Hypothesis Driven Design - 17th May 2018

Slide 1

Slide 1 text

Hypothesis Driven Design 17th May 2018 - DWP Product Owner Community Day

Slide 2

Slide 2 text

Tom Morgan Service Designer @tsmz

Slide 3

Slide 3 text

Bruce Lee

Slide 4

Slide 4 text

Hypothesis Driven Design @tsmz

Slide 5

Slide 5 text

less finger more moon

Slide 6

Slide 6 text

why hypotheses?

Slide 7

Slide 7 text

Communication and visibility

Slide 8

Slide 8 text

Rigour, discipline and rhythm

Slide 9

Slide 9 text

Setting the right context for design

Slide 10

Slide 10 text

Recording design decisions and research findings

Slide 11

Slide 11 text

how?

Slide 12

Slide 12 text

Hypothesis Driven Design @tsmz The Scientific Method (look familiar?) https://en.wikipedia.org/wiki/Scientific_method

Slide 13

Slide 13 text

SCIENCE IS NOT AN EXACT SCIENCE

Slide 14

Slide 14 text

Observations are key

Slide 15

Slide 15 text

… so you need something to test

Slide 16

Slide 16 text

Start with user needs

Slide 17

Slide 17 text

writing hypotheses

Slide 18

Slide 18 text

Hypothesis Driven Design @tsmz We’ve seen… We think this is because… So if we… We’ll see... ahem!

Slide 19

Slide 19 text

Hypothesis Driven Design @tsmz We’ve seen… We think this is because… So if we… We’ll see... observation

Slide 20

Slide 20 text

Hypothesis Driven Design @tsmz User research isn’t about finding what users like, but what works best for them

Slide 21

Slide 21 text

Hypothesis Driven Design @tsmz We’ve seen… We think this is because… So if we… We’ll see... explanation

Slide 22

Slide 22 text

Explanation

Slide 23

Slide 23 text

Hypothesis Driven Design @tsmz We’ve seen… We think this is because… So if we… We’ll see... design

Slide 24

Slide 24 text

Time for design

Slide 25

Slide 25 text

Hypothesis Driven Design @tsmz We’ve seen… We think this is because… So if we… We’ll see... observation

Slide 26

Slide 26 text

Focus on what’s testable

Slide 27

Slide 27 text

Hypothesis Driven Design @tsmz

Slide 28

Slide 28 text

User research defines the rhythm

Slide 29

Slide 29 text

ceremonies discipline & rhythm

Slide 30

Slide 30 text

Hypothesis Driven Design @tsmz Affinity sort Collect and group observations What are the problems?

Slide 31

Slide 31 text

Hypothesis Driven Design @tsmz Then... Prioritise problems Design solutions

Slide 32

Slide 32 text

Hypothesis Driven Design @tsmz UR prep What’s ready to test? What are we looking to see?

Slide 33

Slide 33 text

Hypothesis Driven Design @tsmz Affinity sort What did we see? Anything ready to build? Everything else, back to start.

Slide 34

Slide 34 text

columns organisation & visibility

Slide 35

Slide 35 text

Hypothesis Driven Design @tsmz Backlog

Slide 36

Slide 36 text

Hypothesis Driven Design @tsmz Priorities

Slide 37

Slide 37 text

Hypothesis Driven Design @tsmz Being designed

Slide 38

Slide 38 text

Hypothesis Driven Design @tsmz Ready to test

Slide 39

Slide 39 text

Hypothesis Driven Design @tsmz Ready to build

Slide 40

Slide 40 text

No content

Slide 41

Slide 41 text

No content

Slide 42

Slide 42 text

No content

Slide 43

Slide 43 text

crossing the rubicon

Slide 44

Slide 44 text

A conversation

Slide 45

Slide 45 text

Hypothesis ≠ User story

Slide 46

Slide 46 text

A hypothesis might translate into multiple user stories

Slide 47

Slide 47 text

Business analysts

Slide 48

Slide 48 text

JIRA! built it!

Slide 49

Slide 49 text

what to record?

Slide 50

Slide 50 text

Hypothesis Driven Design @tsmz Title User needs Hypothesis Link to prototype pages Screengrabs & photos Research methods & findings Link to JIRA ticket(s) telling the story

Slide 51

Slide 51 text

What are we aiming for?

Slide 52

Slide 52 text

Hypothesis Driven Design @tsmz A complete picture How were design decisions made? What research was done? What was the outcome?

Slide 53

Slide 53 text

Hypothesis Driven Design @tsmz Audit trail Does the final build match the intention? How did the product end up like this? Useful for new starters / stakeholders alike.

Slide 54

Slide 54 text

Hypothesis Driven Design @tsmz Show and tells

Slide 55

Slide 55 text

Hypothesis Driven Design @tsmz GDS assessments

Slide 56

Slide 56 text

Storing knowledge

Slide 57

Slide 57 text

Examples

Slide 58

Slide 58 text

Check Your State Pension Steven Borthwick

Slide 59

Slide 59 text

Hypothesis Driven Design @tsmz Problem Not all users understand what the most State Pension they could get is. This results in enquires to the digital service and contact centre.

Slide 60

Slide 60 text

Hypothesis Driven Design @tsmz Hypothesis If messages are personal and direct, users will understand if they can increase their pension as a result enquiries should reduce.

Slide 61

Slide 61 text

Hypothesis Driven Design @tsmz Before

Slide 62

Slide 62 text

Hypothesis Driven Design @tsmz After

Slide 63

Slide 63 text

Hypothesis Driven Design @tsmz Measured ● Enquiries and comments submitted to service ● Satisfaction rate and survey ● Skype/screen shares with users ● Telephone calls to operations

Slide 64

Slide 64 text

Hypothesis Driven Design @tsmz Results Enquiries dropped by over 50% Satisfaction increased by 3% (90%+)

Slide 65

Slide 65 text

Budgeting Loans Richard Smith & Phil Swan

Slide 66

Slide 66 text

No content

Slide 67

Slide 67 text

No content

Slide 68

Slide 68 text

Hypothesis Driven Design @tsmz

Slide 69

Slide 69 text

Hypothesis Driven Design @tsmz

Slide 70

Slide 70 text

Hypothesis Driven Design @tsmz Even though it increased drop-outs, there were more eligible claims received” – Richard Smith “

Slide 71

Slide 71 text

Access To Work Rebekah Barry

Slide 72

Slide 72 text

Hypothesis Driven Design @tsmz

Slide 73

Slide 73 text

Hypothesis Driven Design @tsmz Errors on initial travel question 3.95% Loops back to initial question from next page 2.6% Users changing their mind 12.9%

Slide 74

Slide 74 text

Hypothesis Driven Design @tsmz We’ve seen: 15.4% of users either change the radio button they’ve selected or return to the page and change their answer and we’ve seen: validation errors increase on the initial help With Travel page from 0.24% to 3.95% We think this is because: users are struggling to understand the initial travel question and so they are; 1) changing their mind about what their answer should be 2) trying to continue without answering it 3) answering that they don't need help when they really do So if: We change the wording of the question, and add text to the 'yes' 'no' radio buttons to make it clearer Then we’ll see: the numbers of users changing their answer reduce, the number of users seeing validation errors (after travel) reduce.

Slide 75

Slide 75 text

Hypothesis Driven Design @tsmz “There are too many saccades and no fixations at the beginning of the sentence - the eye is only drawn to ‘work OK’ leading to regressive reading and confusion” – Rebekah Barry

Slide 76

Slide 76 text

Hypothesis Driven Design @tsmz

Slide 77

Slide 77 text

Hypothesis Driven Design @tsmz Errors on initial travel question 3.95% → 3.01% Loops back to initial question from next page 2.6% → 1.6% Users changing their mind 12.9% → 7.82%