As a software developer working in the finance industry, I had always wanted to do something more creative with my coding skills. Finally I took a break from my job and developed a game app, launching it on Android and iOS. I wrote blog posts about the development process and UX Design. Meanwhile, my friends downloaded Bug Hop and left reviews. I tweeted about it, got some retweets, submitted the game to app websites…But, in this crowded market, how could I really get my game out there?
Without a strong social media presence or marketing know-how, I decided to try Google and Apple’s advertising platforms. The setup on both is straight-forward, given a little reading up on terminology, but understanding how best to optimise ad potential is another matter.
After a number of beginner attempts, I decided it would be interesting to set up an ad campaign on both platforms and compare results.
Due to my lack of expertise, I wanted to keep things simple: accept default options, use proprietary tools, spend small. My aim was to collect enough data to create a baseline, identifying where improvements could be made next time.
I hope the below provides a helpful case study. I go through setting up each campaign, keyword optimisation and analysing the results.
How it works
First let’s look at the Cost Per Click (CPC) ad model, used by both Google and Apple. How does it work? When and what do you pay for?
Say, for example, someone looking for a new game app, goes to the Google Play Store and types popular puzzle games in the search bar. This will return a list of puzzle apps but the top of the list will be reserved for ads. To decide what ads should be shown, the search phrase will be matched against ad campaigns that are associated with puzzle games. An auction is run between the campaigns — or more precisely, the algorithms bidding on their behalf. The highest bid wins the space and gets to show its ad. This is considered a cost effective form of advertising, as the campaign only pays for the ad if the viewer clicks on it.
Bid prices can be set by the campaign or calculated by the algorithm. Note that when an ad is clicked on, the price paid is not the winning price but the second highest competing price. In that way, campaign spending is not more than necessary.
Keyword optimization
A crucial component of CPC advertising is accurately matching search phrases with campaigns, so that your ads are shown to the right audience.
This table lists phrases I thought described my product. Bug Hop is a game which tells a story over a series of puzzles. It’s fun, mentally challenging and tests problem-solving skills. When any of these phrases were searched on, I wanted my ad to be shown. This data was gathered using Google’s Keyword Planner.
We can see how often these phrases are entered, how much competition there is for ad space and the bid price that wins the top spot.
Generating keywords
There is a lot of information online around how to identify keywords for your product. The advice given to me was: Ideally, to start with, you want to find keywords that are a good match but also don’t have high competition. So when someone enters those words into a search engine, there’s a strong chance they’ll see your ad.
To create my list above, I found these steps helpful:
- Make a list of keywords and enter them into Google’s Keyword Planner, or a similar ASO tool, to see how popular they are. I removed some phrases from my original list as they didn’t have any occurrences.
- The Keyword Planner also allows you to specify a website URL to see what searches led to that site. I entered URLs of puzzle games similar to Bug Hop, like Candy Crush. Most of those searches included the words "candy crush", so people are looking specifically for that game. But the search phrase puzzle gaming also led to the Candy Crush page. It has a high occurrence and would be apt for Bug Hop too.
- Type phrases into a search engine, such as ‘best game app for adults’. With predictive text, you can see how popular a phrase is.
- Look at the landing page or website of similar products to see what phrases they use in the blurb.
Android ad
Google has multiple ad setups: Search Ads allow the advertiser the most control, over keywords, bidding and other criteria. Smart Campaigns are for non-experts who want Google to handle everything. App ads, which I used, are a third category, similar to Smart Campaigns.
You can create multiple ad groups in one campaign, to experiment with different content. Keeping things simple, I set up one group.
Google allow you to set a daily average budget. I set this to £2. It can be exceeded, depending on activity but the overall cost of the campaign should not exceed the average multiplied by the average days in the month. So that’s £60 here. Far from the budget of a professional campaign which would run into the thousands.
Google asked for multiple ad captions, descriptions and images. These would be used to create varying ad combinations and I would be able to see which content was most popular.
Finally I set location and duration. The campaign would run in five English speaking countries for the month of January 2021.
These sample ads were generated:
Google’s Search Ad campaign allows you to set keywords and corresponding bid prices but the App ad campaign does not. Instead, Google handles keyword matching on the campaign’s behalf. For example, if someone enters a phrase in a search bar which is also on an app’s product page, the bidding algorithm makes this match. So before going live with my campaign I needed to update my app’s play store description with the most popular words, taken from my keyword list above.
iOS ad
Apple Search ads provide two plans: Basic and Advanced. Basic is the most cost effective because you only pay for an ad if conversion occurs, which means I would only pay if Bug Hop was downloaded. But there is limited keyword optimization so no guarantee your ads will be shown to the right audience. The advanced option which I chose is similar to the Google App campaign.
Apple allow you to set a daily max spend. This is different from Google’s daily average spend. I chose £3.20. With Apple you can set an explicit campaign budget. I set £60 to equal the Google side. I also set the same duration and location.
Again, multiple ad groups can be created with different content. I set up one. Apple don’t allow you to add new images and captions. Instead they generate ads using existing content from your store page.
Here is an example of what the ads would look like.
Apple group images together in Creative Sets. You can define your own and then see which images are most successful in getting clicks. I made two Creative Sets, in addition to the default one.
Unlike Google's App campaign, Apple allow you to provide keywords and associated bids which I did, referring to my keyword list. I set bids competitively where I felt phrases were most relevant to my product. You can also enable Apple's Search Match facility, similar to Google's keyword matching. This is a good idea because Apple share keyword results so you can see what worked.
Results
At the end of my one month campaign, here are the metrics I am interested in. I can't show exact numbers due to confidentiality, but you can see the differences between Android and iOS performance.
Impressions
An impression is when the ad is shown to a customer. I had four times as many impressions with Google compared to Apple.
The main reason for this is that Google doesn't allow app advertisers to set bid prices. Instead pricing is controlled by an algorithm which can keep bidding low. The Apple Advanced campaign does allow advertisers set bid prices and this appears to drive the price up. Furthermore, iOS apps tend to be more expensive than Android apps, so advertisers could be willing to pay more for ad space. Therefore, the same budget goes a lot further with Google.
Note that the data supplied by the Keyword Planner, which included bid prices, is derived from Google Search campaigns. Those bid prices are set by advertisers, not algorithms. This is an important distinction not made clear on the platform.
There are other factors to consider:
- Google does all the keyword matching on my behalf using a number of sophisticated techniques. Apple does some matching on my behalf and some matching using my keyword list. Perhaps my list needs developing.
- Maybe my iOS campaign wasn’t always bidding high enough to win ad space. But results throughout the campaign did not indicate my bids were too low.
- We might consider Apple is being more discerning with who it shows ads to. But if Apple were using more accurate targeting, then those ads would attract relatively more clicks. As we can see, the click through rate (CTR) is actually higher for Android.
- While the highest bid should win the ad space, remember the platform only gets paid if the user clicks on an ad. A competing campaign may be offering a lower bid but if its ad is more successful it might be more attractive to the platform. So low success with click through rate can negatively impact impressions.
Click Through Rate
Expressed in percent, this captures the number of times your ad is clicked out of the number of impressions shown. A high CTR is a good indication your ad is reaching the right audience. Apple calls this metric Tap through Rate (TTR).
It’s hard to say what would be considered a good CTR. It varies for different products and on different platforms and at different times of year. I am seeing very different benchmarks across websites.
My Google CTR was over 6% which would generally be considered a healthy number. But why was it higher than my Apple TTR?
- It's possible the iOS ad is not being shown to the right audience. Note my keyword analysis was done on the Google platform. It looks like platform-specific analysis could have helped identify more effective keywords for Apple, which I could have added to my campaign and store page.
- Currently, I don’t have ratings on my iOS store page but I do on my Android page. This gives my Android ad an advantage. Those stars are eye-catching!
- For the Apple campaign, the ad caption is truncated, as you can see in the sample. Google ask for specific ad captions with a character limit, so truncation doesn't happen.
- Google allow you to supply multiple captions and images within the one ad group. With more variety in content, there’s a better chance of creating a successful ad.
Conversion Rate
This refers to the rate at which my app was downloaded out of how many times the ads were clicked on. A good conversion rate reflects positively on the product page.
I am seeing wildly varying estimates on what constitutes a good rate. My numbers were healthy on both platforms. This makes sense because the ad material matches the product pages. So if people like the ad there's a good chance they'll like the product. But we see a stronger result for Google. I will look at improving my Apple product page next time.
Overall Cost
Android £61.36
iOS £60
iOS £60
The iOS campaign was frozen on the 18th January as my budget was already exhausted. Why did I run out of money so quickly?
For my iOS ads I set a daily max of £3.20 with an expectation sometimes it could be less. For Android I set a daily average to £2, with the expectation it could be more. In fact both daily spends were consistently on these targets so my iOS budget ran out earlier.
Average Cost Per Click
This value is the overall cost divided by the number of clicks.
Looking online, I have found the average iOS CPC (or Cost Per Tap CPT) is quoted at $0.84. Mine was a bit high.
With Apple's greater campaign flexibility there's more room for error. So some experimentation is definitely needed. For example, Apple allow you to set a daily target Cost per Acquisition (or Cost per Conversion). This adds another control factor, along with the max daily spend and individual keyword bids. By combining different controls over multiple ad groups, a campaign can learn which works best for a particular product and market.
Average Cost Per Conversion
This value is the overall cost divided by the number of downloads.
According to an article referenced below, the global average CPC is estimated at $0.44 for Android and $0.86 for iOS. These figures vary depending on ad type, style and placement etc and should just be used as a guideline. I can share that my Android CPC was £0.35.
My app is free to download. Ads are shown every 4 minutes. My eCPM is £0.44, which means I earn £0.44 for every 1000 ads shown in my app. In order to earn back £0.35 from one download, someone would have to play Bug Hop for about 50 hours. Alternatively, an ad-free upgrade is available for £1.19.
With word-of-mouth factored in, it’s possible I could make the cost of conversion back on Android. I have been advised to increase my upgrade price and also to experiment with ad types. But improving monetisation is topic for another day!
Keyword Results
I really appreciate Apple’s transparency, providing specific keyword and search phrase results. I can't share full details, but the above table relates to my original list of keywords. Google don’t share these results for an app ad campaign. You would have to set up Firebase Google Analytics to see traffic to your product page.
The highest CTR goes to problem solving game for kids. But in fact I got no installs from these click-throughs. This might indicate people looking for kids puzzles initially liked the look of the Bug Hop ad, but on seeing the product page, decided against it. That connects with a concern I have: Bug Hop is not a kids' game per se but it does have child-like appeal. Does this mean I have a category problem? There wasn't enough data to draw a conclusion. But it does give me an action point for my next game: Have a clearer idea of what audience I am trying to reach and what the successful games in that category look like.
Brain teasers and brain puzzle yielded some interest but didn’t actually result in downloads. Maybe when people search for brain games they want something mathematical or physics related.
Games like Candy Crush didn't do well. But Apple's matching algorithm showed the Bug Hop ad when similar puzzle games were searched for. As an example, some of my installs came from a search for "dop 2". Looking at the DOP product pages, I see similar words to Bug Hop's page, like brain, puzzle and solve, so Apple's algorithm would have put this together. Again, knowing what keywords similar apps use is really worth while.
My ad was shown to people looking for word puzzles. Word puzzles are associated with brain teasers so there is a keyword match but this is really a mismatch. It shows a need for a negative keyword list. I would create one next time, adding words like crosswords, word puzzles etc.
Some of my keywords didn't result in any matches. This highlights the point about using a platform-specific ASO tool to estimate keyword traffic and bid price.
Content
Google allowed me to set rotating ad captions. Here are the results.
There's a clear winner which I've now set as my app store caption!
Apple allowed me to organise collections of images in Creative Sets. As expected a clear winner did not emerge because my content is not that varied. This option would be good for someone with a different look and feel at each level of a game, so you could see if one set drew more clicks.
Conclusion
Both platforms have been easy to work with and there’s plenty of information online for guidance. But the Google campaign has been much better value. Their algorithm to match customers with ads is more active and more accurate. They also control bidding price, keeping it low. While they are secretive about methods and don’t allow much control, with very little experience and a small budget, I can set up an effective ad campaign. Kudos to Google!
Apple boasts a competitive conversion rate and cost per install. They might argue that using the advanced option requires expertise, which is fair. But between a limited basic option and a flexible advanced option, it makes it hard for a non-expert to create a viable campaign. However, their transparency with results means you can learn and improve your strategy each time.
I would love to hear from you if you have had a similar experience, have tips to share, or if you think I’ve overlooked something in my analysis. Like most developers trying to promote their app, I’m on a learning curve, relying on blog posts, forums and platform support to figure it all out.
References
(Featured image by Charles Deluvio on Unsplash)