83% increase in ROAS with Apple Search Ads

In four months, we increased ThinkUp’s ROAS by 83%.

With the Internet spreading linke wildfire and reaching every part of our daily life, more and more traffic is directed to websites in search for information.

The first step was to understand user sentiment and values within the meditation/affirmation space by performing market and competitor research. We also reviewed prior organic keyword performance to identify the top terms with the highest conversion rate (CR).

After prior keyword performance was evaluated, we performed an in-depth campaign audit to identify low hanging fruit for immediate campaign optimizations. The ASA campaigns were set up based on best practices in 4 separate campaign types (Brand, Competitor, Discovery, Generic).

In the first two weeks of the ASA campaign, we conducted specific tests to monitor and evaluate elements of the campaign that yielded the best results. As part of this process, we increased daily caps significantly and placed a high default CPT.

By doing so, we were able to capture surges on priority keywords, and observe key activity like auction eligibility, install costs, and conversion rates.

For best long-term results, the AGN team closely monitored keyword performance and optimized campaigns on a weekly basis. For ThinkUp, our strategy to boost visibility of keywords with low impression was to increase bids by 20% each round.

The rationale behind this is that sufficient traffic needs to be generated to find the optimum level and — with low spend on keywords — it’s hard to determine if a term is successful/unsuccessful or just hasn’t been presented in front of a statistically significant audience.

As such, we adjusted CPT bids based on top/mid/bottom range converters, spend wasters, and late converters. To decrease ad spend, non-performing/irrelevant terms were paused and added as negative keywords.

One of the most efficient ways to monitor ad spend for Apple Search Ads is by leveraging ASA automation tools. For ThinkUp, we set up daily alerts on keywords with:

We generated weekly reports to gauge impression share, measure impact, and protect branded terms. From those reports, we adjusted bids accordingly with the following automation rules:

Tactics were evaluated and executed based on the type of ASA campaign (Brand, Discovery, Competitor, Generic), as each is different and may require a different approach for optimum performance.

Brand Campaign

■ Keywords were untouched and only increased if we lost impression share as we typically assume that branded terms will always have a strong ROAS

Discovery Campaign

■ This campaigns was used to mine keywords for the competitor/generic campaigns — keywords which performed well were moved, while keywords which didn’t perform well were added as negative terms in order to manage spend (stop cannibalization)

Competitor Campaign

■ This was used to boost brand awareness especially against those who were bidding on ThinkUp’s own branded terms around the holiday/new years resolution season
■ Because these types of campaigns are typically the most expensive, eventually only terms which were bringing in revenue were kept active, regardless of /installsCPA in order to try and balance ROAS overall

Generic Campaign

■ Was useful in aligning with our ASO strategy