I read a lot of indie developer blogs and I’ve come to realize that while most programmers are very smart, unfortunately they aren’t very good at market segmentation. Good market segmentation typically follows a few rules:
- Identified segments are internally homogenous
- Segments are identifiable based on a meaningful and measurable factor
- Segments are uniquely accessible
- Segments have effective demand, meaning the group is large enough and has the necessary income to purchase your product
A good market segment is NOT women between the ages of 35-45 because individuals within this segment may have wide disparity in income level (not homogenous) and no access to mobile technology (not accessible). A better way to segment users is by behavior because the purpose of segmentation is to identify the customer qualities that are predictive of probable purchase and use of your product.
For example – you might be interested in isolating differences among individuals who derive more value from your product. Within mobile, traffic source is a major predictive variable. Therefore by using conversion-to-paid as a proxy for value, many successful mobile-focused companies will consider the following segmentation groups:
- Organic traffic – users that search and find your app on their own. These guys generally provide the highest value.
- Social traffic – users that get referred to your app through friends. These guys generally convert at a rate of 60% compared to organic users.
- Burst campaign traffic – new users referred through a paid acquisition channel, such as incentivized download or banner ads. These generally convert at around 30% of organic users.
- Featured traffic – users acquired from recommendation services, such as App-o-day and review blogs. The lowest quality of them all. Featured traffic converts <10% compared to organic traffic.