In a previous post, I talked about how to set up Apple’s new iAd Workbench, which can be the tool that indie developers and designers are looking for to get more exposure to their apps. In this post, I am going to talk about my own initial marketing efforts and show you my actual sales numbers. I will then show you all the numbers and charts from my iAd Workbench campaign performance results and provide my own analysis.
Launch Marketing Efforts
Badminton Central Forum
I have been a long time member on this forum and read a lot of posts, but I’ve never really contributed much to forum discussions. After I released Badminton TV, I just started a thread to announce its availability. The announcement includes the app icon, the App Store logo, the best two screenshots, and links to a dedicated page and announcement post on my blog. A few people immediately responded with good feedback and I wrote back.
I also posted an API request graph to show its popularity on the third day to try to get more people interested.
Badminton Thoughts Blog
Since I really wanted to test iAd Workbench as soon as possible, I had to push the release date at least a week earlier than I would have liked. So I hastily put together a dedicated page for Badminton TV with just a few paragraphs, app icon, the App Store badge, all the screenshots I had uploaded to the App Store. I also added the app icon in the sidebar so that it is visible on every page in the blog.
I figured I should also write an announcement post for regular readers to get the news. It is similar to the dedicated page, except I included fewer screenshots and altered the copywriting. I also included a link to the dedicated page for people who want more information before they decide to download the app.
I played in a few local badminton clubs and am a member in a couple Facebook groups. I have been posting new articles from my blog to these groups for several months. So, as usual, I posted the announcement post for Badminton TV to these two Facebook groups. Club members with an iPhone or iPad can therefore download the app.
I also asked a local official Yonex dealer to post a link to Badminton TV in his company’s Facebook page. I bought stuff from him before and he is the Yonex guy to go to in this area.
I didn’t use Twitter much at all because very few badminton players are Twitter savvy, or even tech savvy. I just don’t see a lot of conversions for the amount of effort I need to put in there. My announcement post was already auto-tweeted when it was published in WordPress. I also tweeted an API request graph to show its popularity in an attempt to make it appealing to potential users, and again after I added videos for a recently completed tournament.
iTunes Connect Sales Data
I created a Sales user in iTunes Connect and used App Annie to pull sales data from iTunes Connect. I found App Annie provides a better UI to extract useful information than iTunes Connect itself.
Let’s first look at the actual sales numbers for Badminton TV in the first week. Total number of downloads is 314 from 6/11 to 6/18. As expected, most sales occurred during the first day when it was launched due to the initial marketing effort and the newness in the App Store. Sales then slowed down. There was a spike over the weekend because I added the videos for the semifinals and finals for the Indonesia Open and posted about it in the Badminton Central thread. People wanted to watch matches after they had just concluded.
Badminton is not a popular sport in the US, so I expected there would be a lot of sales outside the US. It turns out to be the case. If you look at the sales chart and table below, you can see the US still has the most sales. I believe this is because the US market is simply the largest iOS market in the world by far. Sales are then dominated by the badminton crazy countries in Asia and Europe. This is not surprising.
iAd Workbench Performance
If you want a quick glance at my iAd Workbench numbers, here is the summary for the overall performance.
My iAd Workbench campaign starts on 6/11, so I started to accumulate Spend on 6/11 for a total of $74.66 (lower than my initial maximum budget of $100.00). However, the Spend doesn’t approach the Daily Spend Goal of $12.50 until 6/14. I suspect this is due to timezone difference. From 6/14 to 6/18, my Spend hovers around my Daily Spend Goal of $12.50 pretty consistently.
The following shows the number of impressions I got every day.
The following shows the number of taps I got every day. I can see there is a correlation between Spend and Taps. It seems that the more I spend on a day, the more taps I get. I didn’t spend up to $12.50 the first 2 days of my campaign, so I got fewer taps.
My campaign generated 0 downloads, so it is 0 every day.
I used CPA for my campaign. Since there is 0 download, the CPA also remains $0.00.
This is interesting. I didn’t use CPC, but CPC remains relatively flat at around $0.40. It isn’t $0.00.
TTR stands for Tap Through Rate. It is calculated from dividing Taps by Impressions. Taps remain relatively flat after the first 2 days, but impressions were lower on 6/14 and 6/17. Hence the TTR is higher on those 2 days.
Except the first 2 days, there isn’t much correlation between Spend and Impressions.
However, it seems the more money I spend, the more taps I got.
There isn’t much correlation between Spend and TTR just like Spend and Impressions.
The following chart shows that number of Impressions does not translate to number of Taps. It seems my campaign is more cost effective on 6/14 and 6/17.
Again, the TTR is higher on 6/14 and 6/17, compared to number of Impressions.
I am not sure if it makes any sense to compare Taps against TTR, but here it is.
The following table shows my daily numbers.
I got an iAd Customer Billing Notification email from Apple on 6/17 telling me I will be charged $62.23.
Analysis and Hypothesis
Let’s first compare the actual iAd Workbench campaign performance with the performance estimated by Apple when I set the budget:
|Actual||Estimated||Does actual match expected?|
The TTR is calculated by dividing the number of taps by the number of impressions.
- Actual TTR: 1,885 / 765,378 = 0.25%
- Estimated TTR: 2,400 / 790,000 = 0.30%
The actual number of Impressions and Taps seem to match or exceed the expected numbers. However, the TTR falls short but it isn’t too bad compared to number of Downloads. iAd Workbench simply does not generate downloads. This is hugely disappointing as I just spent $74.66 to acquire 0 users.
It can be argued that iAd Workbench is rather new and Apple can and may tweak things to improve the performance, but I don’t know how much they can increase from 0 downloads. Maybe people have become ads-averse even on iOS devices that they develop an instinct to skip ads. But if the number of taps are accurate, people do tap on my iAd banner but they just don’t download them. Maybe people were made aware of my app but decided to download it at a later time due to various reasons. There isn’t enough information to tell the whole story. In short, Apple’s audience targeting estimator works well for Impressions and Taps, falls a little short for TTR, and simply does not generate sales or Downloads.
Several things from this data also jump out at me as questionable:
- Apple seems to push my Spend up to my Daily Spend Goal of $12.50 regardless of how many Impressions, Taps, and Downloads my app gets. It feels like I was telling Apple how much and often I want them to charge my credit card, and they do a good job of that but does not deliver the most important result which is user acquisitions.
- I spent only $0.10 on 6/11, the first day of my campaign. But I set it up on 6/11. So it didn’t get to run for the full day.
- 2,557 Impressions for $0.10 isn’t too bad of a deal, although I only got 2 taps from it.
- It takes some time (3 days) for the campaign to ramp up.
- Average CPA stays at $0.00 throughout the campaign simply because iAd acquires 0 users for my app. Does it mean I should not get charged at all?
- I chose CPA but got 0 downloads, so I do not understand why I got charged my Daily Spend Goal of around $12.50 when my cost of acquisitions is infinity (divided by 0!). I am going to contact Apple to ask about this. I will follow up here in my blog when I get a response.
- Average CPC is $0.04. I thought it would be $0.00 because I chose the CPA option. Why is my CPC non-zero when I choose to use CPA for my campaign?
Things missing from iAd Workbench that I wish it has:
- There is no data as to which apps my iAd was shown. Also, I don’t know the age group of my impressions. For instance, I wouldn’t want to show my iAd to kids who use their parents’ iPads to play puzzle games. If Apple can provide more insights into this data, it would be tremendously helpful to find tune campaigns to achieve better results.
- Badminton is not a popular sports in the US. iAd Workbench initially reaches only the 50 states in the US. If iAd Workbench was able to target potential users outside the US, I imagine it would help me to get more taps and maybe downloads. From 6/11 to 6/18, there were 45 downloads in the US but 269 downloads outside the US.
- When you tap an iAd, it shows your app in a special page within the hosting app. The users can tap the Download button but they are asked for the App Store password. They then need to find their passwords from 1Password or whatever. That is too much of an obstacle. They may also think they can’t see the iAd anymore, so they just search for it in the App Store and download it there. If Apple can simplify this process, it will be easier for users to download apps from seeing an iAd.
- Finally, I am sure everyone would love it if Apple tells us how users find our apps. Is it from search in the App Store? What are our top referral links? What is the demographics of searches? What keywords do they search for?
I contacted the Apple iAd team with the following questions and hope to hear from them soon. I will post an update when I do.
- I chose CPA and set it to $0.83. Why is my Avg. CPC $0.40?
- The campaign generated 0 download, so my Avg. CPA is $0.00. However, my Spend is $74.66. How can this be?
- Why does it take 3 days for the Daily Spend to match my Daily Spend Goal of $12.50?
- I am very disappointed at the performance of my iAd Workbench campaign. Am I doing something wrong? Is there anything Apple or I can do to improve the performance?
- Can you provide more insights into the data such as which apps show my iAd, the demographics who saw and tapped my iAd?
I received a reply from someone in the iAd Sales team at Apple today, but the email was very brief:
The CPA that you entered is a CPA goal. We will optimize towards this goal, but you will pay based on CPC, determined by the auction. You might consider increasing your campaign’s CPC or CPA bid to be more competitive in the marketplace. For useful resources on optimizing your iAd campaign, please refer to iAd Workbench Help > Tips for Creating Successful Campaigns: http://help.apple.com/iadworkbench/
Please let me know if you have any other questions.
So the CPA Goal is simply a ‘goal’ – a goal that you pay after a download is acquired. For those of you that have non-zero downloads, would you mind share your numbers? We are still charged for CPC (in my case the CPC of $0.04 is calculated for me. CPC ($0.04) x Taps (1,885) = $75.40 ~ $74.66) regardless of whether we choose CPC or CPA. It feels like whatever we set as our Daily Spend Goal, Apple will charge that amount, even if we get 0 downloads. I’ve read the help multiple times, but I read it again and it says:
For example, if your goal is to achieve the highest number of downloads for the lowest budget, you can set cost-per-acquisition (CPA) goals to make sure your ad dollars are generating new customers. On the other hand, if your goal is simply to increase awareness of your app in the App Store, you can use cost-per-click (CPC) bidding to ensure the money you spend reaches an extensive audience and brings a high volume of users to view your app.
It is kind of misleading because it gives the impression that you will be charged your CPA Goal for each download, but that isn’t the case. The rest of that page basically says test, test, test your campaign to maximize its effectiveness. However, it is hard for us indies to test, test, test, when the initial effectiveness is 0.
The reply I got did not answer my other questions in my original email. I hope Apple will give us indies more help and data to get more exposure and downloads for our apps. If not, I think iAd Workbench is just a different medium for big companies to spend their ad dollars.