You’ve tried Google Analytics (GA) and now you’re looking for a tool which provides you with other ways to provide more actionable insights. Mixpanel and Kissmetrics are the tools you’ll likely come across. Both have a free version based on the number of events you’ll likely use, so they are ideal to test out, but which one is best suited for you?
2 of the main reasons that you’d want to try Mixpanel or Kissmetrics are because you want to perform Cohort Analysis and/or you want to track individual behaviours. There are a couple of articles out there that already cover some of the differences. This one from March 2011, and this more recent one from November 2012. In brief the first article rates the tools based on -
- Live Data
- People Tracking
Personally, although useful to some extent, they don’t really explain to me how to get the types of answers that I’m asking. For example, If I want to know how long on average it takes for people to hit a certain event, and I want to compile a cohort analysis to see if the actions I’m making are improving this goal, which tool is better?
On that basis, I’ll use Dave McClure’s Startup Metrics for Pirates AARRR and see how Mixpanel and Kissmetrics can be used to extract the metrics you want. These are all personal opinions and you may have a different preference based off how you apply your metrics.
Both Mixpanel and Kissmetrics allow you to plug in campaign data into your pages.
For Kissmetrics, when looking at funnel reports, you can select the campaign properties as check the effectiveness of each campaign source to see which channel/websites are the most effective at bringing in traffic that is most efficient toward achieving your goals. This means that if you were to put marketing dollars into a particular channel/website to acquire users, you should know where to place your bets.
For Mixpanel its pretty much the same, set up a person property called campaign or use utm configurations and when you pull up your funnel reports you get the same display as in Kissmetrics.
Both tools are pretty much equal here, using funnels, you select which events you expect people to perform to go down your activation funnel and you get your conversion and see at which events do people drop out.The main differences between the two is what other data is available in from the funnel report.
In Kissmetrics, you can hover over an event and “view people in this step”. This provides you with a list of people who reached this point in the funnel, allowing you to individually click in and see what events and properties you’ve recorded for them
For Mixpanel, you are provided with trend charts, you can see what percentage of people passed this part of the funnel in a day to day map (or week or month if you so choose).
Which one is more useful? For me, Kissmetrics provides an easier way to see who has performed this part of the funnel which lets you dig down into the behaviors of individual users.
Retention graphs by both tools are similar but have enough differences for you to be able to pick a preference.
Kissmetrics power is the ability to create retention tables beyond grouping by date. You can select advanced options, and group people by their properties or events they have performed, this can provide some powerful insights, such as figuring out which campaigns provide higher retention. Kissmetrics also let you see the absolute figure when you hover over the retention chart, this can make processing the data easier to digest.
Mixpanel’s power is the ability to segment. After selecting an event, you can choose which properties that an individual can possess and the retention chart is filtered this way. In some ways similar to Kissmetric’s advanced options, but a bit fiddlier because you would have to create a report for each segment to be able to compare.
For me, in terms of the retention tables, Kissmetrics is better than Mixpanel, But Mixpanel also provides a retention tool via their engagement mailing tools, this is useful creating retention actions. Intercom.io is a better solution for this at the time of writing.
Configuration will allow you to track referral. One way would be to set up a persons property to have a referral variable and if they are referred by another user, attach their id to this property.
Kissmetrics provide a dashboard where you can compile different metrics to see how things are going. You can use this to determine how ‘viral’ your app is. A way to do this is to create a referral event, so when someone signs up to your app through a referral, the event is created and attached to the referring user. You can then track how many times a single user has successfully referred someone and use the “Average Number of Times Event Happened Per Person” metric
You can use the same technique for mixpanel. To display the data, use formulas where you can select the number of times the event has been and divide it by the number of people who visited the site and see how viral your app is.
Both metrics are slightly different, but as long as you understand what you are measuring and create actions to drive these numbers there shouldn’t be a problem.
Both tools have dedicated revenue functions.
Mixpanel has this embedded in their people section, you can segment revenue by sources, and other properties and figure out where the revenue come from.
Kissmetrics uses events which you can map to their standardized revenue events, after which you can then see your revenue reports y referrer, lifetime values and other customer data such as churn rates.
One useful function that Mixpanel provides that Kissmetrics doesn’t is the ability to increment properties. This means you can keep track of variable for users such as credits, if someone spends or buys credits you can increment/decrement that variable. This can be really useful, but unfortunately, where it would provide most value is if you could use these properties easily in a cohort table. Thus this feature is most useful when combined with Kissmetrics cohort table which can be grouped by property. Unfortunately, It’s not possible to combine the tools .
It’s important to remember that both tools are still iterating and developing so this is merely a current state of the play.
Photo Credit: Double Faced Matter Horn by Gabriel Zech