More and more search marketers are cottoning on to the ROAS potential of Google Shopping or Product Listing Ads (PLAs). But do we really understand how to master Google Shopping and make our PLA campaigns profitable? In his talk, Christian Scharmüller introduced three dimensions search marketers can use to maximise their Google Shopping performance.
Image Credit: BigCommerce
- Campaign Granularity = Key
When we’re building campaigns for the Search network, it’s easy to organise ad groups and keywords together into a well-targeted structure in which to write relevant and engaging ad copy. But as Google Shopping does not allow you to bid explicitly at a keyword level and your ads are products instead, it can be less obvious to know what a well-targeted Shopping campaign structure looks like.
A common approach when setting up a Google Shopping campaign is organising by broad categories or brands in much the same way as you would in Search campaigns. Using conversion data from these categories or brands, bids can be adjusted to maximise profitability based on the ROAS of each. Sounds good, right?
Sadly, however, it’s not that simple. On the Shopping network this approach is far too broad: in each brands and categories there will almost always be clear winners but also certain products which are costing your campaign a lot without generating any sales, and therefore dragging down the performance of an entire brand or category.
When it comes to Google Shopping, as Christian Scharmüller explained, the key is granularity. Segmenting your Shopping campaign not by category or brand but instead by individual product stock keeping unit ID (or SKU) allows you to apply bids at a much more granular level, and gives you greater control over performance differences by individual products.
Christian even suggested splitting products out into individual ad groups or, after a period of data collection, breaking consistent ‘top performing’ product SKUs out into their own ad group in order to apply bids more competitively to these highly profitable items.
- Segment At Device Level
Google Shopping is primarily a mobile-first channel, and the click share now stands at over 50% on mobile devices compared to around 30% on desktop and less than 20% on tablet devices. But it’s often the case that although mobile is the key driver of traffic in Shopping campaigns, conversion rates are much lower than on desktop devices. This observation will often lead marketers turning towards bid adjustments at device level, reducing bids on mobile devices and countering this with more competitive bidding on high-converting desktop devices.
We can, however, take this a step further. Christian suggested experimenting with a device-targeting Shopping strategy by which campaigns are duplicated twice, and in each campaign two devices are entirely excluded using -100% bid adjustments. This approach would leave us with three campaigns: one mobile-only, one desktop-only and one tablet-only. Bids can then be applied accordingly based on historical performance of each device as well as individual product SKU performance, resulting in three highly controlled campaigns delivering profitable results.
- Search Query “Sculpting”
As Google Shopping doesn’t allow keyword targeting, it will match all users’ search terms to the most relevant product in your Shopping campaign. Assuming all search queries were created equal, this wouldn’t be a problem.
As search marketers, though, we know this simply isn’t the case. Different search queries are likely to result in very different ROAS, so when bids are reduced to combat poor ROAS for some search queries will also impact queries generating high ROAS if all searches are filtering into a single campaign. The question is, how do we combat this?
Christian outlined a fairly new Shopping campaign optimisation technique called search query sculpting, which involves identifying the different types of searches users are inputting to click on your Shopping ads and segmenting them by performance to form the basis of your Shopping campaign structure.
Then, using the Google Shopping campaign priority settings and creating lists of negative keywords in the AdWords Shared Library, search marketers can gain control over not only the types of searches filtering into each campaign, but the amount of spend channelled into them.
As the diagram below demonstrates, applying a ‘High’ priority setting and a Shared Library list of brand-related negative keywords to a campaign creates an area in which clicks from only generic, low-intent search queries will be filtered.
Branded queries will then be filtered into the ‘Medium’ priority campaign, having been blocked from showing in the ‘High’ priority campaign.
If there’s a third common query or theme of searches that your Shopping campaigns often attract and that generates a high ROAS, these can also be filtered into a third ‘Low’ priority campaign by adding these high-performing queries as negatives to the ‘Medium’ campaign.
Image Credit: Whoop App
Once you’re confident you have a search query filtering technique that will be profitable, you can apply bids on a priority basis depending on campaign performance. Applying lower bids to more generic queries but bidding more competitively in campaigns filtering branded or higher intent queries will channel spend into more profitable areas, driving the ROAS of your Google Shopping campaigns up.
Find the full slides to Christian’s talk here.
