Hm... we're happy users of both and they have almost nothing in common. Mixpanel is more like a flexible general purpose database which gives you real time "views" of your data. You can chart it yourself or Mixpanel will do it for you.
We (http://mailgun.net) built our email analytics on top if it in a very short period of time and it's much nicer that anything we could have made in-house (and keeps getting nicer).
The announcement they make in a few weeks is going to be huge and disruptive. Some interesting bits from the interview in Startups Open Sourced:
Q: Tell me what your experience was like raising money. Was that positive or negative?
That experience sucked, and I'm still in that process. It's a lot better now but it still sucks. It's still very taxing.
When we went to raise before, we were six months after the recession which meant that it was very hard to raise from anybody because they were being super cautious. Sequoia did this "RIP" slide thing. It was challenging, and so it sucked because you just got a lot of "no's." We were two kids from ASU, never done a company, had never worked anywhere really interesting. We had nothing going for us other than that we were in YC. So raising sucked; we almost died actually. We were probably a week from just dying, thinking "Oh guess we're going to go back to school." I was like, "Tim, if you ever want to build anything, build it now. We might not be alive in the next month, so just build it now." And we actually ended up raising from Max Levchin and Michael Birch at the last second, and that saved us. We were like, "Oh, we finally did it, we raised." But raising was just taxing because you end up talking to VC's who have a million reasons why your idea sucks, won't work, and then you talk to angels who won't invest in you until some other angel invests in you. You have this chicken and egg problem.
Q: Sequoia's leaked RIP slides happened October 2008, and we were in the same YC class which was Summer 2009, so the summer right after. People may forget but it was actually kind of a challenging economical environment.
Yeah. I think the companies that raised--you had to be pretty decent if you raised. I think we were more lucky than decent.
Q: Any points where you thought Mixpanel would fail or where you really had stressful moments?
Yeah. It happened two weeks ago. It happens a lot. Basically when you're doing real-time data analysis, the problem is that you have to pre-compute everything. You have to do it all right when the data comes which means that if you don't scale the rate at which the data is coming in you start to get backlogged. You just have data that's piling up over time and so we definitely had scenarios where we were not real time, we had 24 hour, 48 hour, 72 hour delays for two weeks, and there are definitely points in times where the scaling problems were so bad that I would just lay on the ground and think "How am I going to scale this? How am I going to do this?" And the thing is, if you can't, you die.
Q: I remember the delays in Mixpanel's early days.
Yeah, things like that still happen even today, but they're less severe. We have a better handle on it obviously, but there were definitely days where things just got really, really bad, and if you can't scale you just lose.
I wish Startups Open Sourced had some way to do interactive content; I'd love to be able to reformat it and see just one section (e.g. "how hard is raising money") for EVERY company which answered that.
I've read about half of the dead-tree version at Sunfire while waiting for VM stuff; it's a great book, but this is exactly the kind of data where being able to treat it as structured data, instead of a static book, would be awesome.
You're not the first person to recommend this, I'm actually interested in doing it that way too. I'll get this done at some point when I move it into an HTML format that can be re-arranged and compiled on the fly.
The big differences:
a) More real-time. Nothing is 100% realtime quite yet but MixPanel is by far the best.
b) Event driven - GA is pageviews driven, KissMetrics (also a fantastic product) is funnel driven. If you care about tracking key events and tagging those events with metadata that you can slice-and-dice later, MixPanel is the best solution.
c) API - really great and simple API to get data in and get data out. This makes it easy to integrate MixPanel across multiple touchpoints (mobile app + web) for example, and have control over how you make those big piles of data interact/merge/display.
Google Analytics has a delay. Mixpanel is real-time and specifically aimed at certain events that you specify. So when someone clicks your sign up button, it fires an AJAX call and you see them do it in real-time, instantly.
I like what MixPanel is doing, but I don't think it's fair to say that GA has a 24 hour delay. It's nothing close to real-time, but you can get results from the past few hours.
I'm looking at Google Analytics right now for my web site, but all I can see is yesterday's data. Is there somewhere to see the data collected today?
Edit: I use both Analytics and Mixpanel, they're both great services. Google has its strengths, but I don't think event tracking and immediate feedback is one of them, which is where Mixpanel comes in.
Click the date select stuff, you can choose today and it'll give you whatever data it's processed for today, but by default it only shows up to yesterday.
We use it in our product. It's incredibly easy to set up, and the entire focus is on tracking events happening in your code.
I.e., we have an event for people pressing the login button, another event for people actually logging in, an event for posting a message, etc. Very simple to set up, and now gives us real-time analytics (and funnels) on how many people press login vs. actually login, etc.
And those are just small examples - we have a huge range of events that we track, things we care about.
Congrats on the funding and really happy to see their momentum keep going.