Aug
2010
Get more insight into your conversion funnel
This post is part of our series on progressive sales and marketing tactics. Look for more posts like this one every Wednesday this month.
Much has been written about the conversion funnel, and this post is not going to rehash the basics. Pretty much every analytics tool has some representation of the funnel, and most marketers we work with tend to know their funnel pretty well. Even though it is the most common report that marketers use, there are some basic problems with it that cause problems for even the most seasoned CMO’s.
The funnel is a great tool to evaluate the success of a given campaign. Look at the funnel for the example email campaign to the right. Since an email campaign is a discrete event with a short lifespan, it is very effective to use a funnel to measure its success. In this case, 250 out of 10,000 targets actually filled out the lead form on the landing page…pretty straightforward.
The funnel visualization starts to get complicated when you have a long conversion cycle between steps in the funnel. It is not unusual for a B2B business to have a 180-day cycle time from the time that a lead is generated until the deal is closed. Most companies use the standard Salesforce.com Leads -> Opportunities -> Customer funnel to measure their effectiveness across this cycle. The problem with this funnel is that it becomes very difficult to understand the relative performance of your conversion rates over time. You might not notice that you are closing less opportunities until it is too late.
One way to get a pulse on your conversion performance is by using cohort analysis. The Wikipedia definition of a cohort is a group of people who share a common characteristic or experience within a defined period (e.g., are born, are exposed to a drug or a vaccine, etc.). The most common cohort used by marketers is a campaign. Using campaigns can be misleading because if you are getting different performance over time, you may attribute it to the specific campaign you are running rather than other factors that are contributing. For example, say your company hired a rockstar sales engineer who developed a killer demo. He brings in more sales, which in turn improves your overall funnel. If you only look at the campaigns, you may get a false positive that your campaigns are working even though the sales engineer is the reason that those deals went through. The same can be said for website updates, messaging changes, product improvements, etc.
The easiest way to get started with cohort analysis is to use date based cohorts. In the example to the right, we are measuring a simple funnel but we also list key events that happened that month. By doing so, we can look at previous months to see if there is a correlation with any of those changes to a change in performance. In this case, you might realize that the new landing page designs set better expectations for the customer, which in turn helped the close rate. If you were simply measuring campaigns, you wouldn’t have easily made that correlation.
Let’s look at the same graphic for the next month. In this month, our overall leads to customers rate was the same, but we had less qualified opportunities. The key events are that we had messaging changes that may have contributed to more leads, but we had our best sales rep resign. From this analysis, we can hypothesize why we had less opportunities. Was it the sales rep? Was the messaging sending bad leads? We are not 100% sure, but it is something we can easily test.
In conclusion, traditional funnel reports only tell you so much, especially when coupled with long sales cycles. If you want to dig a little deeper into your funnel, it is helpful to use techniques such as cohort analysis to discover hypotheses on the movement of your leads through the funnel. These hypotheses can be easily tested in future cycles and will help you to optimize your funnel over time.







