Conversion rate measures, what percentage of your users complete significant actions that you want them to take, but an important limitation of conversion rate is that sometimes your most important conversions are actually pretty rare events. Imagine that you have a Web application which allows users to share photos and there’s both a free account and a paid version of the service. The act of upgrading to the paid version might be your most important conversion in terms of generating revenue. But if that action happens only once for each user, there won’t be many data points to analyze, which makes it more difficult to detect changes in that metric over time.
So a good approach is to identify both significant conversions such as upgrading, which can be described as macro conversions and less momentous but more frequent actions which can be described as micro conversions. For example, if you know that people who upload a lot of photos are more likely to upgrade, the act of uploading a photo would be a good micro conversion to track, even though it doesn’t generate revenue directly if the act of uploading happens more often. There are a lot more data points and that will make it easier to detect trends and whether the action is becoming more or less frequent as you implement changes to the design. So your best bet is to identify and track both macro conversions and several different micro conversions.
This approach gives you immediate feedback about specific design changes and will also enable you to predict longer term results without waiting for infrequent data to accumulate.