I recently sat on a panel at the London eMetrics summit. Two panels actually, one on the first day and one on the second.The first panel was a good session where we discussed various topics around web analytics. Issues that people were having discussed openly between myself, Oliver Schiffers and Aurelie Pols.
Jim Sterne had asked us to sit on a panel at the end of the last day to tell everyone what we learned. In preparation for this I made some notes, hence this blog post.
I learned quite a bit. First from my peers, other consultants in the field that we’re presenting educational material and their latest concepts.
I really like the phrase “turning the lights on” from LoganTod, a process Matthew Tod has developed internally at his company to describe the process of enlightening their clients about the value of analytics data. I will steal that phrase, sorry Matthew but at least you got credit! Matthew also gets credit for the great response from one of his clients when receiving a report, “If there is no arse to kick then why is it in my report?” Beautifully put phrase I thought, meaning if there is no-one responsible for acting on the information then why bother reporting the finding!
Matthew also showed a great visualization in Excel that showed a way to present analytics data based on variance techniques.
The idea is that you move the filters on the right hand side and depending on what you moved the visualization changed and showed areas where you needed to act. Very insightful to be able to adjust the number of add to carts of a particular product and see how it compares to other products in a simple excel spreadsheet.
We’ve done similar things in the past in excel but the variance in percentage across x and y was a new idea that I’ll experiment with.
Another idea could be combined with turning the lights on above. Ask your salesperson to use your website page copy in his pitches. If he says no way then you have a problem. The idea being that your website should in effect be as simple as the way a salesperson discusses things with his prospects.
Then there were the tools. Some I’d never seen demo’d quite so well and some I’d never heard of that I’ve now investigated;
- Forsee results > I thought they were just an MVT tool but apparently “impact on staisfaction” is tied into analytics based on Survey data.
- Yahoo site explorer > A tool you have to login at Yahoo to see (which is why I’d never seen it before) but it shows more accurately how many links are posted in your direction. Useful for SEO purposes.
- FaceBook Lexicon > A tool to measure keywords used in peoples walls as an aggregated chart. Shows interesting trends and measures buzz on FaceBook. I didn’t know this was released by Facebook yet though I’d heard it was coming.
- Autonomy > A $4Bn company that deals with unstructured data (unstructured being data that doesn’t have a data model like something you can store in a database).
Finally there were plenty of snippets I picked up that I needed to research. For instance how easy would it be to copy what hotels.com have done with some of my clients. They have for instance a wiki that tracks global spend attribution, so if they had run a campaign 12 months ago they would be able to see why a spike in traffic existed. It also took them 2.5 years to get a good level of implementation and sophistication in their reporting interface (Omniture).
I also was forced to think about how to track things like OpenTable. A great idea. You browse your mobile and find the nearest available restaurant seat! Very cool but for ideas like this we need to track how people have engaged with the restaurants – a mobile/internet mash-up of measurement.
Mike Grehan also discussed cognituve search among other things. What he means is people learn by their searches whether they have found what they’re looking for. So you might start by typing “BMW” and then refine your search to be “BMW 116E with 5 doors”. That is a simplistic example but the idea is if we can find the first term and display the correct ad (in this case a BMW 116E 5 door ad when someone types BMW) then whomever does that is going to end up being a millionaire as they will be able to bid on relatively cheap SEM terms but convert relatively highly. The secret is knowing the 1st term people use. In my example “BMW” is most likely wrong as the first term someone would type. It might be something like 5 door hatchback instead. Mike discussed his thoughts on why Google hold this data from us marketers, but also suggested something like Hitwise might be a solution to finding out the first term.
He reckons Cognitive search results are the future of Google’s algorithms.
Then finally I heard that the long tail is actually very short according to a Harvard report. The argument is that blockbusters still stomp on the long tail and concluded that The importance of individual best sellers is not diminishing over time. It is growing….
You’d need to buy the Harvard report to find out how they came to that conclusion but it is quite convincing. So there you have it. What I learned this year from eMetrics. Hopefully I’ll be at Stockholm in October too learn more! I suggest you join me.