Change is hard. (I think we can all agree to that.)
As digital advertisers, change is also constant – new features, new ways to target, and new ad types.
Not exactly new, yet it would seem that everyone is still figuring out how (or if) to proceed with RSAs.
This trepidation is why I decided to take a look at RSAs.
In my view, responsive search ads are a fascinating opportunity.
Just write up a bunch of headlines and ad descriptions and whammo – up to 40,000 potential ad permutations.
On the backend, all of those headlines and ad descriptions go into a machine learning soup with my keywords, ad extensions, and all the juicy back-end-data that most of us “normal” folks never get to see.
When someone enters a search query, the right message is delivered at the right time for that person.
Does this require some trust in Google and Microsoft’s machine learning capabilities?
Does it work?
The word on the street is this:
If you follow some basic best practices and understand how to analyze RSA performance, yes.
Responsive Search Ads: By the Numbers
Early on, there was frustration voiced that RSAs took away control and didn’t exactly live up to promised performance improvement.
Folks saw click-through rate (CTR) increase, but didn’t see conversion rates or cost-per-acquisition (CPA) improve compared to expanded text ads.
More recently, there has been discussion and data shared that shows RSAs driving conversion performance gains.
This is a trend reflected in the data from Microsoft Advertising’s responsive search ads, too.
In data from January 2020, RSAs on Microsoft Advertising were showing:
- A 10% increase in volume.
- A 6% increase in conversion rate.
- A 7% decrease in cost-per-acquisition.
It goes deeper, though.
Where many advertisers are getting stuck is in looking at individual RSAs to determine performance gains.
Instead, the performance should be analyzed at the ad group level where the sum of RSA + expanded text ads will be reflected.
10 Best Practices for Running Responsive Search Ads on Microsoft Advertising
1. Allot Time to Gather Ample Data
Because RSAs run on machine learning, they require time to reach statistical significance with regards to performance.
You can’t expect to launch a new Responsive Search Ad and determine if it worked well in a matter of days.
Time should be counted in weeks – this can and should vary depending on an ad group’s traffic volume.
2. Avoid Comparing RSAs to Expanded Text Ads
RSAs require you to look at bigger-picture trends and compare data.
Put another way, don’t settle for comparing an RSA to an expanded text ad as a means to determine success.
Look at the combined performance within an ad group before-and-after the responsive search ad was launched.
3. Take Advantage of Campaign Experiments
Want to be absolutely sure that you have statistical significance before declaring whether RSAs work for your campaigns?
Take advantage of Campaign Experiments – “A” (control) has no RSA, “B” (variant) has RSA.
Wash, rinse, and repeat.
4. Marry RSAs With Automated Bid Strategies
Remember that bit about Responsive Search Ads improving conversion performance?
To really amplify this effect, marry RSAs with automated bid strategies like Maximize Conversions or Target CPA.
5. Run Both Expanded Text Ads & RSAs in Your Ad Groups
Be sure to run both expanded text ads and RSAs in your ad groups.
It is worth noting that RSAs are a safe bet.
If the responsive search ad isn’t the best option for a given auction, an expanded text ad will be shown.
A recommended mix to get started is 2 to 3 expanded text ads and at least one RSA (keep the ad group to 5 or less total ads).
6. You Can Pin Headlines & Ad Descriptions in a Specific Order
Is the fact that your headlines and ad descriptions will be mixed-and-matched making you uneasy about running RSAs?
While the best practice is to let said mixing-and-matching happen on your behalf, you can choose to pin headlines and ad descriptions in a specific order.
Just be aware that pinning restricts the number of headline/ad description combinations that can be created on your behalf.
7. Lean Into Machine Learning
Lean into the machine learning behind responsive search ads by loading up plenty of headlines and ad descriptions.
Aim for a minimum of 8 headlines (10+ if you can) and 3 ad descriptions.
8. Don’t Repeat (Or Nearly Repeat) the Same Content Across Headlines & Ad Descriptions
Load up those headlines and ad descriptions, but do so with creativity.
- How can you lay out your product or service benefits and features?
- Have you clearly stated a call to action?
- Did you include shipping or return information?
- Be clear, be succinct, but most of all be creative.
If you need help with this, there’s a clever “Ad strength” meter next to the RSA in the UI.
The better job you do of writing headlines and ad descriptions, the higher the ad strength will go (poor to great!).
9. Leverage the ‘View Assets Details’ Reporting
Leverage the ‘View assets details’ reporting to determine what is working in your RSA’s.
- Assets: This will show you impressions by individual headlines or ad descriptions.
- Combinations: This will show you impressions by top combinations of headlines and ad descriptions.
Impression data is admittedly directional at best, but it can be a signal to determine what is being shown most frequently.
If a headline or ad description has low impressions compared to others, this is a signal that the system has deemed that asset as irrelevant or driving down other performance metrics.
Take that as a cue to modify and adjust.
10. Import Your Google Ads RSAs to Microsoft Advertising
Already running RSAs on Google Ads?
Seamlessly pull them over to Microsoft Advertising with the Google Import feature in the UI or desktop editor.
There is plenty to like about Responsive Search Ads – sophisticated use of machine learning, ad copy expansion, and testing at scale… but perhaps most importantly, an opportunity to increase conversion rates.
Admittedly, there is still room to grow for RSA reporting and data clarity.
Progress continues to be made on that front and I expect that trend to continue as more advertisers lean in and launch more responsive search ads throughout 2020.