Remesh – 2019

Waiting Room

waiting-6.png
 

Increase participant engagement and provide moderators with the demographic balance for their study to be successful

Role

Senior Product Designer ― research, design, and testing‍

Team

James Ong - Senior Product Manager

Dan Reich - Lead Engineer

 

Problem

Low participation rates, opaque process, and unbalanced conversations

 

After tracking the participation data on Remesh, we realized that there is a high number of drop-offs after joining the conversation - participants who join but never engage with the conversation. This hurts the moderator's experience and our relationship with the clients. Also, moderators are unaware of how participants access the conversation and when which leads to more questions to our customer success team and visits to our help documents.

Publishing conversations

Publish conversation becomes the main action users need to make in order to give access to participants and to start the conversation, as opposed to auto-starting the conversation based on time.

Publish also serves as the save action that the date, time, title of the conversation are set, and to change they need to Unpublish.

publish.png
 

Participants waiting room

Before the start time of the conversation, all participants go through the waiting room first. Adding this screen and the Join Conversation made sure that only active participants at the start time are the ones joining so we do not save a seat for passive/idle participants.

waiting-2.png
 

Clear state handling for participants

Give the participants a clear idea of the different states of the conversation: Unpublished, Published, Started, Full Capacity.

waiting-3.png
 

Quota ratios on demographic filters

Give the participants a clear idea of the different states of the conversation: Unpublished, Published, Started, Full Capacity.

quota.png

Research

Analyze the drop off in participation

Our data showed that our first come first serve flow was causing the low participation numbers because we were saving seats for participants who become idle or passive after answering a couple of onboarding polls.

In fact, looking at the data we found that the participants who are arriving closer to the start time of the conversation are more likely to participate in the entire conversation and answer all questions asked by the moderator. Our previous workflow was benefiting and preferring the early participant but they were the most likely to leave while waiting for the conversation to start.

Understand the old UX pitfalls

First, I evaluated the Remesh legacy platform, focused specifically on the transition between Build to Live experience. It was clear that users were not sure what to do once they have created and set up their conversations. The issues they encountered are:

  1. What do participants see once they click on the participant link

  2. Does the link work?

  3. What do I do now? What’s next?

  4. Until when can I keep changing the information of the conversation ie. date, time, onboarding polls?

What is Quota Management? And how does it work?

Quota management and balancing the demographics of respondents is a feature that is only attempted and solved in quantitative asynchronous tools, it aims at giving researchers a way to target specific groups for their studies and to make sure that their work is statistically significant.

When it comes to asynchronous research, it’s easy to only collect the data from the participants that fit the specific group you need as a user or decide based on screeners if a participant will get the full experience or not. The big question was how to control quotas in the real-time Remesh conversation while participants keep joining.

Solution

Introduce explicit actions by the users and build a quota management system

We introduced two actions to be taken by both moderators and participants: Publish Conversation and Join Conversation. This made expectations and flow very transparent between moderators and participants.

We also built a way for moderators to associate ratios to demographics and an algorithm that admits and reject participants based on the demographic ratios.


Outcome

In 6 months from the rollout, we increased the participation rate and engagement by 73% – and quota management helped our client get the perfect balance for their sample.

 
 

Next Steps

We are exploring quota renormalization in the case where the participants do not show up and thinking of more ways to increase engagement