Highlights of XChange Europe 2013 in Berlin

« 5 Questions on the next version of Adobe Analytics to Bill Ingram, VP for Adobe Social & Analytics | Main | Web Analytics Rendez Vous on Segmentation and Attribution featuring Matt Bailey (USA) on October 23rd, 2013 in Lille »

07/16/2013

Highlights of XChange Europe 2013 in Berlin

The second edition of XChange Europe took place with 110 participants from 15 different countries on June 10-12,2013 in Berlin. Xchange was founded in California 7 years ago by Gary Angel, former President of Semphonic and now Partner of Ernst & Young (Semphonic has been acquired by Ernst & Young in April 2013). Xchange is a very unique Web Analytics conference organized around “huddle” discussions with no Powerpoint presentations at all during 2 days. In the opening keynote, Gary explained that he decided to create XChange with this format, when he realized that most of the value he was getting from Web Analytics conferences was happening during peer discussions between Powerpoint presentations.

This year, the event tool place at the Intercontinental Hotel, near Tiergarten and the Berlin Zoo. I had a fantastic view on Berlin from my hotel room in the 12^th floor!

Xchange started by a welcome reception on Monday evening, a great opportunity to meet old friends and to make new connections!

On Tuesday, the day began with a panel discussion between Gary Angel (Partner at Ernst & Young), Christian Sauer (CEO of Webtrekk), John Woods (CEO of iJento), Mathieu Llorens (CEO of AT Internet) and Simon Burton (CEO of Celebrus). Mathieu Llorens started a debate by stating that American solutions have a strong focus on new customer acquisition measurement, while, in AT Internet’s case, the emphasis is more on measuring existing customer behavior. Simon Burton agreed that American solutions are more product-led, whole European solutions are more customer-led.

In terms of international expansion, Christian Sauer explained that WebTrekk has just opened an office in San Francisco, but the US market is quite a difficult one with large competitors and an existing partner network. Also, once you’re successful in the US, wait for the patent lawsuits. WebTrekk is also developing on the Chinese market, which may become interesting in 2/3 years. Since China has a 85% e-commerce growth rate at the moment, nobody cares about Web Analytics for now. What about the Web Analytics market in general? According to Christian Sauer, the market is not growing since the market leader, Adobe, is not growing regarding its Web Analytics suite of products. Google is responsible 90 % for market stagnation by offering a great and free product. According to him, third-party cookies will disappear over time with privacy laws and new browsers default configurations. So there is the need to build a strong first-party cookie data collection.

According to John Woods, the US market is a great growth opportunity for iJento, since their product is complimentary to other analytics suite like Adobe SiteCatalyst, with a great installed base in the US. John Woods is also optimistic about the future of analytics. While Google Analytics is more focused on reporting needs, there are still great opportunities in customer analytics and personalization. On top of this, there will still be needs for expertise to get the best out of this technology.

For AT Internet, the growth is driven by existing customer needs with for instance Brazil and Russia, said later Mathieu Llorens. According to him, Google Analytics Premium has brought a big change in the market since Google has educated the market with the data sampling issue. According to Mathieu, we don’t hear about data quality and reliability as much as we should hear. Also, nobody is going backwards when business has seen the value of data.

Simon Burton explained that Celebrus Technologies has based its US expansion on a partner network. Business is different in every region in the US and it is much better to have partners on the ground.

Gary Angel concluded the panel by saying that Google cannot fill up every niche, so there is still room for opportunities.

After this panel discussion, all participants were split in 8 "huddles", with 15 participants maximum for each one.

The topic of my first huddle was “From web analytics to mobile analytics and back” with Peter Pletsch, Meinestade.de, as leader. The main points from this discussion were as follow:

as analysts, we need to rethink our analysis framework and anticipate data collection needs, because gestures and interactions like mouse tracking are not being recorded for mobile devices.
call tracking does not work out of the box, and needs to be tagged specifically.
there is a big divide between Mobile Web sites (tracked by JavaScript) and Mobile applications (not tracked by JavaScript). Build tracking when you build the app. Don’t retrofit it.
Page names and page views are not really relevant for applications. SDKs are very influenced by Web behavior
Responsive design forces to rethink measurement on the web. Report on use cases instead of page views (same thinking with apps).
Tag Management Systems can really help mobile tracking.
all mobile traffic is not equal: tablet traffic is much more similar than desktop traffic. Tablets are usually used sitting in front of television.
differences in emerging markets: more iPhone users in Brazil because they are the ones who can afford a data plan.

After a networking lunch, a group picture was taken with of MeasureBowlers from various cities in Europe. The first European-wide MeasureBowling event took place a week before on June 6th.

My second huddle was on "Creating Strong Foundation for Analytical Success" with Ross Mc Donnell, Walt Disney Company, as leader. The discussion was very well organized around 6 main topics:

1. KPIs

Define a KPI framework (business goals -> measurement goals)
Driver KPIs / driver tree -> what can you influence
If well designed, these KPIs “waterfall” at every level, so that the corporate strategy is steadily expressed, mapped, and expanded in measurable terms for every person at every level of the organization.

2. Implementation QA & testing

Use ObservePoint or Selenium for quality assurance
Document all technical analytics implementations so you are not left confused when someone leaves the organisation
UAT for metrics on real-life data

3. Reporting & dashboarding

Build a web-reporting portal, provide everyone with access, put analytics on it & see who is looking at what
Dashboard TVs -> good tip to disseminate data within the organization
Trying to reduce number of dashboard and empower users to create them
Stop management analytics reports & start reporting to people that are actively/daily spending money to drive outcomes-thoughts

4. Insights & actions

Different layers of insights: 1rd layer: observation (analyst), 2nd layer: insight (category manager), 3rd layer: recommandations (UX experts, online marketing)
Analyst need to educate what could be the most impactful things that you could do
Let people think about the business impact perspective
Only makes sense if upper management buys into the approach

5. How to keep up with KPIs, collection, … as business moves

Go back a year later, look at the data and decide on what needs to be improved / anticipated

6. Training

Internal “Certified trainers” + wiki for FAQs + weekly webinars to get people up to speed
Basic introduction on web analytics but using introduction to clarify the role of specialists in the analysis
Identify key power users

My third huddle discussion was on "When the customer struggles in your website : how to diagnosis ad act using new customer focused analytical tools effectively" with Etienne Cox, Dell Corporation, as leader:

Starting with clickstream to narrow down issues -> content groups
Need to look at overall funnels -> go deeper with segmentation and microconversion-> struggle events
You can’t tell if users are struggling unless you know their intent
Voice of customer analysis to capture customer intent
Be careful with decisions based on heatmapping
Online chat for capturing user experience
Use forms analytics
Be careful of the context in which customers are conditioned to use or not use the web site
Gathering data on social networks about web site issuers -> to feed optimization process
Customer survey feedback is surprisingly not look at very often
Don't forget heuristic analysis!

The second day finished with a great networking party, as well as indoor golfing!

On Wednesday morning, my fourth huddle discussion was on "UX with numbers : How can you triangulate the numbers with human behavior" with Craig Sullivan, Rush Hair, as leader:

Not changing your homepage regularly is like keeping the same
What’s not working with your current approach regarding UX optimization? Ownership split between different teams.
How to save money with UX when budgets are getting tighter ? Remote user testing. Online panels are usually cheaper.
Use agile techniques -> decide earlier than making it perfect. Session replay. Eye tracking is still valid even if they are mouse tracking tools like Clicktale, Crazyegg, Sessioncam.
Be careful with computed eye tracking. The real value gets in analyzing the data with visitor intent, perception, persuasion and conversion.
Time to complete an event is a good indicator to start with, but it is not necessarily a guarantee of good experience experience.
Net Promoter Score -> drawbacks is that you need to get additional information on the service itself to really get value out of it.
Page Load Time, using tools like Compuware Gomez
Different feedback questions according to the page
Cultural differences according to countries with user feedback
Always new features on websites. No features are removed, so the user interface gets cluttered with too many links.

My fifth huddle discussion was on "SEO Analytics for Larger Websites" with Piotr Poznański, Eniro Poland, as leader:

Keywords clustering needs to be used to find patterns
Segment keywords into 3 groups : navigational keywords (domain name, brand name), informational keywords (how to do this), conversion keywords (specific product, buy)
Different search results if you use a mobile or desktop. -> could use segmentation

Unfortunately, I had to leave XChange earlier than planned on midday, because of an air traffic strike in France.

Many thanks and congratulations to the organizing team for a great event once again. Really looking forward to attending next year's event!

Comments