When it comes to common practices in digital marketing, practitioners should embrace a “trust, but verify” mindset. Specifically, adopt best practices when possible, but don’t assume that they’ll improve performance – every account exists in a different context with a multitude of different intervening variables that can affect the impact of any initiative.
One such best practice that deserves interrogation is the adoption of promotions in ad copy in Google Ads. As advertisers, we’re taught that conversions are a result of putting the right message in front of the right person at the right time, and it is often assumed that time-sensitive promotions highlighting cost-savings can contribute to the “right message” part of the equation.
But does that data support such an assertion? For this case study, I’ll be analyzing a client that leases multi-family luxury apartments across the United States. Those properties will periodically run promotions, most of which advertise a certain number of months free (usually 1 or 2) when the renter leases within a specific timeframe. So, for example, one promotion that was included in ad copy had a first description line that read, “Up To 2.5 Months Rent Free. Contact The Leasing Office For Details.” The ads were exclusively ETAs, and, depending on space considerations, the promotional offers appear in either the headlines, the description lines, or some combination thereof.
For this client, when a promotion goes live, new ads are created, labeled and inserted into existing ad groups to compete against the evergreen ads. The ads are set to “optimize for performance” rather than “rotate evenly”. New ads were inserted throughout the property’s campaigns and ad groups.
To collect the data for this case study, I pulled the last 180 days of data for apartment communities that ran promotion ads and compared the performance of those promotion ads to the evergreen ads only for the dates which the evergreen ads ran. So, for example, if an apartment community had promotion ads live from January 15th, 2019 to February 15th, 2019, the data used for this case study would only be taken from that period – for both the promotion and non-promotion ads. In other words, in order to better control for seasonality, I’m not comparing promotion ads to evergreen ads for periods in which the promotion ads did not run.
In order to determine whether or not the promotion ads “won” we compared the ads’ cost-per-conversion, which is the KPI most important to the client.
I’ll be examining results both in the aggregate and the apartment-community-level data to determine whether or not the promotion ad copy has been effective and whether or not it should continue to be used in the future. First, let’s take a look at the aggregate results, which combines the data from 29 different apartment communities that had run promotions in the past 180 days:
The initial results look pretty resounding in favor of the evergreen ads. The CPCs for the promotion ads are significantly higher, the click-through-rates significantly lower, and, most importantly, the CPL is substantially higher. However, our ultimate goal here is to determine whether and how to highlight promotions in the future – to answer that question, we’ll need to go a little deeper into the property-by-property data. Ultimately, even though the aggregate data looks pretty resounding, if there are a few outlier apartment communities where the evergreen ads vastly outperformed the promotion ads, we may still want to utilize promotion ads for most properties in the future. The table below includes a count of apartment communities that saw lower CPLs for their promotion ads:
Still, it looks as though the evergreen ads were overwhelming more effective than the promo ads. 21 out of 29 (72%) properties saw their evergreen ads outperform the promo ads.
Before calling the promotion ads a failure, let’s take a look at one more dimension by limiting our data to non-brand campaigns. Since we would expect these sort of promotions to be more effective users who weren’t searching for a specific apartment community but instead were in the comparison stage of the customer journey, it’s possible that the promotion ads would out-perform the evergreen ads for non-branded searches. The table below has data from branded searches removed:
Looks like that hypothesis may have been partially correct, although not to the degree that the promo ads are the winner in the aggregate. While the metrics are much closer than they were when branded searches were included, we still see higher CPLs from the promotion ads. On the other hand, in this comparison, there were more properties that saw performance boosts from the promotion ads in their non-brand campaigns.
So what gives? Why didn’t the promotion ads improve performance? I have a couple of theories:
- Perhaps because this particular client exists in the luxury rental industry, price-point is a relatively smaller concern for their target demographic. Most everyone likes a deal, but for some, that deal might be less alluring than if that space had been used to advertise the posh features and amenities of their potential new home.
- Perhaps the relatively short time frame that these ads ran put them at a disadvantage algorithmically. Remember, our ad rotation has been set to optimize for performance, and these ads would have had less historical data behind them than the evergreen ads. Could that have affected the system’s ability to enter the ads in the most effective auctions?
So, for this specific client, it looks like the best practice of advertising promotions does not always apply. In the future, I’d love to dig deeper into how long running a promotion affects performance, and also the performance of promotion extensions throughout the account. Still, this initial analysis should yield some actionable insights that we can apply right away.
Whether or not you work with promotion ads, I hope this case study serves as a reminder to interrogate common practices to make sure that the practice is best for your context. Remember – trust, but verify!
Thanks for reading! Reach out with questions or feedback to @ppchero!