Is Google Analytics bad?
06/04/2011 | Written by | Categories: Analytics Philosophy

Abigail tweet referring to Google Analytics being bad

This is a tweet from Abigail Harrison, the MD of the specialist PR agency thebluedoor during a session at GeeknRolla.  It was flagged up to me by a friend in common (via Twitter) as something I would disagree with.  But I find myself unable to disagree with the sentiment that Abigail is expressing.

Business owners and managers don’t need lots of data.  They need insights, intelligence and recommendations.  They don’t need to be spending time navigating through Google Analytics reports (or those from any other web analytics tool).  They need to be told what they need to know in order to make the decisions they need to make.

It isn’t Google Analytics that is bad, it is too much data that is bad.  The solution is to let someone else who is a specialist in that field do the heavy lifting for you.  Either by setting up dashboards within Google Analytics or in Excel so you immediately get the key numbers you need or by interpreting the data for you.

Google Analytics is great because it captures so much data.  It’s just that not everyone needs to be exposed to all of it.

10 responses to “Is Google Analytics bad?”

  1. Such as statement is like blaming the car for speeding and having an accident… the culprit is GA (or any other tool), it’s the person using it who probably doesn’t have the right process in place to make analytics really valuable for the business.

    People shouldn’t get blinded by too much data, just as they shouldn’t get tempted to speed just before their car can do it.

    Throughout my research with the Online Analytics Maturity Model ( I have seen that very often. If Abigail did a maturity assessment (available on my site), it would likely show their objectives are not well defined and/or the scope is way too big for their capacity. I suspect the “Continuous improvement process and methodology” dimension would also be lower than the other critical process areas of the OAMM.

    Stéphane Hamel

  2. agreed 100% – in fact as a noob in the analytics arena I prefer services such as Postrank where numbers make much more sense

  3. Abigail just said what she saw ie. too much data – there is no crime in that.

    What is a problem/concern/need for her is to have the GA data nicely manipulated into a shape which allows her to glance at it and make informed decisions on behalf of her clients. This is a totally understandable and frequent request!

    The only question is who is gonna help her and her colleagues out?

  4. Shame that the live tweeting /reporting included a typo in my tweet! Sorry all.

    @Peter – thanks so much for the blog post: your last line sums it up for me.

    @Stephane – I think that you’ve hit the nail on the head re: the person using it and their ability to derive meaningful insights from the data that is there. Clients are so often bombarded by the complexity of the data (i.e.: we’re really working hard for you / your budgets) thrown at them, rather than being given suggestions, insights etc. – the answer to “Tell me what this means.” Thank you for pointing out your Online Analytics Maturity Model – I will certainly be rolling up my sleeves with that one.

    @Julius – great comment! 🙂

    @Rob – Because of where we sit in the industry (PR / Social / SEO) luckily we are able to help ourselves on this – thank you for the offer though! And if not, that’s where brilliant people such as Peter etc can help us.

  5. It isn’t the fault of Google Analytics – it is merely the tool (and a brilliant one for the price ie free).

    As usual, Avinash Kaushik has the best word on this: less web reporting (data puking) and more web analysis. That’s down to people, not GA. Web analysts need to be more familiar with the real goals of the business – and execs need to more clear about what those real business goals are.

  6. I just love data, so the thought of too much is hard 🙂 But it’s very true that a trap we can fall into is lots and lots of numbers, but very little insight.

    I’ve found in my experience the trick is to use a number of tools, and be sure to include some that perform “out of the box” standard metrics, as well as having access to others that let you dive into the data and get your hands dirty. While the standard metrics can be very powerful, sometimes you need to create new metrics- and you need to make them using raw data.

    But I agree that the goal in doing analysis of both types is the same- end up with a simple, clearly explained analysis that in real business terms defines ACTION for improvement- not a huge pile of data.

    The challenge of this is what I think makes being an analyst so rewarding- when you get it right, the business understands, listens and acts.

  7. Peter O'Neill says:

    Thanks to everyone for their responses and the conversation on this topic. It appears everyone is broadly in agreement on the key points.

    GA is just a tool and, like any tool, what counts is the person in control of it.

    And that data by itself is not of any use, it needs to be transformed into insights to make it understandable and therefore valuable to a business.

  8. Agree with the broad consensus here, that too much information is a subjective statement. If I buy a car, there’s no way I’m reading the owners manual, however my friend who is a mechanical engineer, will absorb it cover to cover and use the information to ensure that their vehicle is in pristine condition years down the line.

    An additional point to make regardless of who I am as data interpreter, is that all data is meaningless without context. E.g. a single line of data on a graph is pretty useless without a comparator, or an understanding of the journey to that data point.

    As an example “bounce rate” is a metric that can be indicative of two entirely opposing outcomes. 1. The page I arrived at did not answer my question, or did not appear to be able to answer my question in the time I tool to make my assessment. Or 2. The page I arrived at completely and entirely answered my question, thus satisfied, I left.

    Web analytics; the amount and variety of quantitative data available therefore has an additional danger. Not one of information overload in my opinion, but information over-reliance. Without the support of qualitative data, an understanding of the user-journey, motive, problem etc. our assumptions and interpretation may on occasion be flawed.

  9. Charting clean data is the easy part, figuring out which parts of the large amount of data trending towards messy matters is the hard part.

    Some free tools which you can use to analyze GA data include Weka, R and NumPy.

    Lots of resources from universities are available freely if you have the desire to expand your toolkit to include real analysis tools.

Leave a Reply

Your email address will not be published. Required fields are marked *

Looking to improve your digital analytics or conversion rate optimisation?
Contact us today on +44 (0)20 7148 5970

Alternatively, come visit us in London and we’ll buy the coffee!
Get in touch
Leave us a message
Use the form below to leave us a message.

Lindsey Street