Navigation path

Top level architecture EC

In line with our approach of following the evidence and involving our stakeholders (both the DGs and the people using our websites), we used the following process to create the top-level classification:

  • develop a hypothetical classification (using a "card sort" to get insights into the ways in which our audiences structure and group information).
  • test the hypothesis (using a "treejack" test to gather evidence on where people would look to find specific information)
  • refine the hypothesis (2-3 times) until our testers achieve an 80-90% success rate.

Between September 2014 and February 2015, we carried out 6 test iterations, with people representative of our typical audience profiles (as evidenced by the top-task poll). A total of over 6000 test volunteers were invited to take part.
Each test iteration was followed by a collective review meeting with colleagues from the DGs (a total of 5 plenary meetings) where we presented the results and our analysis and took decisions before proceeding.
We achieved success after 4 iterations.


Full details on building the top level information architecture for the EC


Presentations used in workshops with DG stakeholders

12 February 2015Agreement on level 1 of the common architecturePresentationpdf(5.49 Mb)
13 January 2015Presenting results of the treejack test and agreeing hypothesis #3 for testingPresentationpdf(8.76 Mb)
4 December 2014Presenting results of the treejack test and agreeing hypothesis #2 for testingPresentationpdf(4.3 Mb)
14 October 2014Common architecture building, preparing hypothesis #1 for treejack testingPresentationpdf(1.35 Mb)
9 September 2014Common architecture building kick-off & workshop, preparing for online card sortPresentationppt(2.83 Mb)