How can we learn about what s best for our citizens and customers? Toby Blume Ex-London Borough of Lambeth Now Social Engine

Save this PDF as:
 WORD  PNG  TXT  JPG

Size: px
Start display at page:

Download "How can we learn about what s best for our citizens and customers? Toby Blume Ex-London Borough of Lambeth Now Social Engine"

Transcription

1 How can we learn about what s best for our citizens and customers? Toby Blume Ex-London Borough of Lambeth Now Social Engine

2 Today Using Behavioural Insights Randomised Controlled Trials Putting it all into practice Key learning points

3 Using Behavioural Insights Randomised Controlled Trials Putting it all into practice Key learning points

4

5

6 The neural tug of war

7 How do we choose?

8 Messenger Incentives Norms Defaults Salience Priming Affect Commitments Ego

9 Using Behavioural Insights Randomised Controlled Trials Putting it all into practice Key learning points

10 How standard evaluation works People People Results Baseline measurement Intervention Time Post- intervention measurement

11 What s really happening? External environment Intervention Outcome Other agency actions

12 What s different about an RCT?

13

14 Using Behavioural Insights Randomised Controlled Trials Putting it all into practice Key learning points

15 We ran a series of trials to: Use behavioural insights to change behaviour Encourage new ways of working Embed the use of RCTs and evidence-based decision making Increase revenue and reduce costs

16 Putting it into practice 1: Nudging residents to increase Council Tax collection

17 Do small changes to Council Tax bills affect payment levels? Three treatments tested:- Simplification - to encourage compliance. Design based on eye-tracking research Social norm people strongly influenced by what those around them are doing Combined both simplification and social norm in one treatment

18 Simplification - eye-tracking research

19 Running the experiment

20 Results Simplifying Council Tax bills significantly increased the proportion of people paying by 4 percentage points The effect was consistent across ward, deprivation and council tax band Social Norm did not affect payment levels although there was variation among council tax band

21 Results not paid 70% paid 60% 56.68% 57.63% 52.91% 52.39% proportion of sample 50% 40% 30% 43.32% 47.09% 42.37% 47.61% 20% 10% 0% Control (no treatment) Simplification Social Norm Simplification + social norm Treatment group

22 Putting it into practice 2: Encouraging channel shift in Blue Badge renewals

23

24 Trial design Can we nudge residents to renew their Blue Badge permits online rather than by post channel shift Three types of message were tested for 5,817 Blue Badge permit holders over a period of four months (Dec 2014-Mar 2015)

25 We tested: Simplification Treatments Messenger a testimonial from another Blue Badge holder of how they have renewed online and encouraged others to do the same. Incentive - appealing to common good

26 Incentive letter

27 Results Simplification reduced paper renewals by 8.6 percentage points An intrinsic incentive reduced paper renewals by 7.7 percentage points. The use of a peer messenger did not appear to have an effect These results, if they continue, would translate into cost savings for the Council equivalent to around 20,000 a year, contributing towards existing savings targets.

28 Blue Badge RCT treatment effect on renewal 100% 90% 80% % 60% 50% % 30% 20% 10% % Control Simplified Incentive Messenger Online Paper Not renewed

29 Putting it into practice 3: Personalising text messages to encourage tax compliance

30

31 Can personalised text messages encourage people to pay their Council Tax? We tested :- Amount of council tax owed Name of the account holder Combined name and amount

32 100% Response by group 90% 80% 70% 51.4% 50.8% 51.4% 60% 50% 40% 3.7% 3.8% 5.0% 30% 20% 44.9% 45.4% 43.6% 10% 0% Control (amount only) Combined (name + amount) Name only Paid Part paid Unpaid

33 70% Proportion paid by month with linear trend lines 65% 60% 55% 50% 45% 40% 35% 30% 25% Control name Combined Linear (Control) 20% Aug Sept Oct Nov Dec Jan Feb Mar Apr

34 Results We found no statistically significant difference in encouraging people to pay But Name-only treatment encouraged nonpayers to part pay The effectiveness of (all) text messages degrades over time. Using them in short bursts works best

35 Using Behavioural Insights Randomised Controlled Trials Putting it all into practice Key learning points

36 Simple RCTs offer a simple, reliable and affordable way to measure outcomes and do things better Small things can make a big difference Focus on business as usual to make it as effective as possible going with the grain

37 Tactical Target low hanging fruit for a positive and immediate impact Tangible financial returns (c. 1m plus reduced costs) Reassure Members and managers

38 Scaleable Potential for adoption and adaptation: orevenue and rent collection odemand management ochannel shift owaste management ointernal systems and procedures

39 A culture of experimentation Constant improvement test, learn, adapt Delivering tangible outcomes no more unsubstantiated assumptions Managed risk doing things differently whilst ensuring accountability and VfM

40 Toby Blume