’re able to buy traffic at a 30% discount compared to the rest of the market, if your quality score is just 5/10 then you have to pay 40% more than the rest of the market to buy traffic!

RedBrain serves billions of Shopping Ads weekly and has over 800 million products in our Merchant Centres, so we’re in a unique position to be able to find the most and least significant factors to build Quality Scores and help you increase them.

Take this same product that we had in our Merchant Centres, one had a Quality Score of 10/10 and one had a Quality Score of 4/10.

Other CSS

Title: iPad
Price: 499.99
URL: https://somestore.com/ipad
Color: Grey

RedBrain Managed

Title: Apple iPad Pro 2018, 32GB, Space Grey
GTIN: 00888462760799
Brand: Apple
Price: 449.99 (updated to match site)
URL: https://somestore.com/ipad
Color: Space Grey
Product Type: Smartphones > Apple, iPhone Operating System (iOS)

The difference? Well we’d run our unique feed optimisation service on it!

Google builds Quality Scores on many factors which aren’t publicly known, but over time we’ve worked out how to build them – the average quality score for our own campaigns is 9.2!

(Red bar across the page) One quick note – this only applies to our Managed service or if you’ve opted into our Feed technology, if you’re using our Self Managed service then we don’t ever touch the product data you give us, you’re in full control.

As we see so much product data from so many retailers, we’ve found everything you could possible need to know about most products out there. We know the screen resolution of your iPads or which cars the headlights you sell fit. In fact, we see over 1 billion products daily, so we’re always up to date.

After processing all this product data we then match the product data we get from you and then look at all the descriptions and attributes that you’ve missed. Essentially it’s a big game of fill in the blanks.

By adding in all this extra information, Google starts to favour our products and Quality Scores on the products increase; driving more ad impressions. Several retailers we work with have seen our Managed campaigns outperform their own direct Google Shopping activity even though they were bidding more than us!

Let’s look at a couple of specific optimisations that can lead to higher Quality Scores.

Product Type

Merchant Centre products have a field called product_type which Google says let’s Merchants structure products according to their own internal category system. Through our testing we found that well structured product type fields could have a significant impact on quality.

Take for example an Adidas trainer, a simple implementation of the Product Type field might look like:

Trainers
>
Adidas

It gives Google a bit of context to the type of product and which category it might be found in, but we can definitely do better. A top performing Product Type for that product looks more like:

Trainers
>
Adidas
>
Mens
>
Originals
>
Dragon

We’ve added in the gender, so that Google can also show the product on search terms like “mens trainers” that it might not have ranked for before, and we’ve also got the product range (Originals) and style (Dragon) so that branded keyword searches should drive more impressions too.

Product Title structure

Probably the most important ranking factor for products is their Title in Merchant Centre. That’s really where Google first looks to try and match relevant products to users search terms.

Shopping Ads only have a small visible title so picking the right combination of terms to match user queries and still maintain a great Click Through Rate (CTR) is really difficult. Internally, we map products to about 32,000 categories, and different product categories have different best practices for what a title should look like.

Let’s look at two top level categories, Electronics and Fashion.

For Electronics products, the best performing titles tend to follow the format:

Brand
+
Attribute
+
Product Type
+
Model #

So for a TV, it might look like: Samsung 88” Smart LED TV with 4K 3D Curved Screen (UN88JS9500)

Whereas for Fashion products, the recommendation is to follow a different format more like:

Brand
+
Gender
+
Product Type
+
Attributes
color, size, material

So for a sweater, they will appear as: Ann Taylor Women’s Sweater, Black (Size 6)

These formulas change the further into the categories you go, for example a mobile phone will have a different title formula than the television example above. The most important thing is that we’re always benchmarking our performance against alternatives to make sure we’re generating as many clicks and impressions for our retailers as possible!

Want to learn more about how our feed optimisation technology can work for you? Get in touch!

Want to learn more about how our feed optimisation technology can work for you?
Get in touch!