*August 2018*

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FDtJg9QGeS4O0KokUuaKz%2Fdd0ef68e-3dda-4e69-b067-bb3255683f7f.png?alt=media&token=5960593f-5931-45d8-a7e4-8320c829ffc5","id":"dd0ef68e-3dda-4e69-b067-bb3255683f7f","width":1730,"height":1382,"filename":"Screenshot 2023-04-26 at 9.48.22 AM.png","type":"image/png","caption":"Selective send (CC) and Presence","border":false}]}
```

**Summary**

⇾ In between Instagram, Facebook, Snapchat, redddit and Twitter, we saw an opportunity for a communication service that afforded users **reach** *and* **privacy**

⇾ Of all the features included in early builds, the “CC” message distribution model had the biggest impact on Jot’s trajectory

⇾ With usage, it became clear the **instant, conversational** nature of Jot was the most compelling part of the experience

⇾ Despite **strong retention** and a **deeply engaged** user base, we did not figure out how to grow Jot quickly enough to achieve scale

**The beginning**

Jot began with a series of questions: *Why do we find ourselves dropping the same link or thought into multiple Gchat and iMessage threads? Why is it not clear where the best place is to talk about what we’re thinking, as we’re thinking it?*

Yes, existing services offered broad reach, but it came at the expense of privacy. A trade off we found drove our sharing elsewhere, considering our sharing came from **wanting to talk** rather than wanting to spread a message.

**Product**

Of all the features included in Jot’s early builds, there was one that defined our course: **CC**. All persons attached – CC’ed – to a Jot were a part of the ensuing group conversation. Our most active users were drawn to the threaded conversational nature of Jot. It became clear the **interaction that followed the original post** was the most compelling part of the experience.

Peoples’ preference for chat became abundantly clear when we discovered the majority of our most active users were sharing each of their Jots with their **entire friend list**. Their goal, as we found out, was to maximize the exposure of their Jots to increase the likelihood that a conversation would ensue in that moment.

This insight around our most active users **seeking immediate interaction** led to Presence, Discover and Subscribe. In effect, Jot evolved into a service for **instantly chatting with like-minded people**, whether you knew them already or not.

**What changed?**

Because the types of topics being discussed on Jot varied so drastically, we found it **difficult to isolate potential users** and thus scale our acquisition efforts.

We thought making it explicit who Jot was for would make user acquisition more straightforward. To test our assumption, we introduced **Networks** and **Spaces**. **Networks** were **private, university-specific “instances”** distinct from “main” Jot for students to talk. We launched Networks at Emory University first, and then at UC Berkeley and SFSU. **Spaces were sub-communities** within Jot intended for conversation around a **particular topic or idea**.

At Emory, we managed to drive **10% of entire undergraduate population** onto Jot within **hours of launch**. But, this proved to be too many people too soon; the community was not able to sustain a “heartbeat” of its own. Learning from our experience at Emory, we rolled Jot out more gradually at Berkeley and SFSU. While communities at both campuses emerged and retained a core group of active users, they, like the original Jot community, did not grow as we needed them to.

Ultimately, we did not figure out how to grow Jot fast enough. While many users shared threads to Twitter and Facebook, these shares did not drive new signs ups at a high enough rate. In fact, most times, the conversations users would share **continued on the service the thread was being shared to**. Secondarily, **users did not invite their friends directly to Jot**. In large part – we think – because users were not incentivized to. Jot gave people instant access to spontaneous conversations, without needing to know anyone beforehand...how would having your friends on the service make the experience any better?

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FDtJg9QGeS4O0KokUuaKz%2Fa06c3fef-e109-494a-9cdb-16ee580b6c62.png?alt=media&token=486bc04b-9a43-4f03-a965-dde676f004eb","id":"a06c3fef-e109-494a-9cdb-16ee580b6c62","width":1726,"height":1376,"filename":"Screenshot 2023-04-26 at 9.49.07 AM.png","type":"image/png","caption":"Spaces and Networks on Jot","border":false}]}
```

**In the end**

We built a service that encouraged a small, but highly engaged community of people to flourish atop of it. We proved there was **value in a service focused on making starting and participating in conversations effortless**. What we did not figure out was the **right mix of incentives to achieve the viral growth** we needed to support our eventual advertising-based business model. Chalk would become our exploration of this last point: *How could we maintain Jot’s real-time conversational core, while remixing the surrounding mechanics to achieve viral growth?*

Jot Retrospective

*August 2018*

**Summary**

⇾ In the process of researching live video for Chalk, we noticed “live sales” emerging on Periscope and Facebook Live. Some live sales garnered **hundreds of live viewers** who were spending **$15,000 in 2 hours**

⇾ We built a **shopping cart for Facebook Live** that helped SMBs’ **grow their live sale revenue** by automating their sales tracking and invoicing processes

⇾ **SMBs paid us monthly for access to Baskets**. In sum, Baskets helped our customers process **$400,000+ in sales**

⇾ Baskets became unsustainable. Frequent, unannounced and undocumented **changes to Facebook’s API** made supporting existing features and promising new ones difficult

**The beginning**

In fall 2016, we introduced live video to Chalk. We thought room owners should be able to express themselves through different media (text, images, recording video and now, live video) and move fluidly between them. Around this time, adjacent services – namely, Instagram and Musical.ly – began introducing live video as well.

For Chalk to succeed, we understood its live video experience needed to be dramatically different from the live experiences on these other services. To differentiate Chalk, we sought a niche live video behavior that was a) **growing**, b) **constrained in some way by existing tools** and c) had the **potential to become mainstream** if supported properly. Of all the behaviors we saw, live selling was the most promising in our view.

By the end of 2016, dozens of live sales were happening on Periscope and Facebook Live everyday. Some streams attracted **thousands of live viewers**. People were selling everything from leggings to collectible comic books and sports cards.

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FJK2TihfQHuQFWIKOPiuB%2F1769c99f-dcb4-47e2-a8cb-e047cc564833.png?alt=media&token=120606d9-3b5c-41cd-9957-5e5abc9cadad","id":"1769c99f-dcb4-47e2-a8cb-e047cc564833","width":1600,"height":1432,"filename":"Screenshot 2023-04-26 at 9.56.56 AM.png","type":"image/png","caption":"","border":false}]}
```

**Product**

After watching many live sales and talking to sellers, we were **confident the format and behavior was “real.”** To understand the behavior more deeply, we had a few of our most popular teen Chalk users try hosting live sales with their audiences on Chalk.

In total, we had 10+ influencers host live sales. Not everyone had success, but a couple did. @Codyorlove and @Hknter each sold $200+ worth of clothing in under 30 minutes without ever having sold anything in this way before. There was something here, enough to keep exploring. We thought our greatest likelihood for success would be to **help existing live sellers generate more revenue**, rather than convert an unfamilizarzed population of teen Instagram influencers into live sale hosts. And so, we shifted our attention from Chalk to this new tool, Baskets.

Initially, we focused on the comic book community. There was **a thick population of sellers struggling to keep up with the growth** in sales volume and viewership. Between describing the comic book for sale, answering buyer questions, keeping track of and announcing each high bid and logging the winner of each item, sellers were overwhelmed. This frustrated buyers and sellers alike.

With the problem scoped, we built **a tool that integrated with Facebook Live** to track incoming bids, post real time auction updates to sellers’ audiences and log the final sale details for each item. Hosts loved having more time to focus on book selection and entertaining; the audience loved the instant feedback of knowing their bids had been acknowledged and documented.

We gained good traction in the comic book community, with **>50% of the most active sellers using Baskets and paying us for it every month**. We then expanded into streetwear and womenswear. To broaden our customer base further and service bigger retailers, we added functionality to support fixed price style sales. While auction-style sales were great for collectors selling one-off collectibles, auctions were limiting for bigger retailers whose business model depended on moving large volumes of product quickly.

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FJK2TihfQHuQFWIKOPiuB%2F908d2ae1-dc46-4f68-b18c-f135de6b875e.png?alt=media&token=49149cbd-7f37-4c0d-8dff-c7146d04e4bf","id":"908d2ae1-dc46-4f68-b18c-f135de6b875e","width":1576,"height":1090,"filename":"Screenshot 2023-04-26 at 9.57.41 AM.png","type":"image/png","caption":"Order dashboard ","border":false},{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FJK2TihfQHuQFWIKOPiuB%2F82e7b303-081b-43ed-9985-e0ab5d265a73.png?alt=media&token=4d4453c2-5109-4a91-af97-f6ec0e83a705","id":"82e7b303-081b-43ed-9985-e0ab5d265a73","width":1588,"height":1090,"filename":"Screenshot 2023-04-26 at 9.57.29 AM.png","type":"image/png","caption":"Live sale dashboard","border":false}]}
```

**What changed?**

Eight months in, we had developed a modest roster of customers and recurring revenue. With a degree of stability, we began to think on a longer time horizon. We started to design more involved features that our customers had been needing (i.e. integrated payment processing). Features that would, we thought, make Baskets a) **more valuable to existing customers** and b) **attract larger businesses** to become customers.

As we were building product, Facebook started to more frequently make unannounced **changes to their API that broke existing Baskets features**. Service outages ensued and it became increasingly difficult for us to guarantee uptime for Baskets’s existing features and promise delivery of future ones. We also had difficulty signing bigger stores who were skeptical of the live selling behavior.

With our momentum slowing on the back of Facebook’s shifty API practices and cash dwindling, we asked ourselves: *Is it sustainable and worthwhile to continue trying to grow a business whose success depended on a platform that could, and had started showing signs of, cutting off our service?* No. We didn’t think it was. At the end of 2017 we informed our customers that we would stop supporting Baskets at the beginning of 2018.

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FJK2TihfQHuQFWIKOPiuB%2F17466721-ac6a-4aa2-8797-59225a82baa0.png?alt=media&token=a40ebeff-2af7-4f03-b3bb-73800f4a9308","id":"17466721-ac6a-4aa2-8797-59225a82baa0","width":1606,"height":1372,"filename":"Screenshot 2023-04-26 at 9.58.53 AM.png","type":"image/png","caption":"Baskets-powered live sales","border":false}]}
```

**In the end**

All together, **Baskets helped sellers generate $400,000+ in revenue** through Facebook Live sales alone. However, we became increasingly uncomfortable with trying to build a business whose success wholly depended on a platform whose interests were not aligned with ours, while struggling to scale the business ourselves.

Baskets Retrospective

*August 2018*

**Summary**

⇾ For a small community of people, **Jot became one of their most intensely and frequently used daily apps**, but it failed to motivate users to share the service with their friends

⇾ Our focus with Chalk was to **achieve viral growth while maintaining Jot’s instant chat core**

⇾ By making Chalk **user vs. topic-centric** and shifting our target audience, we **unlocked viral growth** for a time

⇾ Unable to attract further investment, we doubled down on live video and **discovered a nascent “live selling” behavior** in the process

**The beginning**

Jot developed into a community of people who **talked for 30+ minutes each day on the service**, self-organized offline meetups in New York, Berlin, San Francisco, Atlanta and Washington D.C. and built relationships that persist today. Despite the intensity with which this community enjoyed Jot, the service did not grow as quickly as we needed it to.

We attributed Jot’s slow growth to there not being a compelling reason for people as to why the service would be demonstrably better with their friends on it, not to a deficiency in the core interaction. As such, our goal with Chalk was to experiment with the surrounding mechanics.

**Product**

Broadly speaking, the changes we made to Chalk were intended to make users’ experience with the service better, the more friends they had on it. To encourage this kind of behavior, we thought about Chalk as a game wherein the users who were able to create the **biggest conversations**, were rewarded with the **broadest exposure** on the service, **the most interaction** and thus, **inspire new users to strive for the same**.

Our first step in creating this cycle, was to bring more attention to existing Chalk power users. We did this by introducing **profiles** (which would evolve into user- owned chat rooms) and **performance metrics** (“Chill” and friend totals) to signify their skills for starting and carrying big conversations. Furthermore, we changed how chat rooms were ranked in the feed to prioritize the rooms that had the most people in them at any given moment. **Chalk was big, real-time conversations.**

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FpbhKCHNvbD9tZlhKitwU%2F0b6d4c2e-0abb-4016-880a-a467fcf39a94.png?alt=media&token=ee692c1c-997f-4aff-ae4a-fe68f1cef3bd","id":"0b6d4c2e-0abb-4016-880a-a467fcf39a94","width":1134,"height":946,"filename":"Screenshot 2023-04-26 at 9.52.35 AM.png","type":"image/png","caption":"Personlized chat rooms for each person","border":false}]}
```

At this point, the majority of Chalk’s active user base was made up of Jot’s most active users. But, despite the above structural changes, these users’ behavior didn’t change in the way we intended it to.The people who were on Chalk did just as they did on Jot: carry on great conversations with each other without bringing new people onto the service.

This led us to wonder: *Who were people that would value a tool to help them host big, real-time chats? Maybe there were people with existing audiences who were overwhelmed by inbound requests from their followers to talk, but lacked a space where they could do so at scale without the commitment of needing to be on camera? And if we were able to find people like this, how would their audiences respond? Would the chance to interact live with their "senpai" be enough incentive for them to sign up for a new app? Would the interaction be fulfilling enough for them to continue using a service like this?*

These questions led us to Owen, a teen Instagrammer with 30,000 followers who, in March, 2016, hosted his first live chat on Chalk. **With one Instagram post,** Owen brought on **60 new users in** a matter of **minutes**.

Owen’s success in bringing new users onto the service was a **turning point**. It led us to start **experimenting with Instagram influencers across a variety of subcultures** (net artists, Art Hoe Collective, self-care, makeup, body positivity, travel photographers, food bloggers, trans, LGBTQ, etc.) to see if other people could replicate Owen’s outcome. And they did. Over the next 6 months we worked with **400+ teen Instagram influencers** who helped drive **75,000+ new users sign up**. In the process, we built a series of **internal tools** to **automate the entire process**: from identifying influencers on Instagram to onboarding them onto Chalk.

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FpbhKCHNvbD9tZlhKitwU%2F60bbae4e-0f0a-4284-8633-96ba11818e42.png?alt=media&token=8169b8d7-d70f-4957-972c-a838b6ac2ace","id":"60bbae4e-0f0a-4284-8633-96ba11818e42","width":1728,"height":946,"filename":"Screenshot 2023-04-26 at 9.53.31 AM.png","type":"image/png","caption":"Instagram influencer chats and sharing","border":false}]}
```

As Chalk’s userbase grew, so too did the number of people participating in conversations. To help room owners better manage these conversations and inspire new users to start their own conversations, we started rolling out tools to make conversations more manageable for hosts and compelling for participants. Tools like Freeze, pictures, recorded video and Shaking.

**What changed?**

Ultimately, our topline growth was not enough. Despite cultivating a group of highly engaged users who would spend multiple hours on Chalk everyday and send, in some cases, **30,000 messages / week**, we were unable figure out a way to convert new users into daily users at a high enough rate. Our DAU/MAU suffered as a result. And without a growth or retention story compelling enough to overshadow the deficiency in the other, we struggled to attract investor interest. So, we **shifted our attention from fundraising to product**. Chalk stood for big, live conversations and with text and images already accommodated for, we shifted our attention to brining live video to the service.

```cv
gallery
{"layout":"hscroll","images":[{"src":"https://firebasestorage.googleapis.com/v0/b/maitake-project.appspot.com/o/pages%2FSYsPGdxrndMjN2fuM3qDhPSSNTT2%2FpbhKCHNvbD9tZlhKitwU%2Fd18aef47-49d7-424b-a289-574be74a519a.png?alt=media&token=49b4f12a-4d33-489f-86e3-f6666368cf5f","id":"d18aef47-49d7-424b-a289-574be74a519a","width":1736,"height":956,"filename":"Screenshot 2023-04-26 at 9.54.15 AM.png","type":"image/png","caption":"Deep engagement ","border":false}]}
```

**In the end**

Our belief had been that if we were able to attract enough users to the service through influencer-hosted chats, a big enough percentage of those new users would be compelled to aspire to become power users themselves. Ultimately, this did not happen with great enough frequency.

**Users acquired through influencer chats did not convert into power users frequently enough** to maintain Chalk’s 5% - 10% week-over-week growth rate. To remain competitive as a service for live interaction, we knew **we needed to support live video** and for the live video experience on Chalk to be markedly different from that on other services (i.e. Instagram, Facebook, YouTube and Musical.ly). Our quest for **differentiation in live video would lead us down the live selling rabbit hole** that would eventually become Baskets.

Chalk Retrospective

Product at Wikipedia

Baskets

Run in real-time with anyone in the world.

Chalk

A camera roll for your thoughts, Ping Practice is a nonprofit journaling method and app.

Ping Practice

The Power of Community

"How might software help volunteers, across experience levels, make edits that fit Wikipedia's policies and help newcomers learn these policies along the way?"

Editing the Wiki Way: software and the future of editing

Co-Founder and CEO

Building communication and editing tools to help volunteers keep Wikipedia accurate, exhaustive, and up to date.

Peter Pelberg

Side Projects

Work Experience

Projects

Speaking

Contact