Day 2 Stream Sessions – The Fundamentals

The fundamental elements that are critical in developing behavioural insights trials are the three D’s: Data, Design, and Delivery. Select one stream to hear three experienced corporate leaders and academics talk about their experience and specific case studies. 

Stream 1 - Data

Mr Adam Welsh, Senior Director of Regulatory Affairs APAC, Opower

  • View Adam Welsh's presentation (PDF)
  • Mr Welsh covered how behavioural insights can be harnessed to generate real energy savings. 
  • Reaching the residential energy sector has traditionally been considered too difficult, with energy providers instead targeting industrial users to save energy. Behavioural insights around normative comparisons, combined with big data and cheap channels of communications has energy efficiency to be achieved at scale in the residential sector.
  • When Opower has communicated energy usage data which includes behavioural insights, other behaviour changes such as:
    • higher participation in other energy efficiency programs 
    • people opting in to time of use tariffs – 60 to 70% higher increase where they are available
    • higher usage rates of energy efficient appliances and home energy management systems
  • Opower has undertaken significant A/B testing to perfect its approach, right down to the impact of saying whether a comparison is between households or neighbours. Literal comparisons (between neighbours) do result in people saving more energy but increases their annoyance.
  • The effects from normative comparisons tend to have a strong persistence as people are continually ‘nudged’ with new information.
  • The Opower example shows how a business model can be built which has positive outcomes for society as a whole but also is a profitable company.

Professor Robert Slonim, University of Sydney

  • View Robert Slonim's presentation (PDF)
  • Professor Slonim spoke about how behavioural insights can be used to increase blood donations
  • The context in which incentives are used appears to be the most important factor in whether they are successful. In an example cited by Professor Slonim on the American Red Cross, blood donors were given a given card to use at local merchants in exchange for their donation. This saw a sustained increase in donations with no noticeable effect on safety of the blood supply.
  • Professor Slonim discussed reciprocity and a study which tested whether a gift (a pen) from the Red Cross would motivate people to donate blood. While the gift did triple to the donation rate (from 0.5 to 1.4%) it was not clear whether this is cost effective given the cost of the pens relative to donations. 

Dr James Guszcza, Senior Fellow, Deloitte Analytics Institute

  • View James Guszcza's presentation (PDF)
  • Dr Guszcza spoke about how data can be leveraged to unlock value and enable people to make better decisions.
  • Predictive models can be used to create ‘mindtools’: eye glasses for the mind, which cut through the clutter of cognitive biases.
  • There is a last mile problem for predictive models: going from the model to taking action. Once the models point you in the right direction, what do you do? That’s where behavioural insights come in.
  • Predictive models and behavioural insights are two parts of a greater whole –the real end game is to change behaviour. 

Facilitated by Stephen Brady, Deputy Secretary, NSW Department of Premier and Cabinet