ForumSG

The Opportunity

Problem: scalable product operations

Product operation teams spend a significant amount of time providing quality support. Especially for OGP platform products, admins often write in with complex, high-context questions which require time and effort to understand their particular use-case and craft an accurate response. Furthermore, existing support channels do not keep a persistent, searchable record of solutions, which can limit the multiplicable value of crafted solutions and also result in repetitive, recurring similar questions asked over time. This ultimately does not encourage scalable operations support that is important for lean product teams to operate efficiently.

Solution: The ForumSG experiment

ForumSG is an experimental product aiming to solve the problem of scalable operations support by providing a community space where public officers can find, share and discuss solutions.

What will the forum try to solve?

1. Offload product ops support by encourage peer-to-peer support

2. Keep an organized, searchable history of solutions to reduce repetitive contextual questions

The hypothesis is that a forum will reduce the effort product ops spend crafting repetitive or context-heavy solutions and reduce the time it takes for users to receive answers to their questions, and provide an opportunity for the community to help each other.

The Velocity

ForumSG serves as a community-based discussion platform organised around communities and conversations. Users can join communities, post conversations (questions or discussions), and subsequently engage in conversations through likes and replies. Conversations are searchable within communities (through a search bar) and can be marked as ‘Answered’ to easily inform users that a verified solution exists to the posed problem.

Our initial release to market started with 1 community (FormSG x MOE), aligned with FormSG’s drive to digitalize school paper processes where they experienced these pain points of handling repetitive high-context questions across digitizing similar workflows like bus booking and incident reporting. This also took into consideration the capacity of the team not to overstretch by supporting multiple communities and contexts.

The team had a meaningful time figuring out certain parameters for the product, making decisions through a calculated balance of assumptions and validations through conducting a total of 9 user interviews and usability tests.

Anonymity

Should we show user emails in conversations and replies?

UTs and feedback garnered mixed responses. On one hand, people felt less hesitant to make posts under the cover of anonymity. However, anonymity would increase the barrier if someone wanted to respond or follow-up directly to another person’s reply.

Upvotes and/or downvotes

The team decided to only allow the action of upvoting, since knowing that a post wasn’t helpful would be deduced from having a lack of upvotes. However, initial feedback was negative - users were confused with the purpose of upvotes and its implications.

Iterative research and development

User journey - A search bar

We learnt that users would come onto ForumSG either:

1. They have a specific question they want answered

2. They are trying to figure out how to do something, or if something is allowed.

This led us to priortise building a search bar to support this primary flow. We also observed how people would search for questions in the search bar if they typed using keywords or natural-language sentences.

ForumSG entrypoint + SSO

We decided early on that an entrypoint into ForumSG was on FormSG (a user having trouble building a form would look for the forum there). Since ForumSG required authentication so as to keep a BE record of user information for posts, single sign-on (SSO) was a potential opportunity to ease the friction navigating between FormSG and ForumSG.

Notifications

Another key concern users surfaced were how easily and quickly would they know if their questions were answered. We determined that it was crucial to implement ‘push’-based actions to inform the users of these updates, rather than relying on ‘pull’-based actions (requiring them to stay on the platform and track the conversation updates themselves).

Propagation

Our propagation strategy consisted of a total of:

The Traction

ForumSG received decent traction over iterative propagation. Although a significant number of users logged in, user activity was not frequent, alluding to most users being passive consumers of the platform.

* where user activity = no. of organic posts + no. of organic replies

ForumSG did however receive positive feedback about its potential and interest in the community, which was encouraging.

What's Next?

• Introduce the concept of roles and verified users - community experts that can serve as moderators, determine the validity of solutions, drive active membership, conversations and engagements.

• Expand the types of communities:

  • Onboard other products with active communities to have create their own forums (maybe plumber, isomer, pair etc.)

  • Explore agency-based communities on top of product-based communities, and introduce the concept of private & public communities

• Introduce global, community-agnostic search

The Team