How to do mobile marketing: target, measure, analyze

Recently, I heard from a colleague that Facebook Mobile Install Ads was performing terribly for them. This was a few weeks after I ran a test campaign on Facebook for Chinchin and found great results. I took a look at how they were setting up their campaigns and noticed a few areas for improvement.

There are articles everywhere discussing how we should be measuring customer acquisition cost (CAC) and lifetime value of users (LTV) when doing online marketing but there don’t seem to be many resources that truly get into the nitty gritty.


Many seem to have the notion that performance based mobile marketing either works for you or it doesn’t but this isn’t necessarily true. If you have any experience with keywords ads on Google Adwords, you will know that some keywords work and some don’t. It’s not necessarily a matter of whether or not Google Adwords works for you or not but which keywords.

When I advertise on mobile ad networks, I make sure to segment my targets as much as possible to determine which targets work and which don’t. For Facebook specifically, I split up my campaigns based on OS, gender and age range. With these basic targeting options, I had over a dozen segments such as:

  1. Males between 20-30 using iOS
  2. Females between 20-30 using iOS

After you add additional targets such as interests or pages people like, you can really get into the long tail. I thought back to my colleague’s statement about Facebook and I looked at results from my own campaigns. I realized how targeting incorrectly or not targeting at all could make a good ad network look like a bad one. More on this when I discuss analyzation later in this post.


A couple of years ago, I was in charge of marketing at a startup and mobile marketing was a bit of a mystery. With pay per click (PPC) ads, you could measure everything but with mobile marketing, the app stores were black holes. You would send traffic to the app stores but from there, you would lose track of what’s going on. This is why boost marketing (trying to get your app into the top 10 or 25) was so popular at first.

The technology has improved rapidly since then. At Chinchin, we integrated HasOffer’s Mobile App Tracking (MAT) service. MAT can track attributions through the app stores but it is also beneficial for another reason. Each ad network has its own SDK you have to integrate but if you are advertising on 10 ad networks, that means 10 SDK’s to integrate. With MAT, you integrate their SDK once and you can advertise on the 10 ad networks since MAT is already integrated with them.

So once you integrate with a service such as MAT, you can track attributions. You can track CAC by ad network or even specific ad campaigns. MAT also allows you to track other events such as registration, in app purchases, etc. From there, you are able to track LTV. We were already using Mixpanel for tracking in app events but luckily, MAT and Mixpanel easily integrated with each other. No matter what solution you choose, tracking attributions as well as CAC and LTV is not a problem on mobile anymore.


After you have the SDK integrated, set up your targeted ad campaigns and run marketing, it’s time to analyze the results. Going back to Facebook Mobile Install Ads, I looked at my own results and noticed something striking. CAC for Chinchin via Facebook ads ranged from $0.27 to $1.90 depending on the target.

If I had a limit of $1.00 for CAC and had not done any targeting, my CAC could have easily come out to over $1.00. This would make Facebook a bad ad network for me. With the targets I set, I was able to find out that 2/3 of my targets had a CAC of under $1.00. Facebook went from a marketing channel that might be cut entirely to a highly effective one.

So there you have it. Target, measure and analyze. What we’ve been used to when doing marketing on the web is now available for mobile.


Now read this

Making a mobile app? Iterate with mobile web first

There is a lot of talk about the Lean Startup in which you create a minimal viable product, get it in front of customers and iterate rapidly. With the Web, this process was much easier to implement as you could make one product and it... Continue →