Magento Google Shopping Best Practices - The Definitive Guide

About Vervaunt

Vervaunt is a London-based Paid Media and eCommerce Consultancy who work exclusively with retail brands to support and drive growth. The company specialises in managing paid media accounts (with a focus on Google Shopping) and supporting replatforming projects (from scoping & requirements gathering through to project support). Vervaunt’s paid media offering spans across all paid channels (search, social, shopping, display, Amazon etc) and is geared around being more hands-on and much more technical than your average team. Some of Vervaunt’s clients include MUJI, Sunspel, The Conran Shop, Antler, PMT Music, Lulu Guinness, Pet Supermarket, Self Portrait and David Austin Roses.

About the Author (Josh Duggan)

Josh Duggan is the Co-Founder of Vervaunt and runs the paid media offering and team. With more than 8 years experience, Josh has predominantly worked in the eCommerce space across all main RTB channels incl. search, shopping, social, video and programmatic display. At Vervaunt Josh leads the management / optimisation capability and the development of all Vervaunt proprietary technology. Josh has spoken at multiple global conferences.

An introduction to Google Shopping

Google Shopping is a form of paid advertising specifically for online retailers, that allows users to search for products across the web and compare prices between different vendors.

Google Shopping displays rich product listings within standard results pages, allowing consumers to compare pricing and click directly through to the product detail page on the given website, as can be seen below. These listings  also serve within the shopping results page, which allows users the option to browse products and filter results. 

Over the last few years the visibility of Google Shopping listings has increased considerably, with shopping ads now driving 80% of Google Ads traffic across non-branded, retail queries. In terms of the benefits for merchants, Google Shopping typically delivers lower CPCs and a higher conversion rate than standard text ads, and is generally one of the quickest and most efficient ways to promote products.

Google Shopping delivers ads by indexing product data within the shopping feed to match a user’s search query. The initial overhead around shopping is often lower (dependent on product data, which will be covered  later in this guide) than for search ads, as they do not require the granular build of thousands of keywords. For retailers new to paid media, launching Google Shopping is one of the most efficient ways to build out your activity; you can run a search term report to identify high-performing queries which convert to sales, gradually building out product level text ad campaigns based on these searches. As your customer base and audience knowledge grows, you can begin targeting broader generic keywords, overlaying audience data to ensure these are only served when conversions are most likely. 

The number of shopping products served in the SERP has steadily increased from 5 to 30 with a range of new ad formats being rolled out. In addition to this, an increase in mobile traffic (where shopping ads are more prominent) has also contributed to the massive increase in Google Shopping traffic.

An introduction to the Merchant Centre and building a shopping feed

In order to set up Google Shopping, retailers need to initially set up a Google Merchant Centre account, which houses the product feed and product data. 

Here merchants can upload product data in various forms or, depending on the ecommerce platform, connect a catalog via API. If you’re using a data feed (most common), this must be saved in one of the below formats:

.xml: XML (not supported for local product inventory feeds)

.txt: TSV

.gz: Gnu zip, compressed TSV (recommended) or XML.zip: Zip, compressed TSV or XML.bz2: Bzip2, compressed TSV or XML

Product data can also be sent via an API as mentioned above, however it is recommended to use a fetch of the product data in the above formats so that it can be easily downloaded to debug and check for issues. This also provides an opportunity to optimise the data - you will likely want different data serving to what is simply promoted on site. Depending on requirements, the Google shopping feed can be used for other purposes, such as Facebook, Amazon and Bing advertising. 

Building a shopping feed via Magento extensions or third party solutions

Magento's API solution for Google Shopping

It’s worth mentioning Google Shopping ads Channel which was disabled on April 28, 2020. Although this was a good option for smaller retailers, due to its quick and easy set-up, it did have a few drawbacks. It required advertisers to create a new merchant centre which would unlink any existing merchant centre accounts. New merchant centres can take up to 7 days to be approved and for many businesses, 7 days without Google Shopping would have a negative impact. It was also very limited in terms of optimisations and managing product data, as it would simply input data from Magento directly to the merchant centre. If the Magento API does return, it is worth reviewing these points before deciding on this option.

Magento shopping feed extensions

In general, it’s much better to use a third party extension to create and generate a shopping feed, in comparison to relying on a direct API. We’ll use Wyomind’s Data Feed Manager extension as an example in this section, however other suitable extensions include those developed by RocketWeb, Amasty, Data Feed Watch and Mirasvit. These solutions all have relatively similar offerings and allow for multiple data feeds to be created based on attribute values and hard-coded values.

They are cost effective (normally with a one-off fee), allow the management and optimisation of the feed within Magento, and are easy to install and set up. Please refer to the next section ‘Installing a module and configuring the feed’ for examples on how to set up and optimise a product feed via an extension. 

Third party shopping feed solutions

Operating as a separate platform which will import  data from Magento, third party feed solutions are the most sophisticated of all the options. They give the most control over the feed optimisation and the ability to create additional custom fields. For example, you could import data from multiple live sources to refine and adjust your feed which is harder when managing the feed purely in Magento. As a further example, dynamic cost price can be pulled from an external source, allowing the third party solution to update the current profit margin into the feed, which then helps to optimise activity accordingly. 

There are a number of third party feed solutions, Feedonomics, Intelligent Reach, GoDataFeed, Channel Pilot, Shoptimised and Feed Optimise being a few we have worked with. Within these solutions, we typically recommend the smaller more specific feed management solutions, for example Shoptimised and Feed Optimise. These have low ongoing costs and almost everything  mentioned in this guide is possible with these options. The enterprise level solutions (Feedonomics and Intelligent Reach) are slightly more sophisticated but are only really needed by larger merchants. 

For simple feed management it should be a relatively low ongoing cost. If you are using the solution for broader purposes, for example managing Amazon inventory and passing back orders into Magento, it’s worth speaking to a number of solutions to understand their costs and offering. 

Installing a module and configuring the feed

There are two ways to install an extension in Magento 2, similar to Magento 1: either by using the Extension Manager (which is the Magento 2 equivalent of the Magento Connect feature) or via a manual install. The former is very straightforward; all that’s needed is an admin account and the Magento installation access keys (to connect to Magento Marketplace where most extensions reside). Once authenticated, available extensions can be browsed and simply installed.

If you choose to install the selected extension manually you will have a further two options: download the files and place them in the app folder (following a similar approach used by Magento 1), or use Composer (a package manager for PHP) which is the preferred method.

Not every third party extension allows installation via Composer although more and more vendors take advantage of this method as they get more accustomed to it. Installing extensions via Composer requires technical skills and you should always ask a developer to perform the installation.

The main benefit of using Composer is that with each further deployment to the website, it will automatically receive the latest version of each third party extension, therefore keeping up-to-date for security reasons, as well as having the latest features without having to manually check for each new version (especially beneficial to sites that take advantage of many extensions).

There are further technicalities to consider when choosing the Composer approach to install extensions, which will depend on your development and deployment strategies - you can always consult the Magento devdocs to learn more.

Once the module is installed, you will need to configure it and build out the feed data. This will likely require new attributes to be created if you’ve not done any form of paid media in the past. Once you have the feed created, you can then upload this to the Merchant Centre and wait for initial feedback. There may be a number of issues with the feed and these can be resolved by adjusting the configuration of the feed in Magento. 

In Magento you will essentially map your Magento product attributes to the most relevant fields within the shopping feed. Please see an example below where we map the Magento “google_feed_title” attribute to the Title attribute in the feed. You may also want to map other Magento attributes such as “quantity” (which would represent stock availability) to a custom label allowing you to upweight products with more stock, for example. 

The key benefit of creating a second product attribute for things like optimised titles and long descriptions, is that you don’t need to edit the live title or description on the website, in order to have the data optimised within your shopping feed. 

The above screenshot shows how product attributes are used as variables within the XML markup. These can then be changed into new attributes and created and populated with optimised data.

Optimising the product feed

Summary of product attributes

A list of all the required and additional product attributes can be found here: https://support.google.com/merchants/answer/7052112?hl=en-GB. 

Include as many product attributes as possible to allow for filtering (within the shopping results page) and to rank for more specific queries. Please see a top-line summary of the key attributes below (Shopping feed attributes and optimisation), as well as recommendations for optimising these to increase coverage.

Shopping feed attributes and the GTIN (Global Trade Item Number)

GTINs are a common issue for many advertisers when creating a Google Shopping feed. Either this data has not been added to the product data in Magento, or it can be time-consuming to source. It is best practice to include the GTIN, however there are two workarounds when this is not possible. Here is a simple workflow to follow:

• Add GTIN when possible (this would be added as a product attribute in Magento and the values would most likely be imported if they are not already available in Magento).

• If no GTIN is available, ensure your Brand and MPN (Manufacturer Part Number) values are pulled into the feed. These should still be included even if the GTIN is available.

• If no GTIN or Brand and MPN – Add [identifier_exists] and a value of ‘no’ into the feed. 

If using bundled products or grouped products, use the GTIN of the most significant product. If a bundle consisted of 'buy a pair of shoes and get two free pairs of socks’, it would be the GTIN of the pair of shoes that would be most significant, for example. 

Optimising the Product Title

Product titles have the largest impact on product relevance and how often your ads appear against searches. One initial thing to remember is that the title for Google Shopping doesn’t necessarily need to be the main title of the product, it could be a separate attribute in Magento. You could, for example, add in a new product attribute called ‘google_title’ and then configure the feed to pull in this value, rather than the existing title.You will most likely want to drive as much volume through Google Shopping as possible. If so, you should use the keyword planner to identify the search terms with the highest volume which match your product. Titles can then be amended to increase how often they appear for these terms. For example, the following title: High Waisted and Skinny Fit Ripped Jeans - the keyword planner can be used to identify which terms would be best suited at the start of your product title. Within this example, Ripped Jeans has the highest search volume and should be at the beginning. The product title Women’s Ripped Jeans - Skinny Fit High Waisted Jeans can then be created. 

Google has outlined industry recommendations and you will want to include gender for clothing products, for example. The key thing here is to test different product titles and see how your impression share and volume is impacted. It might also be that users search more often for a synonym version of your product. For example, Sandals is searched 83% more often than Flip Flops; therefore, product titles should be amended based on search behaviour. 

For clothing & accessories, Google recommends the below format for titles:

• Brand, Gender, Product Type, Attribute

As per the above, include brand where there is high search volume. For example, a Levis title should be: 

• Levis Women’s Ripped Jeans - (Colour) Skinny Fit High Waisted Jeans

There are many feed tools and optimisers which will essentially add attributes such as gender and product type to the beginning of your title. This guide is designed to provide you with the insight to optimise your product feed internally. 

“We recommend optimising your 10% top sellers manually using the recommendations in this guide. You can then automatically optimise all other products by appending the most relevant attributes. It might be that you want to append Gender and Product type to your title so you can quickly optimise hundreds of products at once. 

For example, using ‘Women’s Jeans - High Waisted and Skinny Fit Ripped Jeans’ would help you rank more often for the core term ‘Women’s Jeans’, now at the beginning of the title. With bulk changes, be sure you sense check the logic, for example do not end up using gender twice within the title” - Josh Duggan, Vervaunt

Optimising the Product Description

As much detail should be added to product descriptions and other attributes as possible. Google suggests including more detailed descriptions and extra attributes, such as material and colour, can lead to 10%-20% more impressions. In general, product descriptions should be at least 500 characters and ideally  1,000 - 3,000 characters. You should also use the key terms from your product titles  multiple numbers of times within the product description. 

“To easily expand your product descriptions, we recommend writing at least 1,000 character descriptions per product category. You can then simply append your existing product descriptions with the relevant category descriptions you have built out. 

You should cover synonyms which users may be searching for, and also all high volume queries you pull out from the keyword planner, for example ‘Smart Casual Shoes’, ‘Work Shoes’ etc. Once appended to your products, you will have a fully enriched description for each item. You could again build out specific descriptions for your top 10% of products. Make sure you include the search terms which are driving sales for your products to ensure  maximise visibility on terms you know convert” - Josh Duggan, Vervaunt

Matching the Google_product_category

This is another key attribute and it’s imperative to ensure products are categorised into the most specific and relevant groups possible. You can find the list of all categories here: https://www.google.co.uk/basepages/producttype/taxonomy.en-GB.txt. 

If you sell guitars for example, the category should be: Arts & Entertainment > Hobbies & Creative Arts > Musical Instruments > String Instruments > Guitars. 

Setting the Product_type

Although an optional attribute within the feed, the product type allows more descriptive terms to be added to your product category in comparison to the Google product categories, and will also help search rankings. Product type is often overlooked, however, Google recommends: “When you submit a product_type, we can better understand what you’re selling, and when we understand what you’re selling, we can help connect users with your products.” It is also a key factor in your Google Shopping set-up within Google Ads.  Using product types will allow the creation of a more granular campaign structure and also ensures that new products will filter into the correct campaigns. Below shows an example of a product type and how this can be used to build a campaign.

For Books > Non-Fiction > Sports > Baseball, you could use this structure:

• Campaign – Books

• Ad Group – Non-Fiction

• Product Group level 1 – Sports

• • > Product Group level 2 – Baseball

  • > Product Group level 3 – Item IDs

One tactic we have recently begun testing, is to add in competitor terms to the final levels of the product category. This will help you appear against competitor terms (if relevant) without impacting anything the consumer sees, for example title and description. 

Optimising Images

Product images need to be high quality. At a minimum, use high-res images that fill an 800 x 800 pixel space. You can test varying camera angles of images to make products stand out and look different to competitors. Don’t use watermarked or edited product images though as these can be disapproved. 

The maximum image and file size is 64 megapixels and 16MB. Make sure the image is clear and it is typically recommended that it take up no less than 75% of the background. Below is  an example of a common product image style. 

Sale prices & promotions 

Sale prices

When changing prices during a sale, ensure the feed is updated correctly to display the promotion. Please see Google’s example below: 

By simply changing the price in Magento, the Google Shopping price will automatically update as well; however it will not show the price drop and highlight the sale. You should use the sale price attribute in Magento so that the cost saving is pulled into the ad (as shown above). To show the price drop, simply update the sale price attribute with the new price. You may need to update your feed app for the sale price to update. Example below:

Alternatively you can manually add this attribute and schedule the sale price using the  ‘sale_price_effective_date’ attribute. 

Google does sometimes recognise if your feed price drops over a 90 day period and will highlight this. However, it is automatic and not guaranteed and the best option is to use the sale price attribute within the feed. 

Promotions

Promotions are added within the Merchant Centre; you can either discount by a set amount or percentage, or offer a free gift or free delivery. The above promotions are either applied to all products or a selection. Example below:

To apply this to all products, simply select ‘Choose all products’. To apply the promotion to specific products, the Promotion ID (seen above) needs to reflect the promotion_id within your product feed. The Promotion ID needs to be applied to only the relevant products via the shopping feed. This option allows multiple promotions to run across separate product groups (promotions can also be scheduled in the Merchant Centre). You can now also filter products when adding, the promotion so you may add a 10% promotion to products where brand = Nike, for example. 

Feed diagnostics & common issues

There are a number of alerts within the Merchant Centre to watch out for. These are categorised as Errors (red - which will likely lead to disapproved products), Warnings (yellow - which will negatively impact your ads), and Notifications (blue - which are suggested optimisations).You will need to fix all Warnings immediately. These will generally be issues with the required attributes within the feed. Most of the errors should be relatively self-explanatory and easy to amend.

Missing microdata is a common Error notification. You can easily add the required markup (this is the same as structured data commonly associated with SEO) to the templates based on your attribute values, or by using an extension like MageWorx SEO. The missing microdata will most often be “condition”, “price” or “availability” - MageWorx SEO allows the chosen markup to be added directly onto the product detail pages.

One extremely important aspect to note, is the possible issues that may occur during replatforming. The Google Merchant Centre houses your historical product performance data within the SKU ID. This means that should you ever change product IDs (if moving platform or for any other reason) that you need to ensure the IDs remain the same in your feed. You can do this by simply updating your feed export with the previous product IDs, rather than their new, different IDs. 

How to build your shopping campaigns in Google Ads

Account structure will be fully dependent on the advertiser, however there are overall principles that can be followed when creating campaigns. Campaigns should be segmented by Product Type, for example using the navigation bar of a website to separate out Cycles, Trainers, Clothing, Football, etc. You may wish to split this down further if specific products are driving a significant number of conversions and therefore warrant their own campaign and budget, for example, Nike trainers. 

As above, campaign structure may follow the navigation bar on a website. A sports retailer for example may have a campaign for Shoes, Tees and Shorts. Ad groups will need creating based on the second Product Type layer, for example Shoes split by Running Trainers and Tennis Shoes. They could alternatively be split by brand: Shoes (campaign) > Adidas (ad group), for example. It may be more relevant however to build campaigns by brand, segmenting ad groups by Product Type, for example. Adidas (campaign) > Jumpers (ad group). Many retailers segment campaigns by brand, working towards different KPIs based on the respective profit margins and stock, etc.

At the campaign level, it’s possible to optimise by specific locations, devices, time of day and also easily manage negative keywords (through the shared library) - these should be taken into account when building campaigns. Product categories for example, will see different performances across different devices. A basic example is that high value products often convert less on mobile devices. In this scenario, you would want to break out campaigns factoring in the product price. Device bids can then be measured and optimised separately. Within  ad groups, product groups are the final layer which contain either a group of products or individual products. Bids are applied at the product group level, and you would therefore generally want to break out specific products into their own product groups. When products are driving a lot of volume and/or you want to push or de-prioritise products, then you will need to ensure the products are in their own product group with their own bid. In general, we recommend breaking out product groups as granularly as possible and only group products together if they are very similar and low value, for example packs of golf tees. 

It’s possible to have multiple product groups within each ad group and these should be further segments of your ad group category. If your ad group is Shoes for example, you can break that out into product types >  Sprinting Shoes, Running Shoes, Track Shoes and Road Shoes, etc. Five levels of product groups are possible. 

In this example you would break out another product group within Sprinting Shoes - the Item ID to allow SKU specific bids. Please see an example below which has broken out Yamaha (Brand), Bass (Product Type 1), Bass Guitars (Product Type 2), Electric Bass Guitars (Product Type 3) and then the Item IDs (2006 and 642). 

If you do break out Item IDs (individual products) within product groups, all other SKUs will pull through within an ‘Everything Else’ product group. It is important to monitor this closely to ensure that products are split out and optimised where needed. If an individual product within an ‘Everything Else’ group is proving expensive and not converting, this will need to be broken out into its own Item ID product group within Google Ads. You can then apply a reduced bid, or even exclude it. Conversely, top sellers should also be separated and upweighted. You should also be aware that new products will be assigned into the ‘Everything Else’ category. If you were to add new Sprinting Shoes to the feed for example, you would need to break these out immediately to optimise the bid.

How to optimise your campaigns in Google Ads

The search term report is one of the most important aspects of Google Shopping. Unlike search campaigns where you decide the keywords, Google Shopping serves ads when a user’s search matches your product feed data. This however can lead to ads showing for a number of irrelevant and under-performing keywords. It is therefore imperative to review your search term report, filtering by impressions and spend, to make sure the worst performers are excluded. 

Bids are placed at the product level rather than at keyword level, which can cause issues. For example, it is likely that someone searching for Buy Nike Air Max is more likely to purchase than a user searching for Trainers. However, with bids placed at the product level, you assign the same bid for both queries.

“One common tactic to optimise keywords within Shopping is to segment campaigns by branded and generic terms. This will allow you to maximise coverage and impression share when users are searching for your brand and then optimise more specific bids to the generic terms. 

Simply create 2 variations of your campaign: Brand and Non Brand. The branded campaign should be Low priority, and the non branded, Medium. Run an SQR (Search Query Report) over the last 6 months and add all searches which don’t include your brand as a Negative Exact Match keyword within the Brand campaign. Then add your brand name, including misspells, as Phrase Match Negatives to your Non Brand campaign. This should filter brand and generic queries into their respective campaigns, allowing you to apply more relevant bids. 

A custom script can also be used to add in the generic terms as negatives to your branded campaign. This will ensure you maximise coverage on brand on not waste your higher bids on the generic terms” - Josh Duggan, Vervaunt

This also works well in the case of product specific names and model numbers. For a sunglasses retailer, a user searching for Ray Ban Rb3025 will likely be more valuable than a user searching for just sunglasses. You can create a low priority model name/ number campaign which negative exact matches all generic searches that do not include the model number terms, e.g. Sunglasses. The generic campaign would then be medium priority and would exclude all model number terms (such as rb3025) as phrase match negatives. 

“For more advanced advertisers, there are multiple scripts available which will move your campaigns towards being keyword specific. You essentially add in all terms you wish to appear for, e.g. ‘buy bed for bad back’ and all other terms which trigger your ads are then automatically added as negative keywords (via the script). After a month or so of the script running, you have a campaign where the majority of traffic goes through your selected terms delivering a higher return, e.g. model number queries (RB3025 for Ray ban Aviators). You can then increase your bid here, and keep your previous campaign live, with a lower bid, adding in those terms you wish to appear for in the script campaign as negatives. This would be the generic campaign, not including top-performing terms, where you would again apply lower bids”  - Josh Duggan, Vervaunt

Another important aspect of Google Shopping is cross match negatives. Google serves products with the highest ad rank (relevance and bid), which may not be the best product to be serving to the user. Therefore it’s important to make sure campaigns are built in a fashion that allows you to apply cross match negatives to ensure the most relevant product is shown. A simple example of this would be a music retailer, who should have one ad group for Drum Products, and another for Electric Drums. Electric could then be added as a phrase negative to drum campaigns/ad groups, ensuring searches including Electric trigger the electronic drum set products. 

“It is also worth creating a high priority campaign for top-selling products. This will ensure these products are shown most often for relevant queries and also allows you to monitor performance and spend separately”  - Josh Duggan, Vervaunt

At Vervaunt we use a feed solution which automatically populates top sellers (with higher than average ROIs) with a custom label, and use a high priority campaign which targets all SKUs with that custom label. Top sellers are then pushed into this campaign with a higher bid and priority setting to increase coverage. If the ROI drops the label is removed and the SKU begins serving in its original campaign. 

Similarly, we use our Google Analytics API to automatically monitor SKU coverage on top sellers in GA. If top sellers on the site are driving only minimal traffic, an alert will be provided to review this. We can then ensure best sellers on site have good coverage across Google, where relevant. 

Additionally, a number of scripts can be used to automatically monitor and apply negative keywords; we have seen many examples where new SKUs can appear for broad, irrelevant terms. Although this should be monitored regularly - the scripts flag under-performing search terms so they can be quickly identified and reviewed. Likewise, a SKU level script flags when a product spends over a certain amount and does not convert. 

As mentioned above, it’s important to ensure products are broken out from Everything Else groups, and this should be an ongoing process. Within the below example, products 522 and 658 have been broken out allowing the optimisation of the bids. By first breaking down by category, and then by brand, table football table, and then toulet, you can see there are Everything Else groups which will house the products not broken out. You want to first segment by product category,brand, etc, so that your bids can be optimised for similar products. In this example, set a bit for everything else in toulet, which should then be different to the Everything Else bid for rs barcelona. 

It’s important to avoid pulling too much volume through Everything Else groups where one bid is being applied to multiple products where performance can vary. It’s best to make sure that the top spending and converting products are being regularly separated, allowing you to apply more specific bids based on their performance. 

Campaign level optimisations should also be regularly applied which include device bids, geographic bids, ad schedules, and remarketing bid adjustments, for example upweighting users who have recently abandoned cart. At the time of writing,demographic and in-market audience bids aren’t currently available in Google Shopping, however, audience signals are becoming increasingly important in search and there will likely be developments with what’s possible in the future.

A good use of technology and data within Google Shopping is where we have applied logic, based on the stock available. There are hundreds of possibilities allowing more sophistication with APIs and data sources in Google Shopping. For example, optimising  profit margin, weather, competition, site data, etc, but one example that works particularly  effectively is stock. At Vervaunt we offer sizing analysis for clients looking at purchase rate by size, but also at returns rate (to help support sizing information on PDPs). For one client we set up a script which alerts when the higher performing small sizes go out of stock, and spend increases on the larger, under-performing SKUs. We also implemented this strategy on shoes, and have seen considerable improvements to ROI by reducing spend when popular sizes are OOS. As product size cannot always fit into a product title, it’s important not to not waste ad spend where users could land on the site via an unpopular size, only to realise the most common size is out of stock before bouncing. 

Pros and Cons of Smart Shopping

We will first cover the difference between Google Shopping campaigns and Smart Shopping campaigns. In May 2018, Google released its new solution for Google Shopping, this is what we have come to know as Smart Shopping. What is Smart shopping? Google takes your product feed, campaign objective, budget and country settings and uses its machines learning capabilities to do the rest. Google's algorithm decides where and when to serve your ads, how much to bid, and which audiences to target. It essentially takes all management and optimisation away from the retailer, which is great for smaller brands, but limits all of the pre-mentioned tactics above. 

Pros for smart shopping

• Simplified campaign management – by leaving the optimisation to Google, this reduces the resource required to fully manage the campaign yourself, and Google’s machine learning capabilities allow for large amounts of data and optimisations to be processed quickly.

• Smart Shopping combines search with display which allows advertisers to expand their reach, which includes both dynamic remarketing and dynamic prospecting.

• Smart Shopping uses machine learning to serve the most relevant combinations of your visual and textual assets to deliver the highest return.

Cons for smart shopping

• Reporting is limited - placement, search query and audience data is not available. This lack of visibility means you have no knowledge of what may and may not be working. These could also be learnings you could have applied to other campaigns.

• Lack of search query reporting is significant as you won’t be able to see which products are serving to which queries. Some brands would naturally want to avoid appearing against certain terms and Smart Shopping does not allow you to add negatives, i.e. irrelevant or low intent search terms. You cannot identify top performing terms to add into search campaigns either. 

• No control over your products - Google recommends advertisers target all available products in one campaign which Google then manages. However this means all products will be treated the same, i.e. Google won’t factor into bidding the following: products on sale which an advertiser may want to push more aggressively, products with higher margins, high value products which you want more exposure on, or products with few sizes left in stock which you may want to pull back on, etc. It also wouldn’t know when products are on sale. 

• We often find that Smart Shopping spends more budget and drives higher CPCs on under-performing products. When visiting Reports > Pre-Defined Reports (Dimensions) > Shopping > Item ID, make sure that the top spending SKUs are the products driving the best return. We often find that higher ROI SKUs and the GA top selling products are not being aggressively pushed by Google. They will often find a number of SKUs which achieve around the ROAS target and then simply allocate the spend here, whereas there is much more opportunity to push products driving multiple sales at a strong ROI but only have a low impression share.  

Best Practice for eCommerce PPC

There are a number of ways to manage search text ad campaigns and although these normally follow the same principles as mentioned earlier in this document, best practice differs between businesses and agencies. In general, you should be looking to drive as much traffic as possible through exact match keywords. These exact match keywords (although the same principle applies for all match types) should be in highly relevant ad groups, where the ad copy and landing page closely reflects the keyword.

For higher value and high volume terms, advertisers often use a single keyword per ad group (SKAG) to ensure this is exactly included in the headline, description and display URL.

Ad Extensions have become much more significant over recent years, especially since these became a factor in quality score. These essentially expand your ad with additional information, which gives people more reasons to select your ad.

Below is an example of sitelink extensions used for Nike, which gives users the opportunity to click straight through to specific landing pages. In this case, Men’s or Women’s running shoes.

Nike have also used a structured snippet (Types: running shoes, shorts, tights, tops), a location extension and seller ratings in this ad. 

You can apply extensions at the account level, meaning the same messages are presented across all categories. However, it is much more effective to build out category level extensions ensuring users are served the most relevant messages. For example, sitelinks for an advertiser’s Dress category could be Party Dresses, Maxi Dresses, Midi Dresses, Day Dresses, etc. For specific campaigns e.g. Day Dresses, you could apply Summer Day Dresses, Occasion Day Dresses and Casual Day Dresses sitelinks etc. 

Pulling product level performance to optimise your products shown in Google Shopping

Data should be used from both Magento and Google Analytics for analysis and optimisation. Magento, along with GA, provide the overall product level performance across the site, rather than purely PPC results. Magento Business Intelligence (BI) also includes the added data of Time to Purchase and average Customer Lifetime Value (CLV). This data can then be fed into Google Ads, ensuring bids are increased on the keywords delivering the most value based on where the traffic is being driven. 

Summary

Google Shopping has become more and more important for retailers, now accounting for 80% of paid ad traffic, across non branded retail queries. For text ads, optimising within ad groups, targeting more specific and lower funnel keywords, and testing different ad messaging tailored to those search queries, is recommended. This article presents the most effective way to improve your Google Shopping performance via both feed optimisation and tactics within Google Ads.

Within your Shopping feed, the most important aspects to optimise are the product titles and descriptions. Google indexes product feed data to serve the most relevant product to a user’s query. Your product title and description are most important in serving your ads to either more queries or more specific queries. For product titles use higher volume terms (or the terms you wish to appear on) at the beginning. There are also specific Google best practices per industry that should be followed. Many clients pull in titles from their website which may limit coverage. A title such as Aeroplane Jumper for example, will rank a lot less often than Boys Cotton Jumper - Aeroplane Jumper for kids since more users search for boys jumper than aeroplane jumper. 

Product descriptions need building out as much as possible. Add in synonyms as well as general terms with high volume; for example Work Shoes, Smart Casual Shoes would not be in the title but would work well in the description, which customers wouldn't see. We advise using at least 1,000 characters per description to ensure you’re including as many relevant terms as possible. If you’re selling womens ripped jeans then you could include the below terms in the description:

Womens ripped jeans, womens jeans with rips, girls ripped jeans, womens blue ripped jeans, womens ripped jeans.

Your Google Ads structure can really impact performance. Split product types by ad group so that you can apply cross-match negative keywords. For example, adding Electric as a negative phrase match keyword to the Non-electric guitar ad group. You can control budgets, audience, location and device bids at campaign level and it typically makes sense to segment campaigns by the top-level product categories. A structure which allows you to filter more valuable searches (lower funnel/brand) into a campaign with a higher bid is the ultimate goal. You can then restrict spend on the campaigns appearing for more generic searches. 

You may also want to create a high- value campaign which is set as high priority. This will ensure that top sellers, products with the highest profit margin or new seasons, etc, are prioritised across your products. In terms of optimisation, product group bids and negative keywords will be the most significant changes to focus on.