0.25
0.5
0.75
1.25
1.5
1.75
2
Learning from our learners: harnessing explicit and implicit feedback to boost engagement and improvement
Published on Apr 08, 2019144 Views
Related categories
Chapter list
Learning from our learners: harnessing explicit and implicit feedback to boost engagement and improvement00:00
EnetCollect00:18
Learning from our learners00:27
Bare bones of Cambridge English Write and Improve02:56
Cambridge English Write and Improve updates06:28
How sticky are our learners?07:32
Since soft-launching on 5th September 201607:38
Registered account growth is highly consistent08:02
The active base grows more slowly08:17
The active base is constantly evolving: Part 108:55
The active base is constantly evolving: Part 209:36
W&I base is heterogeneous09:58
I speak your weight!10:59
Crowdsourcing feature ideas and UX insights 111:49
1.1m-word Corpus of Review Essays12:06
What do our users say about Write & Improve? 12:23
Corpus analysis for insights into Learner eXperience13:19
What they like (in approx. order of priority)13:34
Quicker than a teacher14:17
What they don’t like (in no particular order)14:34
It’s just a robot15:47
Responses to review feedback16:14
Just for fun!16:52
Reasons for cancellation17:34
Cancellation Feedback17:44
Coming back: Part 118:04
Coming back: Part 218:26
Responses to churn feedback19:04
Support emails19:42
Support email questions and requests19:50
Using support emails to inform FAQs20:26
FAQs21:08
Top 10 FAQs21:11
Responses to FAQ feedback22:03
Jobs to be done!22:39
Crowdsourcing feature ideas and UX insights 223:07
Implicit feedback23:26
Primary methods24:26
Over 100 event types captured25:02
A/B testing25:49
Probability distribution quantifies uncertainty26:39
Comparison quantifies confidence in difference27:22
Bandit testing exploits expected payoff27:58
Bandit testing examples28:24
Choose your level prompt28:47
Learner level response survival curve29:43
Bandit testing criteria30:40
Beware the Golem31:58
Proportion "spam" submissions 201733:19
Proportion of per-user spam: Part 133:55
Proportion of per-user spam: Part 234:32
Type A examples34:36
Type B examples34:51
Survivorship bias35:22
Highlighting prompt relevance35:59
Primary methods36:32
Our learners tend to improve37:02
More writing sessions, greater improvement38:07
Progress reporting email38:40
Progress reporting improves progress38:51
Onboarding emails provide inspiration39:16
Reactivation emails provide motivation39:35
Email programme metrics are promising39:55
Q&A40:17