I've been working with Google Analytics since, well, before it was Google Analytics (remember back to the days of Urchin On-Demand?) and I've seen the product evolvetime and again.
While everyone has their gripes with GA (I certainly have mine), I've
seen time and again that people take shots at Google Analytics with
uninformed ammunition. There is a long list of myths circulating out
there about GA - in this post I hope to dispel a few of them.
This time around, I'm going to use the recent post by Annie Wallace at Search Engine Journal entitled "9 Sins of Google Analytics".
Now, I don't know Annie - never met or talked with her - and I believe
she meant well in what she wrote. In fact, I agree with a few points
that she makes, however teach of the 9 "sins" pointed out are, I
believe, incorrect. Claims of GA's failings are common, and I think
much of that stems from the assumption that it's a free tool and can't
be that good, plus a lack of knowledge and expertise in setting things
up, so the outcome is sub-par. Thus, let's look at each claimed sin
and review the facts to see if we can't bring some redemption to the situation.
Redeeming Sin 1: Data Ownership
Who owns your Google Analytics data? Do you control it, or can
Google use it for other purposes - sell it, mine it, etc...? I've
heard concerns and complaints around this day after day. A common
belief is that Google can and does do whatever it wants with your GA
data. But is that true?
First off, let's look at the terms of service for Google Analytics.
I'm not an attorney, but I don't think that's a requirement to read
these terms - they are pretty straightforward. Right at the beginning
in section 1 "Custom Data" is defined as "the data concerning the
characteristics and activities of visitors to your website..." I.e.
"all your data." Reading on in section 9 we see that any 3rd party
(yes, that refers to agencies, web developers, etc... who might setup
GA on behalf of a client) must have a legal agreement with the company
GA will be used for that they (the company, not the agency) "owns any
rights to Customer Data in the applicable account..." So, here in the
legal Terms of Service, we have Google defining anything collected by
GA as your data and setting the precedent that the data must belong to you.
Now, granted, it would be nice if there was a crystal clear line in
here that says "all your data are belong to you" to clear up the
misconception that "all your data are belong to us."
Second on this point, whether or not the TOS is crystal clear, if
you've Looked at the settings in the last, oh, 4 years, you will see
that there is a really nifty feature for controlling data sharing and usage.
If you don't want your data shared to other Google services then it's
not shared. If you want to share only to other Google services that
you use, so that features like AdWords integration can work, then you
may activate that. The third option opts you into benchmarking reports
- but even that doesn't share anything beyond aggregated,
non-site-identifying data points. Bottom-line, all your data belongs
to you and you decide what to do with it.
Now, I'll end this point with the observation that there is a drawback in all this: while your data belongs to you, you don't have full control
over it since the data resides on Google servers. For example, the raw
tracking hits aren't available for exporting - just processed report
data via the API. But, even this has a solution: Urchin. Yes, it is still around and can be used very nicely in conjunction with GA to act as a backup and re-processing sidekick for GA users.
Sin 1 Redeemed: You, not Google, own your data
Redeeming Sin 2: Trouble with Support
I just have to chuckle when I hear this one. Really? Serious
competitors to GA border on obscenely expensive and still don't have
free support, and people expect Google to provide excellent support to
all users free of charge on a product that is free of charge. The math just doesn't add up.
What would be nice is if there was a paid
route to support - at least then you could get help when you need it
and the cost would still be far less than the other major web analytics
tools. I have good news: you can do just that through with network of over 200 consulting firms
worldwide that Google has vetted, trained, and approved to provide
professional services on its behalf! Ladies and gentlemen: meet your
friendly neighborhood GACP.
This is where I should disclose that I work for a GACP, so I
recognize I'm inherently biased. But, in five years of doing this,
I've met literally hundreds of people who tell me "GA can't do this"
and then I show them that it in fact can, and hass been able to for
years. The response is usually "oh, I didn't know that, but that's
awesome!" My point: the capabilities of Google Analytics aren't
limited to your experience with it.
Now that we've cleared things on that, let me preach a bit about this.
The GACP network is truly like nothing else in web analytics. This
is a vibrant global organization larger than any other web analytics
consultancy filled with people who are fanatic about the Google
Analytics product and love to help people. Where do you have a forum
where these partners help each other daily, answering questions and giving free advice to their competitors? We have that in the GACP network. We love to have people use GA.
Now, we don't work for free, as I'm sure you don't either. I think
that's where a lot of the misunderstanding comes in. In some weird
way, because GA is free, people expect help for it should be free to,
but if it were to cost a billion dollars, or even just a thousand,
people would probably be less intimidated by the prospect of paying for
help. I don't understand why... but it is. Imagine if Airlines
suddenly made all seatson their flights free, would you expect a free
lunch too? Actually I think a lot of people would. But, when we pay
hundreds for a seat, paying $10 for a bag of chips and a soggy roll
doesn't seem so bad. Strange.
If you don't want to pay for a product or pay for help using it,
then you'll need to become an expert yourself. Which, in fact, you can
do. There are tons of resources on Google Analytics - you just have to put in the time. I like the training videos for the GAIQ
the best. Personally, it took me many hundreds of hours working with
the product to reach what I look back on as a solid "intermediate"
level of expertise. If you're there, or working there, more power to
you - and send me your resume!
Sin #2 Redeemed: there is help available, and lots of it!
Redeeming Sin 3: Reporting Limitations
This one really gets me going, particularly that the claim here
states "for example, the conversion data is provided in percentages
instead of numbers." Yes, that's true, if you only look at one report. Out of the box conversion data is listed in specific counts not just ratios under the Goals report section where it can also be added to the dashboard.
Furthermore, Custom Reports have provided a means of getting conversion counts for years.
But about those limitations...
I do have to say that I partly agree on the general point
here of a more rigid box. I'd love to be able to create "custom
dimensions" and "custom metrics" within GA and then build a dashboard
that incorporated those custom dimensions and metrics with custom names
and annotation data to explain everything.
That said, if I'm going to do all that in-tool, I might as well do
it externally in Excel. Which clears the dust on this one: the data
export API is awesome, and the exploding ecosystem of applications
that are built atop it provide ample ways to get and report your data,
however you like it - flame broiled, with extra cheese and a Coke
Sin 3 Redeemed: reporting is only as limited as the box you choose to stay sitting in.
Redeeming Sin 4: Limits to Defined Variables
Claiming this "sin" borders on irresponsible. Way back in October
of 2009 Google released a major feature improvement to user-defined
segmentation variables called Custom Variables.
Writing that "Google Analytics allows only one user-defined variable
for segmentation" when that feature has been deprecated and surpassed
with a new feature that allows 5 variables with 3 scopes and virtually
unlimited name/value capabilities is very misleading.
But, Custom Variables are hard
Yes, Custom Variables are hard setup - but that's the point, they
are an advanced, powerful feature and will do exactly what people used
to complain the old single-dimension user-defined segmentation
couldn't. If you take a few steps back and think about what your
business needs are before you start implementing things, you'll be able
to plan usage for Custom Variables that will make you dizzy.
Sin 4 Redeemed: There are more than one dimension for user-defined segmentation
Redeeming Sin 5: Slow Data Delivery
I partially agree with this. First, though, it's important to note
that data usually only runs an hour or two behind in Google Analytics.
Just set your date range to include today's date and you'll see what's
current up to about 2 hours prior.
That said, real-time data is indeed absent from Google Analytics.
However, I have to ask: what does real-time do for you? It's super interesting, even addicting (I like Woopra - ooh, look at all the pretty dots that are blinking), but is it valuable? What action
will you take based on real-time observation of your data? If you have
any amount of traffic, you can't watch hundreds, thousands, or tens or
thousands of data streams at once. If you're a small site and like to watch people as they browse your site because you have nothing else to do, fine.
At the end of the day, real-time doesn't matter until you can start
to do things like content customization, instant trending analysis, and
the like. I'm not sure there are any web analytics tools right now
that even do this - if you know of one tell me!
Sin 5 Redeemed, partially: not real-time, but real-time isn't all it's cracked up to be
Redeeming Sin 6: A Lack of Log Files
Umm, I really am not sure I know what this point is about in the original post.
It seems to be the issue Annie is taking is that Google Analytics
can't re-process your original raw tracking data, which is true. But,
that is and has always been remedied by adding a simple addition to
your GA page tags and sending a copy of tracking hits to your own logs
or an external source for collection and processing with Urchin.
If you want to retain your data, you certainly can.
As for the other points here - filtering? Check. That's available.
Filtering retro-actively? Not on raw data, but the Advanced Segments
apply backwards to all your data and for most needs, that will suffice.
Sin 6 Redeemed: you can keep your logs!
Redeeming Sin 7: No Spiders - who likes spiders anyway?!
Right. Again, this claimed sin leaves me shaking my head. A lack of spider data is a good thing for web analytics data. Who wants to pollute your data about people with data generated by robots?
If you do want to know when spiders are crawling your site, sure, get
a logfile analysis tool like Urchin or even AW stats. But, with a
tag-based, hosted web analytics solution the point is to avoid data
pollution from non-human visitors!
Sin 7 Redeemed: it's supposed to exclude tracking of spiders
Redeeming Sin 8: Paid Campaigns Showing as Free
This is a problem indeed, if you don't read the instructions on the label!
I get a lot of questions about this toppic, so I'm going to explain
how this works in GA. Basically, Google Analytics goes through an
order of operations when trying to identify traffic sources. If you
don't tag the traffic sources you control with something that
identifies it, GA will do its best to guess at what the traffic is.
Here is the Google Analytics traffic detection methodology, in plain
- Look for campaign tracking
parameters in the URL of the very first pageview. Those
"?utm_source=someplace..." bits are tracking parameters, and GA looks
for them every time it runs. If it finds them in the URL, it will
update the __utmz cookie with the information contained in the
parameters. What exists in the cookie is what shows up in reports.
- If no campaign tracking parameters
techno-speak this is the document.referrer property. This is generally
set to the page you were on immediately prior to the current page, if
you clicks from that page to the current page. There are cases where
this doesn't work, like when a link bounces through a meta redirect or
for links going from HTTPS to HTTP pages.
If a Referrer value
is found, Google Analytics will use the information to try and
determine if the visit is coming from a known search engine, or not.
If the referrer matches the list of known search engines, the "medium"
will be set to organic, the "source" to the search engine name, the
"campaign" to organic, and the "keyword" to the search query used at
the search engine. If, however, the referrer does not match a known search engine, then the medium will be set to "referral" and the source to the domain of the referring site.
- Lastly, if no tracking parameters
were found and no referrer was present, GA will set the __utmz cookie
to report "direct/(none) for the Source/Medium. This is essentially a
"catch-all" identifier. If you consistently tag all your campaigns and
don't use redirects that corrupt referrer strings, then "direct" can be
trusted to consist primarily of peoplewho typed in your URL directly.
So, will GA mis-attribute paid traffic to "unpaid" sources? Sure, if you don't tag it. But it does this by design
- if you don't name it, the thing won't have a name. So, don't go
around believing GA can't track your ads - it can, and very nicely I
Sin 8 Redeemed: GA can track paid traffic, and do so very well indeed
Redeeming Sin 9: Revenue from Email is unknown
This one is totally random! This last claimed sin again
baffles me. Once again, what wew have here is a failure to follow
directions. If you don't setup your email to be tagged and you don't
setup Google Analytics to capture revenue, you won't get revenue
attribution for email. Many things in life work like this. For
example, I just learned (the hard and very expensive way) that if you
don't change the filter in your furnace, really bad things happen to it.
Why in the world would you expect Google Analytics to
attribute revenue to email if you don't (A) record revenue to GA and
(B) record traffic from email? You can't track what you don't tag.
Sin 9 Redeemend: what more can I say here?
I appreciate that people lend a critical eye to Google Analytics -
that is what makes the product better, because the team at Google
listens to that feedback. But there is a distinct difference between
informed criticism and spreading uninformed assumptions. So, to all
the Google Analytics critics out there, please be diligent in your
fact-checking, then bring on the criticisms and we'll see Google
Analytics get even better.
And for those of you who are thinking "but can Google Analytics do ________", try me! I bet it can be done.