Google Ads is in a state of transition, due to technological developments in machine learning and AI and partly because of privacy concerns that users and large governments bodies are making Google very aware of.
Performance Max campaigns is Google’s latest development on their advertising platform.
On paper, a Performance Max campaign looks like a hugely simplified campaign type. It’s reminiscent of the old “smart” campaign Google developed for ad novices in previous years. An all-encompassing one-size fit’s all campaign.
Digging a little deeper however, it would appear that these campaign are far from simple and far from dumb.
What is a Google Ads Performance Max Campaign & How Does It Work?
Performance Max campaign Credit: Google
A Performance Max campaign is the ultimate machine learning experiment.
You are the advertiser will fill the campaign with all the creative, copy and assets you can get your hands on; from headlines and descriptions to images and video.
From this, rather than creating a campaign for each type of Google advert such as Search, Display, Gmail or YouTube.
Instead it will create adverts which can be be presented on all of these platforms.
You tell Google what goal you want to achieve e.g. leads, sales or phone calls etc and you leave the machine learning to do all the hard work.
Pooling of all ad assets to be used across the Google ecosystem.
To quote Lord of the Rings; “One campaign to bind them. One campaign to rule them all”.
How Much Control Do You Have?
Not a lot.
Even some of the bidding adjustments we are used to seeing elsewhere are not something we can mess with on a Performance Max campaign.
You a limited to bidding adjustments by location only.
You also still get control of ad scheduling.
That’s about it. Everything else is driving through the machine learning bidding which is implemented.
I don’t particular have a problem with this as this is exactly what machine learning is supposed to take care of. Allow the data to run the campaigns.
How Much Data Do You Get?
You get very little outside of top-level numbers and nearly no ability to segment this data by platform.
This makes the top level figures difficult to base anything off, as if you mix search, display, Gmail and YouTube all into the same figures; you are gonna get some high impressions and low conversion rates.
Perhaps, this simply means I need to reset all the internal Google Ad barometers I’ve got in my head, but I would like some more feedback.
Some of the reporting you do get is a little, pointless.
I don’t need to see the best combination of assets which Google has generated, I don’t see how this information benefits the advertiser.
They do provide you with some non-quantative feedback on the assets you provide which give you a little feedback on what to chop and change in your assets groups.
However, this is limited to a “Low / Good / Best” feedback.
Feedback on ad assets in asset groups.
Performance Max: Shopping & Non-Shopping
Performance Max truly are the one-size fit’s all campaign, as you can also feed in your Google Merchant Centre feed and utilise the power of Google Shopping ads too.
This is where I personally have been using Performance Max campaigns mostly and getting sizeable results.
It would appear that Smart Shopping was merely a stepping stone towards Performance Max campaigns, and soon to succeed them by the end of September 2022.
Performance Max is an evolution of this and opens up the same possibilities into a non-shopping type campaign.
Smart Shopping vs Performance Max?
If you are finding good performance in Smart Shopping campaigns, I would very much advise you to run a variant test on Performance Max campaign, however be warned that in my own experience I found that like for like Smart Shopping and Performance Max campaigns don’t like to co-exist, therefore you’ll need to run your Performance Max campaign independently.
I found when trying to run these together the Smart Shopping campaign stopped serving completely.
This might put you off given that this may result in short term reduced results, however I’ve found from my own experiments, the results outweigh this.
Typically it takes a Performance Max campaign a few days to get itself up to spee, I believe that this will be very much dependant on budget and then it had outperformed all other campaigns put against it.
One Large Campaign vs Segmentation
It may be tempting to, give the campaign type’s nature to only have a single campaign running for your whole account.
I’d strongly advise against this. For ecommerce I’ve found that segmenting out Performance Max campaigns has helped my performance.
I’ve segmented out my campaigns by product type / category or my brand priority etc. This has allowed me to allocated relevant budgets to each campaign, and gain a better ROAS.
For a non-ecommerce, I would most likely segment my campaign by lead generation type or service to achieve the same outcome.
Audience Signals
Part of the set up process of Performance Max campaigns is provide Google with some audience signals to get a clearer idea of the type of audience you are looking for.
Which is a neat trick and a nice evolution of the “observation” and “targetting” audience lists we’ve got.
The more you can fill Google’s machine learning with pure data and untarnished daya then the better results you will see.
Google themselves have come out and said that this function is merely to give Google a better run-in to the campaign. Once it’s done that, it will use it’s own data collected to start gaining better results.
Which is equally a very smart way of creating a campaign and giving advertisers exactly what they want.
You can use any mix of audience signal that you want, however I would advise to try and keep this as broad as possible and let Google work out the value in the audience.
I typically include a broad array of interests audiences based off the product and service alongside an audience of past converters which has been set up in Google Ads. If you don’t have this audience yet, set it up.
Get The Assets You Need
I believe that the getting the most out of these campaigns will rely on getting your assets right.
Google provide you with feedback on your headlines and descriptions that you will need to continually optimise.
Also, you images and graphics will need updating often to avoid ad fatigue over a period, much like when running social ad campaigns.
Poor visual assets such as images and video are simply going to result in poor results and ads.
Google do automatically generate video ads for you if you arn’t able to provide them, however I would avoid at all cost. They are low quality and templated. I’ve even found some large errors.
At all costs try to provide your own professionally created video assets.
This is an ending splash logo page on a auto-generated performance max video ad – it’s awful.
Results
In my small sample of Performance Max campaigns so far, I’ve had very good results.
However, it’s worth noting that I’ve quite limited to ecommerce campaigns as it stands.
It would appear to me that these campaigns are being manually weighted by Google to provide better results that other campaigns in the algorithm. This will be a mechanism to get more advertisers onto the campaigns.
Which will work, as all advertisers will want to find that cutting edge.
We will see how long this sustains and also how these work on lead-gen type campaigns, which I will have more involvement with in the next couple of months.
I know in the Google Ads industry there is a mis-trust in using Google’s automated bidding strategies. I personally feel that all their testing and learning has come to the new campaign type.
I believe this is a big change for Google Ads as a platform and the only reason to be scared about it is because we believe that we lose granular control.
However I don’t believe that is the case, it merely required a difference kind of management and a focus less on the “dials” and “buttons” but more of the messaging and the assets.