I wrote a detailed post about the Emetrics Marketing Optimization Summit Wrapup and the Eutectic Point over at Webmetricsguru.com but I left out some detailed thoughts about my takeaways from Emetrics that I want to share here.
- Focus on optimizing the entire user experience rather than one specific action – metrics to fall in line with this.
- Predictive Analysis can’t be done from Web Analytics Data (Avinash Kaushik) but statistical analysis can be.
- Engagement Metrics platforms are emerging rapidly and will be ready for prime time in the next year or so (Baynote Systems, Nuconomy (see interviews). Also, Monster Worldwide could give Buzz Monitoring a try – forward looking – looked at Buzzlogic.com in detail (see interviews)
- Engagement is a proxy or events that can’t be fully measured (Gary Angel); Angel distinguishes between three types of Engagement – with Brand Engagement being the most valuable to track.
- Google Analytics now processing log files every hour.
- New York Times uses many tools (New York Times Analytics Session) for different stakeholders but seems to have a problem integrating them- it’s also a very political environment based on the session I attended. However, I was impressed with NYT using historical database data to predict how many extra copies of the paper to print when a hot news story emerges (i.e.: Governor Scandal).
Nuconomy Studio measures semantic level website and rich interactions, Word press, photos, videos, comments, posts, etc. The platform is free for most sites (unless it’s heavily trafficked). Nuconomy can also be run in an Intranet environment on a company’s own servers (but that’s rare).
Nuconomy’s business model is different than other analytics platforms – they believe data is a commodity and allow you to data mine your own data in ways that are not easy to do in most other platforms.
A generic dimension is easily generated is in the studio platform and can read data out of metatags or URL tags (and the SOAP information can be sent directly from the database bank ends.
Formulas can be created on the fly – (i.e.: find community leaders) or can be derived from a correlation metric – depending on what your business model the package will calculate a formula for you (i.e.: against pageviews, visits, etc) – this comes from Data miming data.
Processing data is done every three hours but will soon be done every hour.
You can drill down to any user and get a “interest map” of that user; Nuconomy can also determine the content that is “most engaging” to a particular user or set of users.
Another unique feature of Nuconomy is the ability to analyze your data and tell you what you need to do (keywords, etc) to increase visitors, etc.
Measurements are collected both on the site and user level (visitor and blogger, content owner, for example). I see the possibility of mapping any form (and fields of a form) into Nuconomy.
Nuconomy also has a recommendation engine that is built on top of the analytics but it’s use is optional.
Nuconomy is also built to monitor Social Networks and is able to generate vector maps backed on activities from a specific page (spot) in a Social Network (vector analysis against actions taken from a specific action point).
Basic takeaways from this interview with Todd Parsons is Buzzlogic is a valuable platform to determine influentials in a conversation and the most influential posts or pages of a website for any particular set of keywords.
Functionality was added last fall to key in Google AdWords and several other advertising platforms to run ads against specific properties that Buzzlogic determined were the most influential for those keywords.
It’s also likely that Buzzlogic can target specific posts and run ads against that post (determined most influential).
Once Influentials are indentified, there’s a workflow process that allows users to document and qualify the value of increasing social media traffic from a specific source – it’s more of an annotation service at this point but could be more, down the line.
Finally, while Buzzlogic determines influentials – it does not presently track what happens on a site once the traffic arrives. Work with ComScore and Nielsen is underway to develop the other side of the equation – metrics for onsite behavior that will be merged with the Influentials and Conversation Tracking that Buzzlogic excels in.
I was very interested in what I saw that Baynote Systems did for Expedia.com (it hasn’t been announced formally yet) which jumped conversions up by over 30%.
The results for Expedia were extraordinary (based on what I saw), but I think the same basic approach, with a few changes, could work on many sites – as per the http://baynote.com/tour/ with the “Intent Driven” actions (http://baynote.com/tour/intent-driven.php) that are missing on most sites.
The work to integrate Baynote appears simple and straightforward (http://baynote.com/tour/integration.php) and the use of “implicit” recommendation a powerful combination with the data sites already have on their customers (http://baynote.com/technology/).
http://www.webmetricsguru.com/2008/05/buzzlogic_1.html Buzzlogic Interview
http://www.webmetricsguru.com/2008/05/nuconomy.html Nuconomy Interview
http://twemes.com/emetrics (channel for Twitter Emetrics tweets)