Hi everyone! This is the second update of my product management summer internship at Twitter covering weeks 4, 5, and 6. I hope you enjoy it!
#HackingAway (Week 4)
One of the highlights of my internship so far was the Twitter Blue team’s internal “hack week” during week 4. Hack weeks are very common at tech companies because for one week, teams are able to take a pause with their normal work schedules and are given the chance to test new ideas for the company. The beauty of hack week is that it allows employees to build what they’re excited about into reality when it may not have been possible otherwise. In fact, Twitter itself was born from a hack week- like activity! Odeo, a podcasting company, wasn’t doing so well due to competition so employees split into teams to come up with new ideas. An Odeo engineer named Jack Dorsey (you may have heard his name!) then came up with the initial idea of Twitter.
The purpose of Twitter Blue’s hack week was to spend some time exploring potential new subscription features and enhancements for Twitter Blue v1. We created a list of ideas and the team either worked individually or in small groups on about 7 different ideas. What caught my eye on the list was the idea for collaborative bookmark folders, which would allow Twitter Blue subscribers to view, edit, and share Twitter Blue’s current offering of bookmark folders with other users. Mai Elamin, who is one of Twitter Blue’s awesome designers, and I decided to work on this project together with the goal of fully researching, scoping, and designing what the collaborative bookmark folder experience could be like.
The original idea for collaborative bookmark folders came from the question of “how can we add value to Bookmark Folders to make it an even more premium feature?”. Knowing what our goal was, Mai and I designed what collaborative bookmark folders could be like using this process:
- Conduct research (Who are the collaborative bookmark folders’ users? How do users and other products share information? What do current Twitter users think of collaborative bookmark folders?).
- Document research findings and analysis.
- Collaborate with cross-functional partners (such as engineering and legal) to scope the feasibility of collaborative bookmark folders.
- Create low-fidelity designs of bookmark folders.
- Get feedback on designs, and then iterate and test on new designs.
- Pitch collaborative bookmark folders at the end of the week!
What I most enjoyed about this process was being able to work side by side with Mai throughout the research/design process and collaborate really closely together. We met Monday through Thursday (early morning for me and late afternoon for Mai since she’s based in London!) to talk about the project and work together on our daily goals, such as reviewing Mai’s research documents or going through the potential different use cases of collaborative bookmark folders. Afterward, we would then work asynchronously on our individual tasks before updating each other again in our daily meetings. Not only was I able to work on an exciting potential Twitter Blue feature and exercise some product manager (PM) skills use cases, potential risks, who is our user, etc.) - but I was also able to really take a deep dive into the designer/PM partnership for a week.
#ScopingExperimentingAndIllustrationsOhMy! (Week 5)
Throughout weeks 4 and 5, I continued to collaborate with the Customer Journey and Twitter Blue teams on scoping out the details of upsells. An important prioritization that occurred was optimizing the implementation schedule of our three chosen upsells based on dependencies. To help break down the upsells roadmap, we decided to start with one inline upsell per subscription feature - Bookmark Folders, Undo Tweet, and Reader Mode. Each of these upsells would be inserted into a specific Twitter timeline (i.e. the home timeline for the Undo Tweet upsell) when shown to a potential subscriber. However, we quickly learned that not all timelines on Twitter had the ability to support an inline prompt. The larger timelines, such as the home timeline, could support inline prompts whereas smaller timelines like the bookmarks and conversation timelines where we want to insert the two other upsells, did not have this capability. With this new information, we decided to prioritize the Undo Tweet upsell as our first milestone and then build the Bookmark Folder and Reader Mode upsells after their inline prompt support had been built in by the Customer Journey team.
In addition to upsell prioritization, I also worked with cross-functional partners on illustrations, copy, and experimentation for the upsells. Although the upsell design had been finalized, we still needed to create custom copy and illustrations for each upsell. To do this, I worked with Julia (Twitter Blue content designer) on the copy for each upsell and really focused on what our primary “call to action” text should be. Although it may seem like a minor detail, using specific words can greatly affect a user’s behavior so it was important for us to think about what action we wanted the user to take and have the wording reviewed by cross-functional stakeholders. Then, Tegan (Twitter Blue designer) and I started the collaboration process with the creative team on illustration mockups for each upsell. We started out by sending over some inspiration pictures and discussing the feeling we wanted upsells to create in users, such as excitement, curiosity, awe, etc., to help the illustration designers start some iterations.
Working on defining experimentation for upsells was also important during this time to make sure we could properly attribute any upsell effect on overall subscriptions. With Scott’s (Twitter Blue data scientist) help, we put together some success metrics (how we would tell if upsells performed the way we hypothesized) and experimentation that could be run. To make sure that we did not have any bias on the effect of upsells, we planned to initially show upsells to a test group of eligible users while not showing upsells to a control group. This way, we could see if upsells did have an effect on subscription behavior. Additionally, Scott and I planned out some experiments to run depending on the data we received after the upsells had been live for a while. This included experiments such as testing two versions of copy or two versions of illustrations to make sure the upsell model we had was the most optimized.
Some key takeaways I got from going through this scoping, experimentation, and illustration/copy experience were:
- Learning how to prioritize with project dependencies and working with other teams’ timelines for the first time. I didn’t have a lot of experience doing prioritization while also considering dependencies and other teams’ timelines, so continuing to work with Customer Journey team was valuable because it taught me to expand my prioritization scoping skills.
- Reinforcing the importance of consulting with cross-functional partners early and often. Talking with cross-functional partners, especially Customer Journey and engineering, helped to inform my decisions and shape what the product requirement document looked like. I know it’s said a lot in product management, but collaborating early with partners is super valuable and something that should be done every time when starting out a new project.
#HalfwayThroughMyFlight (Week 6)
Time flies when you’re having fun, right! My internship midpoint came really quickly with lots of new changes and additions around upsells. One of the biggest changes to the upsells roadmap was an addition of a new upsell — the home timeline upsell. As we dove into technical scoping of the Undo Tweet upsell, we learned that Customer Journey’s timeline of building the inline prompt support for the other two upsells would be longer than expected. With Smita’s (my awesome manager) help, I realized that in order to achieve the original goal of upsells, we needed to take a step back.
The goal of upsells is to drive awareness and conversion for Twitter Blue. By going forward with the Undo Tweet, we were making progress towards this goal, just at a smaller scale. This is because not all of our eligible potential customers would see this first upsell — only users who had deleted at least one tweet would see the Undo Tweet upsell. To make the best use of our time and making progress towards that overall goal, Smita and I came up with the idea of a general Twitter Blue upsell on the home timeline. This upsell would help us better drive overall awareness because all potential customers would be eligible to see it, therefore allowing us to take a wide approach with the first upsell. Additionally, this upsell had no dependency since we were placing it on the home timeline and we could use the work already scoped for the Undo Tweet upsell and transfer it over smoothly. With this home timeline in mind, we made this new upsell our first milestone and organized our specific upsells of Undo Tweet, Bookmark Folders, and Reader Mode into our subsequent milestones.
To share this new addition as well as the current state of upsells, I held a PRD review with all upsell cross-functional partners and teams. The purpose of my PRD review was to update everyone on our progress, highlight important changes and prioritizations, and then discuss any questions that partners had. I first shared the upsells competitive analysis that I had put together, which had the goal of understanding the industry standard for upsells in paid subscription products. I analyzed 9 different paid subscriptions’ upsells and found common patterns in the upsell messaging, content, and placement to better develop our own Twitter Blue upsells. Then, I highlighted important changes such as the home timeline upsell as well as our prioritization decisions based on Customer Journey dependencies. Finally, we all walked through the PRD together and really fleshed out any grey areas and questions that needed to be discussed further.
The PRD review helped to realign myself and in turn, my cross-functional partners, on what we were trying to achieve with this particular project. Through each week, upsells has been like an onion and revealed a new layer to dive into, which has helped me to become better at considering all angles (pros, cons, risks, etc.) when approaching a new problem to solve. Although making big decisions, like the home timeline upsell addition, may not always be the easiest decision, it has really taught me to go into future work always anticipating and constantly questioning of what can be done to make sure we’re constantly executing towards the project’s end goal.
Read on for my weeks 7, 8, and 9 update here!
