Hacker News new | past | comments | ask | show | jobs | submit login
Mixpanel Raises $10.25M (allthingsd.com)
111 points by mynameisraj on May 10, 2012 | hide | past | favorite | 23 comments



Been thinking about switching to Mixpanel, but can't see any massive improvements over the new Google Analytics, and it looks a lot more tedious to set up.

Any opinions here?


I've been using both. The primary use case has been from a mobile app where various events are tracked. There are also some extra pieces of data attached to each event (for example the audio volume) called custom variables in Google and properties in Mixpanel.

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.


It's exactly the same amount of effort to setup. I use both GA (for pageview tracking) and Mixpanel (for engagement tracking). Mixpanel's funnel tracker is more robust, and there is no GA equivalent to Mixpanel's retention analysis. Plus, Mixpanel's event tracking is truly real time.


Doshi said he wanted to let it be known that he’ll use the funding to pay his first 100 employees “way above market”

Nice :)

Seems like a good company with a valuable product.


100 employees @ 100k = 10 mil. He will last a year?


I'm assuming his company actually makes revenues since they charge, so most of the salaries may be covered by revenues already and the funding will make up the rest.


    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
I guess they have discounted volume plans, otherwise it's looking too good to be true ;)

UPDATE: using their 20M datapoints for $1600 plan, 7B datapoints per month would bring "only" $560K revenue per month.


Congrats! I love Mixpanel for the results, and also the team is great. I switched from Kissmetrics to Mixpanel and found the Mixpanel team to be solid, smart, and very responsive.


I've known Suhail, online, for quite awhile. Good to see this news. Congrats. I remember when he wanted to learn Django, now look where he's at. :P


I really shouldn't be building this myself. But looking at these prices for mixpanel and kissmetrics, they seem out of whack-- great for profitable businesses and enterprise, but way out of line for startups. Plus what if you need between 25,000 and 500,000 data points per month with mixpanel? That's a huge gulf and you go from free to $150 a 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.


Have you seen what omniture sitecatalyst costs? (may be they've gotten cheaper last I heard, but considering what Adobe paid for them, I don't think.)

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.


Omniture and WebTrends are also steaming piles of outdated features and buggy UIs. MixPanel seems like 100% New Hotness (pardon the cliche).


We absolutely support you: https://mixpanel.com/free/ - instead of a $30-80/mo plan, it's free! Of course, if you're averse to placing a badge on your site, then that's a different problem.


The badge option is very generous IMO. Keep up the awesome work!


I'm glad that you offer the badge option, but I wish that was in addition to a $30-80/mo plan that I could pay for with cash...


Shouldn't that be no-followed?


To defend my post, mixpanel is providing copy/paste html with a link for people to put into their footer in exchange for services of monetary value.

Could someone explain why this does not qualify as an ad?


On Google's published guidelines and guidance, this is white hat, as the links are editorially chosen (you, the webmaster, really are specifically intending to endorse them for their analytics services because you use their analytics services).

The shadow rule is that any linkbuilding method is kosher iff no one has figured out a way to scalably achieve rankings on very competitive terms with it. For example, citations on the footers of blog templates were 150% A-OK right until people found out that was a scalable way to rank e.g. for e.g. [printer cartridges], [payday loans], etc etc. See also widgetbait, quizzes, etc etc.


great for profitable businesses and enterprise, but way out of line for startups

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 also implement sampling in your client/site since you are unlikely to need every detail from every session. For example in my use cases there is a percentage probability that the session events will be dispatched. That starts out at 100% on initial deployment and then goes down over time as the user base builds up. We also mark some serious events as always dispatched.

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).


Im all about frugality (when warranted) in a startup situation, but $150 a month? That's a cell phone family plan bill, man.


You're crazy. You should value your own time at over $1000/hour if you're starting a company. $150/mo is only worth it if you can bang this thing out in minutes. So you're going to waste hours that you SHOULD be spending on your product or marketing or sales on some BS that could be NOT YOUR PROBLEM for $150.


Not everyone has thousands or millions of dollars in funding.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: