News for hackers. Stuff that matters.

Ask HN: Recommendations for mobile analytics product at scale?

We're now at scale and looking for a mobile analytics product that allows us to take a look at our funnels and user retention.

We have the following requirements: -We need to track about 200MM-400MM events/month. -Realtime retroactive funnels. I want to be able to make a query and get an answer back in less than 60 seconds. I also want to be able to add arbitrary key value pairs to each event and filter steps in the funnel based on the presence of a particular value. -We need to be able to apply filters on users to the funnel and retention graphs that we look at. We'd like to filter on standard things, such as country or device, as well as define our own user properties (eg by ad campaign or A/B test group).

Up to this point, we use Flurry and Apsalar, although we've found both to be inaccurate compared to numbers that we track internally. They're also annoying tools to use. We're happy to pay a few thousand dollars a month for this. Mixpanel (who offers the closest to what we want)/Kontagent cost an order of magnitude too much. I've also looked at Localytics, but after talking to a few current users of the platform, it seems rather simplistic and they don't have any advanced segmentation on the funnels.

Do you guys know of anything that fits the bill? We'd rather not have to think about building it in house.

( 18 comments | 23 points | on HN )
Re:Ask HN: Recommendations for mobile analytics product at s... by liamgooding on Thursday July 18, 2013 @04:57AM (#6062537)
Hey,

at http://Trak.io we can definitely handle that data volume, all data analysis is done retroactively and in realtime.

Note: However we don't automatically track any events, as we found most people were collecting a ton of garbage event data that they would never use. So you'll need to setup the events to be tracked at the beginning of the integration.

We don't have iOS or Android SDK's, but a raw REST http API (and JS) but we'd be open to discussing what we could sort out to win your business ;)

A/B tests: date vs. date cohort segmentation, event vs. event cohort segmentation, property vs. property segmentation. And groupby (breakdown by) view.

Trends

Linear funnels are in progress, but again for an app this size let's talk about what we could do.

The main way we differentiate ourselves from Mixpanel (and similar) is that we're trying to go for as simple interface as possible. Essentially, if analytics tools or reports are complicated, they just wont get used. So we're sticking to a very design-led product which may or may not suit you, depending on what exactly you're looking for.

Email me on [email protected] if you want to chat further :)

Re:Ask HN: Recommendations for mobile analytics product at s... by raviparikh on Thursday July 18, 2013 @02:17AM (#6062061)
We (Heap) recently launched our iOS library. In addition to retroactive funnels and arbitrary key/value pairs, we let you retroactively define events without having to ship new code to the app store.

For instance, you can define an event as a "Touch on a UIButtonView". Then, we dig into your users' entire past activity to let you instantly segment on that event or include it in funnels. No new code is required.

We haven't publicly launched our Android integration yet, but we should be able to figure something out - ping us at [email protected]. We're flexible on pricing and would love to help.

Re:Ask HN: Recommendations for mobile analytics product at s... by smilliken on Thursday July 18, 2013 @01:56AM (#6062006)
A few thousand dollars a month is a modest budget if you want to outsource this. Try negotiating price if there's a vendor that does everything you need.

If you only need to analyze <1B events at a time, then an old-fashioned relational database is fine. Beyond that, CitusDB, Redshift, and Vertica are analytics databases that might be in your price range.

Bigger mobile publishers will usually consider analytics a core competency and handle it in-house; if you continue to grow you'll quickly reach a point where it's no longer possible to outsource this.

Re:Ask HN: Recommendations for mobile analytics product at s... by sdossick on Thursday July 18, 2013 @02:32AM (#6062120)
My company ( PushSpring ) has a product that lets you track arbitrary data and events about your mobile app users and then segment them. We also offer the ability to send targeted messages to segments you define via Push Notifications, as a way of keeping them engaged (or re-engaging groups who've left). We're working with some of the largest publishers in the app store.

Love to chat about your requirements -- [email protected] if you'd like to discuss.

Re:Ask HN: Recommendations for mobile analytics product at s... by gorkemcetin on Thursday July 18, 2013 @12:35AM (#6061790)
Countly is an open source, realtime mobile analytics solution with an enterprise plan. Funnels and segmentation are currently being developed and planned to get ready for mainstream due 1-2 months (advanced segmentations being quite earlier). Check http://count.ly for a list of features and benefits.
Re:Ask HN: Recommendations for mobile analytics product at s... by appdeveloper on Thursday July 18, 2013 @12:55AM (#6061847)
Looks interesting, thanks! Questions: for funnels- are you guys planning to make it so I can query on historical data and get the results right away? Flurry is pretty annoying- you have to set up the funnel, and you only get results from that point forward, it doesn't apply to data you've collected in the past.
Re:Ask HN: Recommendations for mobile analytics product at s... by zmitri on Thursday July 18, 2013 @12:22AM (#6061765)
Would using Google Analytics work? I know it's not as easy as Mixpanel but they have a lot of tools. They also have an Objective C SDK.

There's a YC company called Heap Analytics that offers pricing based on monthly unique users. I've never used it but it's the only other thing I can think of.

Re:Ask HN: Recommendations for mobile analytics product at s... by mattdeboard on Thursday July 18, 2013 @09:27PM (#6067778)
What does "we're now at scale" mean? I realize this sounds snarky but I am asking sincerely. Maybe it would help if someone could finish the following sentence in such a way that it's equivalent to "we're now at scale": "We have so many customers that ..."
Re:Ask HN: Recommendations for mobile analytics product at s... by kposehn on Thursday July 18, 2013 @12:23AM (#6061766)
We have a platform that is similar to what you need, but geared to the web. The advantage is adapting to your use case would not be a huge issue and I've been intending to go that direction anyway.

Hit me up if you would like, my email is in my profile.

Re:Ask HN: Recommendations for mobile analytics product at s... by appdeveloper on Thursday July 18, 2013 @12:53AM (#6061838)
Interesting- do you guys have a mobile SDK already? We'd need both iOS and Android.
Re:Ask HN: Recommendations for mobile analytics product at s... by kposehn on Thursday July 18, 2013 @01:15AM (#6061893)
Replied to you with details :)
Re:Ask HN: Recommendations for mobile analytics product at s... by appdeveloper on Thursday July 18, 2013 @02:10AM (#6062045)
thanks!
Re:Ask HN: Recommendations for mobile analytics product at s... by LedZeppelin on Thursday July 18, 2013 @12:22AM (#6061764)
What information are you looking for?

Geographic location, age, gender, education?

Re:Ask HN: Recommendations for mobile analytics product at s... by appdeveloper on Thursday July 18, 2013 @12:57AM (#6061849)
Ideally the SDK would auto-pull from the phone anything it could, such as device type, OS, language/locale settings. You wouldn't be able to know actual demographic information on the person, of course [ :) or at least I don't know of any APIs that exist for that ].
Re:Ask HN: Recommendations for mobile analytics product at s... by endlessvoid94 on Thursday July 18, 2013 @01:44AM (#6061974)
You get what you pay for.
Re:Ask HN: Recommendations for mobile analytics product at s... by marcooda on Thursday July 18, 2013 @01:51AM (#6061988)
appfireworks.com is launching soon - apply for beta.
Re:Ask HN: Recommendations for mobile analytics product at s... by appdeveloper on Thursday July 18, 2013 @02:09AM (#6062042)
It looks like you guys are more on the attribution analytics side similar to Mobile App Tracking. Do you guys do funnels/retention/segmentation analysis like Mixpanel/Flurry?
Re:Ask HN: Recommendations for mobile analytics product at s... by marcooda on Tuesday July 23, 2013 @09:50PM (#6094060)
Focus is on installs, user LTV (so yes segmentation, retention, etc) and crash analytics.