Introduction

Google cites that advertisers can enjoy +30% incremental conversions from running performance max. Our view is that these gains can only be enjoyed if you’re doing a sub-optimal job of running your Google campaigns, and that better performance can be achieved by running standalone campaigns across Search/DSA, Standard Shopping, GDN, YouTube and Discovery.

What is Performance Max?

Performance max is a bottom-funnel advertising product serving on inventory across all Google properties e.g. Search, Shopping, YouTube, GDN, GMail and Discovery. Advertisers set a conversion KPI e.g. target ROAS, and Google fulfils this request by serving ads to users most likely to convert, e.g. relevant product term searchers, or those in remarketing lists. PMax doesn’t serve any purpose for mid or upper funnel objectives. Creative is auto-generated by Google, removing the barriers to entry for video and display marketing. In general, there are very limited options for optimising and managing the campaigns, and as such the campaigns serve two very clear purposes:

1. No frills option for advertisers with low account management resource e.g. in-house teams, or agency clients with low budgets
2. Fill otherwise unfilled Google advertising inventory, to continue greasing the wheels of the corporate monster

The key problems with Performance Max:

• The only way to opt out of particular channels is to manually request this from Google. The opt out lasts for 6 months, and you have to remember to make the request again, otherwise your ads start serving on those channels again.
• There is absolutely zero visibility of which channels your ads are serving on from a reporting perspective. If you have a search-first strategy with Google, metrics such as impression share are missing, robbing you of a key metric in sizing market opportunity. Spikes in impressions and clicks are commonplace, with no explanation of where they come from (probably display inventory).
• The auto-generated creative that Google comes up with is very poor quality, and unlikely to sit well with any self-respecting CMO or business owner.
• There is no visibility of which terms your ads serve against on search and shopping. As such, advertisers can’t set negative keywords at a product level, or analyse an incredibly useful set of data on what people are searching for.
• Campaign priority settings for shopping don’t exist, removing the ability to segment priority products, or run brand vs non-brand segmentation.
• In general, there are very limited options for optimising performance in the short-term. If you advertise for a business operating in a very dynamic environment (99% of businesses), it’s very hard to make quick changes to help hit business objectives.

There is an alternative solution however. At Propel Digital, our preferred approach to running campaigns across the Google suite is to run standalone campaigns across Search, Standard Shopping, YouTube, GDN and Discovery – coupled with Google’s powerful smart bidding technology. This enables the following:

• Full control and visibility over which channels your ads serve on.
• Full control over creative execution across all channels
• More options for controlling and manipulating performances

Example Tactical Execution:

• Search: overall goal is to consolidate campaigns as much as possible, without losing data relevance for the algorithm, and copy relevance within creative. Segmenting product categories with key business themes e.g. price, margin, brands etc. usually makes the most sense. Ad groups should be grouped semantically, to allow for the most relevant copy possible, with exact and broad match combined together. All campaigns should be supplemented with relevant dynamic search ad groups to mop up any remaining searches.
• Shopping: run your top 5% of margin-generating products (bestsellers) in a high priority standard shopping campaign, using single-product ad groups, optimised towards tROAS. Segment own-brand products into a separate high-priority campaign with a different tROAS target. Run all products with a focus of generating cash e.g. end of line in a medium priority max conversion value campaign, segmented by product type, and all remaining products in product type-focused low priority campaigns with an appropriate tROAS objective. 
• GDN, YouTube & Discovery: focus on targeting lower-funnel remarketing audiences, with ad groups segmented by product type pages visited. Attach your product feed to the relevant campaigns to serve for example dynamic banner ads on the GDN, or individual products overlaid on YouTube videos. Utilise tROAS or max conversion value bidding objectives across all campaigns, and focus on high quality creative executions across all formats.

The key drawback of the above strategy is the manual effort involved in achieving the required results. However we believe in the hands of marketing and platform experts, the performance gains are well worth it!

It’s worth noting that all businesses have unique objectives, and the sometimes unpredictable nature of the campaign algorithms means Performance Max may work better for some advertisers than others. The only way to determine the outcome is to test it. There is no perfect way of testing this, however our recommendation is to identify some large geolocations with similar historical performance and run the different campaign types in these places.

Summary

Final caveat is that we work in an ever-changing industry, driven by algorithms that improve over time as they ingest more data. Performance max is a relatively new product in the grand scheme of things, and in our experience Google’s advertising products tend to improve over time. As with smart bidding vs manual bidding, a tipping-point will surely arrive where a version of Performance Max outperforms all other campaign types.