Qubit meta-analysis on website optimization

via Tristan Handy’s newsletter, a meta-analysis on website optimization [PDF] published by Qubit. Some interesting takeaways for product teams.

  • Cosmetic changes are usually immaterial.
    • “We recommend choosing a design and sticking with it based on preference or through a qualitative process.”
  • Most effective experiments affect perceived value of the product or service.
    • “The biggest winners from our analysis all have grounding in behavioural psychology – scarcity, social proof, urgency, and to a lesser extent, abandonment recovery.”
  • Poor implementation can distort results.
  • Not all personalization is created equally, but segmentation helps.
    • “By analysing the expected effect of segmented treatments vs non-segmented treatments we observe a nearly 3 fold difference in expected uplift in revenue.”
    • “Within this analysis we include treatments that can be automated based on visitor context and on-site behaviour : abandonment, product recommendations, social proof, scarcity, weather and urgency. At Qubit we call these ‘Programmatic experiences’. From this analysis it is clear that these are not equally effective.”
Posted Jun. 25 2017, 8:29 am