Web Analytics Industry – International Pulse (US vs. EU)

This is the 1st post in a series of posts about the factual (quantitative researched) state of the web analytics Industry. I also recommend reading the WAA (Web Analytics Association) sponsored qualitative study Global Use and Attitudes Towards Web Analytics - March 2007 by Eric Peterson as well. However; I will be able to answer different questions - simply from super crunching millions of factual report and analysis requests.

As I am participating and speaking in a WAA Webcast titled: “Taking the Pulse of International Web Analytics” (15th Nov. Signup online through WebEx) – I decided for a somewhat provocative (fun) question as the foundation and kickoff of this series of post about the factual state of the web analytics Industry AND as the use-case for my WAA presentation.


Question:
Are US Web Analysts more advanced in their use of Web Analytics than European Web Analysts?


- A “Web Analyst” is defined as an enterprise user who signs into the web analytics tool and request reports and analysis.
- “Advanced” is defined as in the web analysts use of a set of subjective chosen sophisticated reports versus a set of simpler reports

Coming back to the question. Most people would say yes, just because that is the way it is supposed to be, some would say, myself included, that there is no difference (and I have voiced that opinion many times before, based on my experience working with clients in both regions). Let’s have a look at the data:


Dataset:
Period: Q3 2007
Total Report requests: 1,911,913
Total Report requests by Enterprise users: 470,320

- I defined an “Enterprise user“ as users of the high-end (Enterprise) version of IndexTools and as one who built and/or requested a custom report in the session analyzed. Making sure that the results, we are going to look at, are not skewed by the regional distribution of advanced report users versus simpler report users.

Advanced/simple subset:

Total Report requests by Enterprise users: 470,320
Advanced Report requests by Enterprise users: 18,091 (3.85%)
Simple Report requests by Enterprise users: 15,864 (3.37%)

- I defined the “Advanced Report requests” segment as combined usage of the Pivot Table like Custom Report Wizard and the Visitor Segmentation reports
- I defined the “Simple Report requests” segment as combined usage of Most Requested Pages, Referring Domains, Visits and Traffic Summary

This is of course subjective, as I decided for the definition of what “advanced” is. However; I trust we agree that the two pools are quite different and that they indeed represent two levels of sophistication. The reasoning for leaving out the middle segment is that it can be debated, and quite fairly so, whether a scenario analysis or path analysis is advanced or not.

Now that we have the data needed for our question - all we need to do is; to segment the data on the two regions in question.


GEO segmented Advanced/simple subset:
Advanced Report requests by Enterprise users (Europe): 4.04%
Advanced Report requests by Enterprise users (US): 3.45%
Simple Report requests by Enterprise users (Europe): 3.30%
Simple Report requests by Enterprise users (US): 3.59%

- Understand that this is of course not a comparison between “Advanced” and “Simple” – that would not make sense.


Conclusion:
In contrary to what is commonly understood, European web analysts are on an average more advanced than US web analysts. This fact is based on two fundamental conclusions; namely that European Analysts are requesting and creating more Enterprise level reporting and analysis and working less with simpler reports.

Source: 1,911,913 reports requested by Web Analysts using IndexTools (Q3 2007)
(VisualRevenue.com/blog - Dennis R. Mortensen)

This is great fun and you should basically see the above as an appetizer for some of the conclusions that I will provide in this series... I must - of course - skip my trip to ad:tech New York in a couple of weeks, as I am most sure my US piers will hunt me down and force me to render null and void the post. :-)

Signup for the WAA Webcast and be entertained with the results for specific European countries as well, such as UK, Finland, Belgium and Italy – Why? Because that is where my fellow panellists (Steve Jackson, Michael Notte, Alessandro Zanzini and Vicky Brock) are from! :-)

Cheers

Customer Lifetime Value – KPI

This is going to be the first - of a couple of posts - about this metric. Lately, we have seen that Enterprise Web Analytics vendors have put some thought into a Visitor Lifetime Value metric (KPI) – I would like to add the following comments:

First and foremost, I think we should use a stronger description (skipping the traditional Analytics Vendor description of Visitor) and call it: Customer Lifetime Value (which is also the more general and common marketing term) – Secondly, I think one have to accept that a Customer Lifetime Value metric (without even debating how we are supposed to calculate it) directly referenced to e.g. a campaign is far from telling anything about either ROI (we must expect multiple campaigns attribute to a sale/s) or even whether we reached a high value customer segment (as the metric does not recognize whether a customer is frankly lost, have not bought anything in years or otherwise not represented rightfully).

That said; Introducing a Customer Lifetime Value metric, I suggest that the following accompanying metrics be introduced as well:


Customer Lifetime Metrics (KPI's)

• Customer Lifetime Value (definition: Aggregated revenue SUM)
• Customer Lifetime Acquisition Cost (definition: aggregated campaign costs SUM
• Customer Recency (definition: Time since last sale)

Furthermore, given the strong filter and segmentation opportunities in Web Analytics solutions (that compared to e.g. more traditional CRM systems) - defining “Lifetime” (usually a very difficult task) as ALL tracked data so far, seems quite appropriate to me. We can then use this as a proxy for the absolute customer lifetime value (and we probably do not have the time to wait it out OR even the history of enough customer life cycles to have an educated idea anyway).

Then again; coming back to the ever returning dialogue about “attributing” a given sale to e.g. the Original campaign, The Direct Campaign or perhaps something even more glamorous or Intelligent. Does this matter and is this really a concern? – should we not “just” present the aggregated SUM of all Customer Lifetime Values of those visitors who reacted to a campaign. As in, if 19000 visitors reacted to a campaign (as per a campaign summary) then a SUM of all the visitors Customer Lifetime Value is presented alongside the other campaign metrics. The period chosen (e.g. February) is used to present active campaigns and their visitors in the period, but will otherwise not have any implication on the SUM calculation of all the visitors Customer Lifetime Values.

Then more importantly; one can filter/ segment on Customer Recency (dependant on business-model) to create a better understanding of what kind of customers (and their value) we are looking at, thus more an indicator on whether you are talking to the intended audience (as in; is this your typical repeat buyer, the lost customer coming back etc.). AND if one want to generate a true (or probably just better) ROI calculation; a Customer Lifetime Acquisition Cost metric can be applied.

This is a good beginning – leading me into the second post about the Customer Lifetime subject. How one should consider using the described Customer Lifetime Value metric (KPI) as part of ones SEM optimization efforts.

Emetrics Marketing Optimization Summit - Washington DC – 2007

I attended the Emetrics (Marketing Optimization Summit) in Washington DC earlier this week. And as always, with this mandatory event, everybody was there and every single minute was used ~ SUPER EVENT! .. It felt like I spoke to every single vendor, practitioner, industry analyst and blogger in the analytics (marketing optimization) industry, over the course of three days.

The following Web Analytics vendors participated as exhibitors:

• Omniture (OMTR)
• IndexTools
• Visual Sciences (VSCN)
• Webtrends
• Google Analytics
• Unica

You can catch a few comments from the show here:

• Emetrics Summit Presentations on Second Life
• Back from eMetrics Dec '07
• Emetrics Washington 2007: Consumer segmentation, personas and ...
• Emetrics: Blog Promotion and Measurement
• eMetrics Wrap-up: Am IA Psychic Marketer?
• eMetrics: Rethink SEM/PPC Analysis by Avinash Kaushik
• Live From eMetrics
• From Emetrics Summit: Google Analytics Revamps, Microsoft to Compete
• Exciting Announcements at eMetrics Today
  

...See you back home in Stockholm for the next Emetrics - Jim, Matthew and co.. :-)
Cheers

What and how to measure Online Travel and Hospitality websites

In today’ multi-channel marketing world, the Web is quickly becoming not only the first “point of contact” with an overwhelming majority of travel and hospitality customers but also a key tool in achieving incremental growth and competitive advantage. What’s more, an effective Online Marketing strategy in travel and hospitality plays a crucial role in establishing interactive relationships with customers and utilizes all channels available on the Internet.

To gain advantage in this highly competitive industry and win customer loyalty, travel and hospitality website owners need insight into:

• performance of their online business across customer segments, markets, properties, packages, routes and destinations
• effectiveness of their online marketing initiatives by traffic source, online channel, advertising partner or individual campaign
• success of their e-CRM efforts in terms of customer satisfaction, retention and loyalty

But as always - there MUST be a set of measurable KPI’s that are strongly aligned to ones business objectives, like any other organisation that operates with the web being a channel of theirs. Hereby concluding that KPI’s are meant to be acted on and not just reported on. Therefore I suggest optimizing towards the following two business objectives.


Online Travel and Hospitality business objectives:

• Revenue
• Profit margin growth

These are determined by a number of indicators, the most important being:

- Volume of bookings (total number of bookings)
- Bookings (sales) conversion rate (number of bookings/number of visitors to site)
- Average booking value (booking revenue/number of bookings)

For tactical decision-making, you need to break these strategic indicators down into manageable segments and analyze their performance by - for example - geography or property, in order to understand how well each is contributing to the overall result; or, assess them by traffic sources and marketing channels to see where you should direct your marketing budget.

But, let me outline a set of leading KPIs which will enable us to gauge the effectiveness of online initiatives and help us steer a Online Travel and Hospitality websites in the right direction:


Online Travel and Hospitality KPI’s:

• Bookings conversion rate
• Average booking value
• Look-to-book ratio
• Top Geography and Demography
• Return on Ad Spend (ROAS)

 

Online Travel and Hospitality optimization opportunities:

1. Visitor profiling and segmentation in Travel and Hospitality.
   
   Knowing your visitors is an extremely important consideration for the success of your Web site or online marketing strategy. Addressing your key audiences and providing them with relevant information is one of the key aspects of any travel and hospitality site. If your site does not speak directly to each of these audiences, you will lose most of them to the competition.
   
   Segmentation can help in the analysis and understanding of how important groups of people use the site by looking at the page-by-page activity of the group without being distracted by activity from other types of users. Some of the key criteria for segmentation would be:
   
   - Demographics
   - Traffic Sources and channels
   - Activity and Navigation patterns
   - Custom built segmentation criteria
   
   Opportunity: Target the right groups of visitors with the appropriate marketing message, adapt content to users’ preferences and improve usability to fit your visitor’s browsing behaviour.

 

You can find a complete and in-depth PDF version of our inputs for Best Practices for Online Travel and Hospitality here:

DOWNLOAD GUIDE (4 page commercial version | 754K PDF)

 

Book: Fooled by Randomness

When I commented on the book Freakonomics back in May – I had a bunch of entertaining comments about this being a book that had been out for years. With that I mind, I decided to continue my noticeably antique view of the world with a comment on an absolutely FANTASTIC book I read recently: Fooled by Randomness (2001) by Nassim Taleb.

For those of you (who for some reason, like me, did not get to read this book when published) – Nassim Taleb is a mathematics and statistics adept trader, but perhaps more a philosopher who has out of the ordinary comments about probability, psychology and life.

This book should not be recommended reading for a web analyst or online marketer – but mandatory! I simply cannot recommend this enough and I will bring into play a handful of his thoughts over the next coming period. Such as:

• Monte Carlo simulations
• Data snooping
• Causality
• Path dependent outcome

.. and if the above subjects does not turn you on. It’s time to buy a Porsche.

The final question (for myself) - Am I fooled by randomness along with everybody else? hmmm :-)
Whatever the answer is -- GO READ THE BOOK!! -- I will question you in random aspects if I meet you in Washington (Emetrics) next week.

Cheers

Search-dominant visitors vs. Navigation-dominant visitors. (+ how to segment)

I somewhat recently wrote an article on how to determine the width of an internal search box. However; this post do not debate whether you should have an internal search tool to begin with and if so, how prominent you should feature it. So please notice the previous post as a successor to this post, as I suggest one investigate visit search behaviour before branching out into analyzing usability elements such as search box WIDTH. I believe it is possible to segment visits (not visitors) into largely two distinct direction-finding groups:

• Search-dominant visits
• Navigation-dominant visits
• Bounces*

Before debating how we set up the segments described above in our Web Analytics Tool - I would like to add a comment on why I think this is a significant discovery. First of all, I strongly believe and find it common sense, that making a visitor suffer through poor direction finding elements (traditional navigation or internal search) is bad for business. And “poor” is defined as in not being able to find the content one is looking for.

A segmentation suggestion could be along these lines (whereas other metric filters spring to mind as well), but

Search-dominant visit segment

• Exclude all bounces (a bounce used as a definition of a visitor who used neither traditional navigation or Internal Search)
• Include all visitors who searched on the first page of their visit path
  

Navigation-dominant visit segment

• Exclude all bounces (a bounce used as a definition of a visitor who used neither traditional navigation or Internal Search)
• Exclude all visitors who searched on the first page of their visit path
  

Bounce segment

• Include all bounces (1 page view visits)
  

Example
Let’s try it out and have a look at my grand little blog here (which holds basic navigation and an internal search). I set up 3 segments as described above and applied them to a basic visit report.


(As you can see; focus should not be too much on Search, which is generally the fact on most blogs, but I am most sure you see the value of this beyond blogs)


Conclusion
If you have predominately Search-dominant visitors then you should engage in serious usability analysis about how to incorporate internal search into your overall site navigation - this including such elements like how to determine the width of your internal search box. Usability studies that are likely to tell you that you are far better of listing the 5 most searched for terms than expanding you traditional navigation.

On the other hand, if you have predominately navigation-dominant users, you will be wasting valuable screen real-estate on search elements not used - and not only that, you will be sending your users down a path they are not comfortable with. So perform the above analysis to figure out which path (search or navigation) you should focus on.

.. And remember that this is most likely different from section to section of your web property.