I have found this thread fascinating and while I have been on this
group for many years, I am not usually actively replying, however I
thought I would chime in here.
I can claim some experience both using analytics systems and building
analytics systems so the long email below is the summary of my 5+
years of experience.
My very brief background: I coded/built and sold a mobile app
analytics company (2009-2011) and also ran mobile marketing business
servicing and building apps for large brands, i also built some well
used consumer mobile apps and some small/indy games. Now I am building
a mobile monetization solution bringing brand dollars to app
developers - www.alphonso.tv.
Given my background, take this entire email with a pinch of salt. (-:
For the impatient - TL;DR version:
- Mobile app analytics systems are great to have for all app
developers. There is something for everyone in it.
- There are many good options both paid, free and open source - BUT I
do not recommend rolling your own analytics system.
- We worked with some very privacy conscious conservative brands so we
did research (actual A/B test on effect of putting or not putting
analytics systems in apps) - we found no effect on download, usage or
engagement number.
Longer version
App Analytic systems are awesome for app developers:
Let's broadly divide dev's in broadly 3 category
--- New app dev's with some apps (few 1000's of downloads)
--- Apps in the trunk (10k-few million downloads and reasonable actives)
--- Apps in the head (10-100s of millions of downloads)
BTW, you could slice it differently - ad supported / free with in-app
purchase / paid apps OR you can slice it - apps from well know names
(big co and big brands) / small co. etc.
Any way you look at it there are the following use cases for analytics systems:
Product usage - I want to know how the users are using my product. It
is sometime about the color of the button, sometimes about the utility
of a feature, some other time it is about segmenting the user base on
the basis of geo or some feature and and in each case doing something
interesting with the next version of the app - to monetize in-app
purchases better, to place ads differently, to understand which
devices are making your app crash in some corner case etc.
Monetization usage - I want to know more about my users - who are
these users, when do they use the app, how often do they use the app,
do they use a certain feature more than another? a good story here is
from a large entertainment content app we worked with, they were able
to identify that they had a lot of hispanic users (based on analytics)
- so they launched premium app targeted to the hispanic audience. this
was purely a decision driven by analytics and it created a huge new
line of revenues for the company. i cite this example because i am
sure there are many game developers on this group who would be able to
cite a similar example for games (either indy or large)
Marketing / optimization usage - your typical A/B test, optimizing
user flow etc are all done via analytics systems.
Advertising / customer acquisition usage - analytics systems are also
a must for anyone who is looking to acquire users by any sort of
marketing - be it CPI based campaigns, mobile ads or even off-mobile
ads. this is the only real way for app developers to know what is
working / not working or rather even more nuanced what is working when
and where so as to optimize the spend accordingly. this is critically
important for channel attribution as well as to know long term where
the best users are coming from. Also useful for user re-activation.
Privacy concerns - I often get push backs from the smart engineering
folks who base the theme of argument on potential and projected
privacy concerns from the user base - my experience has been these are
always way overblown. this is my experience and not just my opinion -
it is backed by what we saw doing large scale research and A/B test.
We worked with some very privacy conscious conservative brands so we
did research (actual A/B test on effect of putting or not putting
analytics systems in apps) - we found no effect on download, usage or
engagement number. We didn't find any significant comments or even
anecdotal evident from the users about privacy.
Comparing various analytics systems:
There are basically two classes - paid / free.
The paid version the trade is - your data is your (you may or may not
get access to it, but it does not get used outside of your app) - and
you pay for service as a software as a service - typically these are
freemium services - localtyics is a good one, there are others.
The free version - typically has some 'ad' model attached to the
service. flurry or google both will at some point use this data in
some way to make the ads smarter / better.
Rolling your own - as other pointed out, getting a basic analytics up
and running is not a big deal at all. log stuff, do some greping and
you are good, script it a bit and you can make it work well for 100k
users. these home grown systems start giving diminishing returns when
you start broaching millions of users. you will find as a small to mid
size app dev shop you are spending more time on these tools than on
what matters the most - your app. also ongoing feature addition is not
your key priority so you won't do a good job and will find yourself
doing a bunch of scripts or queries as your needs change. this is
where using one of the services which are battle tested work well.
So there. My 2 cents.
Please do chime in with comments, I am eager to learn.
Cheers,
Ashish.
www.alphonso.tv
Cheers,
Ashish
On Fri, Dec 13, 2013 at 3:59 PM, Ted Oliverio <google@playnaked.com> wrote:
>
> On Dec 12, 2013, at 11:50 AM, Doug Dawson <dougdawson@mac.com> wrote:
>> A problem with analytics is that you're (at least technically) obligated to disclose the use of it in your privacy policy, and in this day and age people have a tendency to react negatively to that, to an extent that can dwarf its usefulness.
>
> (a) Someone actually READS privacy policies?!?!
>
> (b) Are you aware of any cases where negative reaction to disclosure of analytics collection has dwarfed the usefulness of such collection?
>
> (b)(2) If yes, please explain.
>
>
> Thanks!
>
> P.S. My experience is that analytics, correctly reported in privacy policies, are amazingly useful. If that's the "dwarfed" version, I'm very excited about the potential of people getting over it. :)
>
> --
> A Code of Honor - Never approach a friend's girlfriend or wife with mischief as your goal. There are just too many women in the world to justify that sort of dishonorable behavior. Unless she has really nice tits.
>
>
> --
> You received this message because you are subscribed to the Google Groups "iPhone Software Business" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to iphonesb+unsubscribe@googlegroups.com.
> To post to this group, send email to iphonesb@googlegroups.com.
> Visit this group at http://groups.google.com/group/iphonesb.
> For more options, visit https://groups.google.com/groups/opt_out.
--
You received this message because you are subscribed to the Google Groups "iPhone Software Business" group.
To unsubscribe from this group and stop receiving emails from it, send an email to iphonesb+unsubscribe@googlegroups.com.
To post to this group, send email to iphonesb@googlegroups.com.
Visit this group at http://groups.google.com/group/iphonesb.
For more options, visit https://groups.google.com/groups/opt_out.
Posts
