Any opinions here?
Google works really well with its out of the box reports. The UI once you are trying to do more complex things gets somewhat bizarre, but you can generally find out whatever you wanted. If you have a numeric custom variable then they only show the average which is spectacularly useless - the distribution is far more important. Their Android library only allows one tracker instance so you can't have two different components tracking events against different ids. There is no support, unless you fork out $150k per year (that is the lowest price above the free plan). There is a way of doing data queries but I don't believe you can get export the raw underlying data that was sent to them.
Mixpanel is very strange. My (perhaps wrong) conclusion is that it was designed by an architecture astronaut. I mean data is just data so why not abstract data into data. There are no precanned reports, and the reporting tool requires conditions to be set when there is more than one item of data being looked at (which is all the time because you are trying to correlate and rank items with each other). It is frustrating to try and find out anything. You do get to see more granularity on numeric values, but not as much as I'd like. They do have support but all we got back was a bunch of platitudes. Fortunately they do have an export mechanism so you can at least download all of your raw data and write your own analysis code.
Nice :)
Seems like a good company with a valuable product.
octave:1> price_per_datapoint = 150 / 500000
price_per_datapoint = 3.0000e-04
octave:2> datapoints_per_months = 7e9
datapoints_per_months = 7.0000e+09
octave:3> revenue_per_month = datapoints_per_months * price_per_datapoint
revenue_per_month = 2100000
UPDATE: using their 20M datapoints for $1600 plan, 7B datapoints per month would bring "only" $560K revenue per month.
Some startups start with large customer bases but operate at a relatively low margin per customer (because its extremely cheap to serve each customer.) For instance, for $250 a month in hosting[1], I can service tens of millions of customers, but buying either of these services for that many customers would cost a lot more than the hosting.
Looking at real world metrics for one of my apps I'm currently collecting enough data points (for free using Flurry) that to use mix panel would cost a sizable (%15-%20) of the apps monthly revenue.
I'm not saying they're not worth it-- I'm sure they are. I just don't see a way for a startup to onramp here. Our primary burn is keeping us housed and fed, operations are being kept as cheap as possible until we have real revenue. $150 is a lot. (though kiss metrics has a $30 and $80 plans which alone means we're more likely to use them.)
[1] in part because hetzner's XE 4S servers are such a screaming deal- that's about 5 32GB servers.
Mixpanel, and a whole host of other services, are doing startups a big favor. Complaining that $150 is "out of wack" is pretty close to trolling.
I'm running Mixpanel on a handful of my sites and its great. Any time I can outsource an important task or tool for dirt cheap, so my developers can concentrate on our core product, I feel like I'm getting it for free.
Good riddance to companies that don't post their prices and give you an 800 number to talk to a sales rep.
And that is how it should be. To build a longterm business, you really cannot rely on catering to startups for whom $150/mo is a huge investment. I lead mixpanel adoption at startup I work for. We're ok funded and the mixpanel cost does not even make a blip on our expenses.
And yet, like you said, when I go home to test out my own ideas outside of work, mixpanel does seem like a cost I have to think hard about before paying for.
You can amuse yourself by looking at Google's pricing. They have a choice of free or $150k per year!
At my startup we would gladly pay Mixpanel's prices. Unfortunately we found the functionality to be poor, and support to be rather bizarre (every response was a bunch of platitudes).