Live data from Hacker News

Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

blog.fleex.tv

1–10 of 27 posts

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#2
The last paragraph is important. I spent some time earlier this week when I learned about Universal Analytics -- but quickly discovered that UserID tracking hasn't shipped yet.

Can anyone on the GA team speculate about a release date for the uid bits?

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#3

The last paragraph is important. I spent some time earlier this week when I learned about Universal Analytics -- but quickly discovered that UserID tracking hasn't shipped yet. Can anyone on the GA team speculate about a release date for the uid bits?

userId bit has been there from the start in Universal Analytics. Its called custom dimensions and can be used to send any property about the user into the GA and then link it to a User or a specific Visit.

https://developers.google.com/analytics/devguides/collection...

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#4
post #3

The last paragraph is important. I spent some time earlier this week when I learned about Universal Analytics -- but quickly discovered that UserID tracking hasn't shipped yet. Can anyone on the GA team speculate about a release date for the uid bits?

userId bit has been there from the start in Universal Analytics. Its called custom dimensions and can be used to send any property about the user into the GA and then link it to a User or a specific Visit. https://developers.google.com/analytics/devguides/collection...

OP (and the article) refer to the uid tracking mentioned here: https://groups.google.com/forum/m/#!msg/google-analytics-mea...

As of Jun 27 it hasn't shipped according to a GA team member.

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#5
> For one, there can’t be 2 [clientID, userID] couples with the same userID: with the way mixpanel does things, this is essentially a technically impossible scenario (...) And yet one user can access your site through different clients, leading to a systematic overestimation of the number of visitors hitting your site.

Really? Anyone can confirm this behavior? I'm pretty sure KissMetrics doesn't have this limitation.

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#6
My gripe about google universal analytics or analytics.js vs ga.js is

broken backwards compatibility (cookie data is no longer stored in the same way) this was an interface many add/systems used and depend on from the days of Urchin.

Otherwise, new interface is pretty slick, features look good, the API to send data server side is so much nicer.

broken compatibility just kinda sucks though

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#7
This article is really making a big deal out of nothing. All the "major issues" brought up here only create problems in edge cases. When you're trying to drive growth or understand your users (the purpose of metrics at the end of the day) you should not be focused on edge cases.

In most cases the reason you care about tracking logged-out -> logged-in behavior is to measure onboarding behavior, understanding what the user does pre-signup so you can do a better job of driving signups. Signup is not a multi-client process in the common case so being able to track multi-client behavior pre-signup doesn't really matter at all.

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#8
I was not aware that the new analytics would track users. One interpretation of section 7 of the Google Analytics Terms of Service is that tracking individuals is not allowed:

http://www.google.com/analytics/terms/us.html

  > You will not [...] use  the Service to track, collect or 
  > upload any data that personally identifies an individual 
http://productforums.google.com/forum/#!topic/analytics/tTaq...

  > you cannot store names or ip addresses in a custom var, 
  > but you can store ids that need your backend to resolve 
  > into a person identification

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#9
post #5

> For one, there can’t be 2 [clientID, userID] couples with the same userID: with the way mixpanel does things, this is essentially a technically impossible scenario (...) And yet one user can access your site through different clients, leading to a systematic overestimation of the number of visitors hitting your site. Really? Anyone can confirm this behavior? I'm pretty sure KissMetrics doesn't have this limitation.

Indeed, and this is why we ended up choosing KM over MP. With KM you just "identify" a visitor whenever you want and if there's already another anonymous cookie, it'll tie together all events retroactively. We couldn't find an easy way to do this with MP when we looked at it.

Re: Why we'll use Google Universal Analytics over Mixpanel and KISSMetrics

#10
post #9
post #5

> For one, there can’t be 2 [clientID, userID] couples with the same userID: with the way mixpanel does things, this is essentially a technically impossible scenario (...) And yet one user can access your site through different clients, leading to a systematic overestimation of the number of visitors hitting your site. Really? Anyone can confirm this behavior? I'm pretty sure KissMetrics doesn't have this limitation.

Indeed, and this is why we ended up choosing KM over MP. With KM you just "identify" a visitor whenever you want and if there's already another anonymous cookie, it'll tie together all events retroactively. We couldn't find an easy way to do this with MP when we looked at it.

Yep, it would seem that KISSMetrics has a better implementation where aliasing can be called several times (as stated here: http://support.kissmetrics.com/apis/common-methods.html).

The fact that it links accounts retro-actively though can be dangerous, in the scenario of publicly-accessed devices. I'll have to admit though, this is not the common case.

I guess my personal gripe with what MP and KM are doing boils down to: if you can't infer stuff about who is visiting my website, be honest about it and don't.

Post reply on HN