Analyze optimize (experiments) data

[anuveyatsu]

As we have setup A/B tests for Download page conversion on 9 May 2018, we want to analyse results and decide on which actions should be taken.

Acceptance criteria

• [x] we have outcomes from the experiments
• [x] we have defined next actions

Tasks

• [x] by the end of the sprint analyse the reports
• [x] come up with decisions and next steps

[anuveyatsu]

FIXED:

• Search page - banner telling users about CLI tool (based on 799 sessions)
  • objective: increase number of visits of /download page
  • result: variant with the banner had 3% to 249% of improvement
  • outcome: added banner - https://datahub.io/search
• Download page - accent on "fetching data" rather than on "publishing" (based on 399 sessions)
  • objective: increase total downloads of the cli tool
  • result: new variant had -66% to 0% of improvement
  • outcome: keep old version
• Core datasets page - button for "fetching data" (based on 4406 sessions)
  • objective: increase number of visits of /download page
  • result: no clear winner found, new variant had -49% to 92% improvement
  • outcome: keep old version
• Pricing page - message about the "CLI" tool: still running and optimize suggests keep running it