Tribe

Behind the scenes of building: Tribe

TL;DR

Tribe was the beginning of my startup journey. I left my job as an M&A Associate in investment banking to start a company with my friend who taught himself how to code. We set out to solve a pain point we personally felt: meeting new groups in NYC.

​We tend to socialize in small, tight-knit groups and meet others serendipitously for the first time in person. It's an inefficient process and often results in short-lived relationships. Existing social and dating apps fell short because they were designed around 1:1 interactions. We built Tribe to empower groups to make meaningful connections based on mutual interests and intent. The platform enables friends to create a group, check into events, see other groups attending, and connect ahead of time.

​Adopting lean startup principles, we launched our MVP on iPhone in December 2019. Within a few months, we grew to 1k users and connected over 250 groups. We acquired customers organically through word of mouth and viral loops built into the product. Unfortunately, COVID hit NY in early March and we decided to put Tribe on hold.

Problem

As a new college grad, I moved to New York City to start work. Every weekend, my friends and I would go to bars, concerts, and apartments of mutual friends looking to meet new groups. We wanted to forge friendships or find partners.

​The process was analog and ran mostly over SMS. We would show up to a place without knowing who else is going to be there. After a few high level conversations over the course of the night, it's over - come back next week, rinse, and repeat.

We tend to socialize in small, tight-knit groups. Yet, there is no platform to facilitate meaningful connections between two friend groups.

Is anyone solving this problem?

With a clearly defined problem in mind, we looked into existing solutions. Most platforms in the market fell into three segments: 1) dating / social; 2) private group messengers; and 3) event curation & ticketing.​After an exhaustive search, we came up empty. Dating apps were exclusively formal, 1:1 connections. Within event curation, Meetup enabled group connections, but was tailored to large, open forums. Meanwhile, social messenger apps like Instagram helped with discovery and planning, but did not facilitate meeting others. 

Another important step was to look backwards in time to see if anyone had tried to solve this problem, but failed. Two platforms came up in our search: Tinder Social and Grouper. Both had been discontinued so we researched them extensively and talked to old users to understand what went wrong. 

Get out of the building

Before thinking of a solution, we needed to talk to more folks like us. Our goal was to better understand the problem from first principles and confirm that the challenge was broadly felt beyond our social circle. I created a call script and sourced 1:1 interviews with friends and second degree connections. Within a couple of weeks we had completed 50+ calls / meetings.

We confirmed our problem was broadly felt. There was a clear desire to connect with other friend groups, yet many struggled to do so. Word of mouth was the main source of finding things to do and making meaningful connections was left to chance.

Ideation

With clear signals from customers, we knew this was a real problem worth tackling. I used the Jobs to Be Done framework to guide our ideation process. We listed out "jobs" that customers had and brainstormed solutions to accomplish these tasks. For example, people struggled to make meaningful connections with other groups because they had to randomly sample whoever was in the same bar, apartment, etc. Without background information, deeper conversation rarely took place.

After reflecting on our top solution candidates, we had consensus on what to build for the MVP: an app where friends can create a group, check into events, see other groups attending, and chat ahead of time.

Scope

Next, I scoped out product requirements based on our ideation sessions using the user stories framework: As a TYPE OF USER, in order to ACCOMPLISH A GOAL, I want ABILITY. 5 User Types emerged including the Organizer, the Tag Along, the Partner Seeker, the Friend Seeker, and the Experience Seeker. An example user story is "Organizers wanted to discover new bars in order to make plans with friends."

Once User Types and User Stories were in place, ​I developed a detailed product roadmap. Each feature would tie back to a "Job to be Done" and include a description, priority, and ranking based on impact vs. effort. After input from my co-founder and a few iterations, we had a list of high priority features to focus on for the MVP. Tribe was ready to move onto the design phase.

Growth loops

Andrew Chen has had a big impact on my startup thought process. He co-wrote a blog post about growth loops, which are "closed systems where the inputs through some process generates more of an output that can be reinvested in the input." These are productized engines that use engagement to drive compounded user growth from one cohort to the next.

For Tribe, I developed an organic growth cycle where users were incentivized to attend events in order to receive inbound requests to chat from other groups. To attend an event, users needed to be in a group of their own, which led to sending invites to friends. To compound this, a percent of those groups who checked into events could share their plans to Instagram from within Tribe. These posts on social media would bring new cohorts to step 1 of the loop.

Wireframe

Now it was time to bring these concepts to life. My preference is to use pencil and paper in the earliest stages of design. This approach maximizes speed and minimizes creative block.​

One of the first challenges to think through was the classic "chicken and the egg" problem. If a new user signed up, but didn't see other groups, they would churn. My solution was to bring event / location discovery to the forefront of the experience. We could curate popular bars and events in the NYC area and encourage groups to check in if they were interested in going. This would help users plan their weekend and get value from Tribe on day one, even with a small number of groups on the platform.​

I sketched out key screens and shared them with my co-founder. The first tab is a map interface called "search" showing nearby bars, events, and neighborhoods for groups to check in. The second tab, "popular", lists curated locations sorted by number of groups attending with filters for group gender and date. Users can see profiles of groups attending each location and request to chat. Successful matches would show up in the third tab, "chat", where each group could get to know each other and coordinate plans.

Design and test

In Figma, I created detailed mockups of the user interface with supporting user flows. Before Tribe, going out was a game of chance. People showed up to bars and events hoping to strike a conversation with someone with mutual interests. Now, groups could connect ahead of meeting IRL and develop more meaningful relationships.

With a clickthrough prototype, we were able to get an early version of Tribe into the hands of 15 testers for usability testing and feedback. The response was overwhelmingly positive.​

One area of critique was a concern about the safety of meeting other groups in private places (ie. apartments). There was a comfort in meeting in public places like bars or restaurants as well as added safety of being with friends at all times. To address this concern, we removed the ability for users to add private locations as places to meet. Instead, groups could attend general neighborhoods and exchange addresses within the group chat once the relationship was more developed.

Build

With our design file updated, my co-founder began engineering sprints. He used Swift for our front-end programming language and Google Firebase for our back-end realtime NoSQL database. We used a kanban board on Trello to track progress on the product roadmap and highlight blockers. As features moved into review status, I performed QA testing to check for bugs and ensure the user experience was flowing appropriately. 

Private beta

Our MVP was built and ready for beta testing on TestFlight. We gave 30 of our interviewees key tasks such as creating an account, making a group, attending an event, and chatting with other groups on the platform. We received A LOT of valuable feedback and Tribe seemed to be solving the user's problem. One of the test groups even hit it off and continued chatting after the beta. Our biggest change from the test was to onboarding. We reduced the profile upload requirement from 4 images to 1 image and cut a number of questions. This resulted in profiles being completed 3x faster and removed a large bottleneck for completing the onboarding funnel.

Launch!

Throughout the development process we had created a small waitlist, which now had ~250 signups generated from word of mouth. In early December, we put Tribe on the Apple App Store and announced our launch party. To get entry, users had to show they attended our event on the app. We chose to launch with a physical event because it would help bootstrap our user base and prove the use case.​The launch party was a huge success with over 300 people attending. It was amazing to see groups of friends that were once strangers making new connections because of our platform.

Analytics

After our launch, I used Google Analytics to monitor user trends and behavior. While planning the product roadmap, I defined several key events to track engagement such as the onboarding funnel, creating a Tribe (personal group), attending a location, requesting to chat with another group, and messages sent to other groups. To establish north star metrics, I defined two OKRs: 7% weekly growth for 1) weekly active users and 2) messages sent to other groups. Making a connection and talking to new people was the "magical moment" for users, which we wanted to optimize for. ​

Although WAUs looked promising with 10% growth in week 3 post-launch, our key engagement activity of messages sent was down / stalling at ~3%.

To dig into the issue, I ran a cohort analysis to monitor retention and activation. The data showed that new customers were churning before ever connecting with a new group and sending a message. We needed to talk to users to further explore the cause of our leaky bucket. 

New insights

I reached out to both active and recently churned users to conduct 1:1 interviews. Our goal was to gain context around why they left the platform and what issues they were running into. Here were the key takeaways:​

1. Although users appreciated the list of curated bars and events, they were hesitant to check into events on the app because plans were too fluid

2. Common interests between groups was very important to users, but this information was hidden behind individual profiles

3. Meeting groups was the primary objective, but finding them on the app felt cumbersome because locations were front and center

Iteration

Sticking to our lean startup principles, my co-founder and I revisited our Jobs to Be Done framework and iterated on the MVP. Based on the qualitative and quantitative data, we shifted away from event / activity discovery within the app and doubled down on showcasing groups with mutual interests. Within three weeks, we had designed, developed, and launched our updated app. 

Events -> interests

We removed the first tab and map interface as we pivoted away from event discovery. This was replaced with a dashboard where users can manage members of their Tribe, set their location, and indicate group interests.

Prioritize groups within UX

Next, we brought groups, profiles, and interests front and center to make finding relevant connections easy. At this stage, we had enough groups on the platform that we were less concerned about our earlier design challenge of users churning due to a lack of groups. We also removed the popular events tab to streamline navigation and replaced it with an email newsletter to provide event discovery off-platform.

New growth loops

With events removed from the core experience, our growth flywheel needed to be revisited. In its place, I created a token system that rewarded engagement and incentivized virality. Each day a user gets 100 tokens. It costs 10 tokens to request to chat with other groups. To get more tokens, and therefore more chances to connect with groups, users can complete certain tasks. Engagement tasks included adding more photos, daily logins, and making 5 new connections. Growth tasks included inviting friends, sharing Tribe to instagram, and rating Tribe in the app store.

Positive results

A few weeks into our product update, we looked at the data and gathered feedback from customers. Our OKRs were improving with WAUs and messages sent growing to 625 and ~350, respectively. New cohorts were also showing improved retention.

We were excited about the future of Tribe and the prospect of helping people to make new and meaningful connections. I published a blog piece sharing our vision of the future and why we were building the platform.

A global pandemic

In the first weeks of March 2020, COVID 19 swept through NYC. The impact on social life was swift with millions forced to stay at home. We had to react fast to this changing landscape. One option was to shift to a completely virtual platform removing any health risks of meeting IRL. However, we felt that groups getting together in person was critical to the value proposition of Tribe and therefore would not be sustainable. Ultimately, we made a tough decision and put Tribe on hold until there was more certainty around the pandemic. In the meantime, my co-founder and I had a few other ideas and projects to explore that were more apocalypse-resistant.

Learnings

A) Do things that don't scale

I have found that the most effective method for bootstrapping a community is to do things that don't scale. For Tribe, this included holding in-person events, pitching the app to strangers at bars, handing out flyers on the corner near universities, and more.

B) Network effects are powerful

A minimum threshold of active users are required for network effects to take place in a social app. New users that joined Tribe would churn if there were not enough groups to connect with each week. In retrospect, I would have liked to spend more time in private beta seeding a robust community before launching in the App Store.

C) Team is everything

As a first time founder, I'm proud of the product my co-founder and I built. For me, the experience reinforced the importance of partnership and team. Open communication, trust, and a lack of ego stood out as core competencies required for success.