Traditional Web Analytics is Dead
We have all lived "web analytics" analytics for a long time. There are practitioners and vendors and a well established "system" of working and thinking. Yet companies are frustrated at their ability to fundamentally understand the data and take action. I have come to the conclusion that there are multiple forces that are at play and the world of traditional web analytics is now dead.What does the world of web analytics look like? If we measure any of the following it is likely that we live in the world of web analytics:
- Page Views
- Path Analysis
- Track HITS
- Daily Unique Visitors (!!!)
- Top Exit Pages on the Site
- Visitor Screen Resolutions
The core outcome of measuring these "KPI's" (Key Performance Indicators) is that we can take very little action and they are distracting from focussing on core KPI's and driving change for our customers.
I believe that we are now living in, heading towards, a world of Web Insights. In the world of Web Insights we don't simply measure KPI's but rather we do Key Insights Analysis (KIA's) such as:
- Click Density Analysis
- Visitor Primary Purposes
- Task Completion Rates
- Segmented Visitor Trends
- Raw Voice of Customer Pareto Analysis
- Multi Channel Revenue Impact
You know you live in the world of Web Insights when you realize that every piece of data you look at drives action and not just action but action that adds to whatever bottom line Outcomes that our companies are trying to achieve for our company customers (note that important difference, not outcomes that your boss wants, not outcomes that his/her boss wants, but Outcomes that your customers want).
The world of Web Insights takes time to move into but once you get comfortable in it you will have achieved a long term strategic advantage.
For more context on this topic you can check out my podcast/interview at Internet Marketing Voodoo or download a mp3 version from the Media Mentions, Podcast page.
Agree? Disagree? Please share your feedback via comments.
Subscribe 
May 15th, 2006 at 13:39
Hi Avinash! It's good to see that the Green Eyed Blog Monster motivated you to join us out here.
These (web analytics are dead) are great thoughts but I don't think traditional web analytics are anywhere close to dead. For you, yes. You are really advanced. But most of the world now has a website with crummy analytics that they still don't even know how to look at. And that includes e-commerce customers (just not all of them.)
This reminds me of Shane Atchinson's comment in ClickZ (which I am too lazy to go and find) about non-linear conversion funnels. Yes, conversion is not always linear and it is a cool idea. But how about all those people who still don't understand what a funnel is?
Of course, when I wrote about this one my blog, it caused a real uproar…
Robbin
May 15th, 2006 at 18:07
Robbin, Thanks so much for your comments.
I think you overestimate how "advanced" I am, that's not true. :)
You do have a great point that most of us are still trying to figure how out to make sense of what Omniture / Coremetrics / WebTrends / Your Favourite Tool is "spewing out". That is the exact reason that we need to abandon looking at "traditional KPI's" (think Visitors, Page Views, Exits, Time on Site, Conversion). We can jump into the world of KIA's (Key Insights Analysis) and rather then think just KPI keep a perspective on our business goals and then try to find nuggets of insights.
Here is a really simple one, rather than measure Average Page Views Per Visitor how about Page Views by Referring URL's or Conversion by xxx and yyy and zzz. Not really smart stuff but not really just the KPI world.
-Avinash.
PS: Non-linear Conversion Funnels are really cheap, just nine hundred bucks, or a free trial, with ClickTracks.
May 16th, 2006 at 08:11
Hi Avinash, it's great to see you blogging. Your ideas and experiences will keep the industry pushing ahead. Cheers!
May 16th, 2006 at 15:37
I am more partial to conversion rate per referring URL and conversion rate by landing page, but maybe that is just my way of slicing and dicing the data to figure out what creates an average conversion rate.
I hope you will take those KIAs you suggested and write a post about each one. How to calculate, what it means, why it's important. I think that this is what we miss the most in WA, that level of nitty gritty "how to" knowledge. And the reason (I suspect) is that once one knows how to do something, it's so easy, one must think, Ah, the whole world knew how to do this except me, that's not blogworthy. Avinash, always remember that you are at the cutting edge and any "how to do this" information from you is valuable.
Robbin
May 22nd, 2006 at 11:13
Hi Avinash,
I think you are right, but you are an advanced user! You will just have to trust me on this – in my role I get to meet about 25-30 of the top companies every year, and having seen what you do I can confirm you are in the advanced cataegory!!
But is web insight not a stepping stone to proper integrated customer insight that includes online behaviour? I know you bring together much more than pure web behavioural data to give yourself a sandbox to play in. Much of what you talk about above is still based upon anonymous user session, where as customer analytics (predictive analytics according to others) deals with individuals.
THanks for the post… gets the brain going!
Talk soon
Matthew
May 22nd, 2006 at 11:50
Matt: Thanks for sharing your thoughts. If I am advanced then there is no hope for our world of web analytics!! :)
Oh absolutely positively online behaviour is a critical component. In my mind it is not where I would start. The ordering is Business Strategy -> Fantastic understanding of your customers -> Qualitative data -> Hard core clickstream analysis -> Money, Money Money!!! :)
At the moment we often, all of us, miss the first few steps and move to "hard core clickstream" and sure enough money does not follow.
May 22nd, 2006 at 12:30
Agreed on your process!
In fact let me paraphrase it, and then request you do a blog on each section.
1.Business objectives – money making plans
2.Great customer understanding (Web analytics?)
3.Qualitative data (Define? What works best?)
4.Web insight – the hard core clickstream analysis
5. Action agenda – doing stuff that will make more money.
So how about a blog on each point? Now there is a challenge you surely cannot refuse!
Thanks
Matthew
May 22nd, 2006 at 12:34
Matthew: Ok I accept. I have thoughts in my mind on each of these topics and it will come out in due course but I will keep your wish list handy as I think of things to talk about. Thank you.
May 26th, 2006 at 08:02
Dear Avinash,
I am a online advertising professional, who has caught the web analytics bug recently and I follow your blog and find it extremely useful. Since I am from the lead generation and hence optimisation background, I would like to see your view on the convergence of online marketing efforts, like banner advertising, SEM, whitepaper marketing, etc.. being analysed through web data and unification of different reporting and hence understanding of customer and their behaviour on the website.
I am particularly interested in understanding the intergration of adserver, other induvidial reports (through APIs) and webanalytics in one single window, to enable us to understand the marketing activity in a single window and a deeper insight in the consumer behaviour.
May 27th, 2006 at 00:30
Jaisri, I am glad you find the blog useful. Thank you for taking the time to comment.
This is not as hard as you think it is, in fact many tools already make this possible. For example the tool we use, ClickTracks, can directly download the data from google adwords or overture and show integrated reporting, as you put it, in one window. The cooler part is that then if I want I can go segment behaviour by that integrated data and understand customer behaviour much better.
The same thing applies for adserver as well. I saw a really nice implementation in Omniture (specially with doubleclick) where data came in and integrated into reports very well.
At the end of the day any anlaytics tool will do what you want. Some better, others will need more work. The most important part is this really simple part: Think about tracking before you implement these campaigns and then ensure your marketing efforts are passing identifiers (url parameters, cookie values etc) specific to efforts to your website (and thus in turn to your analytics tool). Then you are only limited by the ability of your tool to slice and dice, and most do a decent job.
The point of failure usually is that we don't think ahead of time and we don't make our agencies / people working on our efforts make that extra trivial effort to pass us identifiers.
Hope this helps.
May 29th, 2006 at 01:53
Thanks a lot. Does "Google Analytics" support this feature?
Rgds,
jaisri
May 29th, 2006 at 09:33
Jaisri, Google Analytics / Urchin does allow for integrating data, for analyzing campaigns. I am not a power user of GA but would refer you to this blog: http://www.roirevolution.com/blog/ they have great stuff about GA.
Thanks for your comments.
-Avinash.
August 29th, 2006 at 07:02
More than great …
go on
September 10th, 2006 at 03:07
[...] So Web Analytics is not only going to be adopted by companies, it's also going to evolve and embrace other things not just 'web'. The vision will be much more holistic as it will be possible to follow the customer trends on different medias, make correlations and thus improve the experience of the users. In the end this is the only long term strategy to succeed today and even more so to keep ahead in the near future. So here I wanted to name again Avinash who speaks about Web Insights instead of Web Analytics. If you allow me Avinash, I would go a step further and I would suggest that we need to find another word for web. What could it be? Business (Business Insights), Customer (Customer Insights), Behavior (Behavior Insights)??? I guess that we won't find it overnight but here's something the industry (and maybe the WAA) should start thinking about. This new field will in the future be integrated with BPR (Business Process Reengenering) and on the long term, using artificial intelligence (AI) it will adapt itself to report on new findings… But well, that's the future and most companies that I know don't even use basic Web Analytics. Measurement leading to actions is one of the strongest competitive advantages any company should focus on (don't get me wrong I'm not only speaking of a WA package but most important: the knowledge and experience that allows the right interpretation and action decision!). [...]
January 9th, 2007 at 11:49
[...] A dare un'occhiata in giro e a riguardare quello che è successo non si può non dargli ragione: in fondo il 2006 è stato l'anno dei dibattitti sulle classifiche e sulle mappe e d'altro canto, anche altrove, l'esigenza di definire delle metriche per i blog che vadano oltre quelle tradizionalmente adottate per i siti web, da molti date per morte, appare sentita. [...]
June 29th, 2007 at 13:36
i listened to your podcast however i was not able to quite understand when you talked about the limitation of path analysis – you mentioned something like given that 1-5% of the people follow a similar path – there would never be a reason for you to change your website ?? could you please elaborate i am not sure if i completely get it.
August 19th, 2008 at 18:32
[...] nenhuma ação), como Page Views, Hits, Daily Unique Visitors e Principais páginas de saída. Avinash disse que as análises baseadas em métricas como essas estão mortas e que tiram nossa atenção das métricas que realmente importam, que são as que resultam em [...]
January 27th, 2009 at 09:50
Hi there Avinash,
I just wonder, what difference it makes using KIA instead of KPI, when we start crawling around then in mounds of data again?
As you said above:"We can jump into the world of KIA’s (Key Insights Analysis) and rather then think just KPI keep a perspective on our business goals and then try to find nuggets of insights."