Playdate: From 5k Users to Zombie - The startup that burnt $40k and shut down

Hi Logan! What's your background, and what are you currently working on?

Hey there, my name’s Logan. I’m 28 and I’m an engineer in San Francisco. My majors were in applied math and behavioral economics, beginning my career in finance in 2014 and transitioning into software engineering in 2015. I’ve successfully failed 3 companies so far, and I’m most passionate about making life happier and better for others.

I’m currently one of the founding engineers at a small startup called Retain.ai, helping companies understand how they spend their time. If you are a designer or full-stack or frontend engineer looking for a fun startup environment with no politics and tons of learning and mentorship, email me at logan@retain.ai!

As far as side projects go, I’m currently working on a mobile trivia game app for the TV show Friends called Bamboozled, which is in development. I’m hoping to deploy to the App Store and Play Store by the end of 2019.

My most recent entrepreneurial pursuit was being CEO and CTO of Playdate for 2 years. I alone grew our user base to 300+ users and social media accounts to over 300+ hundred followers over a month, before growing the team to 7 at its peak and about 5k MAU (monthly active users).

Playdate was an on-demand social networking app. Our users could pick an activity and at a tap of a button be instantly matched with another person. The idea was that we could use science to minimize “time to meeting” for people looking for short term or long term friendship. The app was free with no premium features or ads, and we aimed to make money by offering coupons from nearby venues that were only revealed upon physically meeting for a playdate.

‍

What motivated you to start Playdate?

I’m strongly introverted so I know all too well the pain of being crippled by shyness and social anxiety that can make making friends incredibly daunting. I tried all the social networking apps I heard of but they simply didn’t work for making friends. Over the years, I started noticing others, even extroverts, having similar problems. As I learned about the loneliness epidemic, I realized the world needs an app that actually works, helping people easily meet others and create friendships. So I began working on this as a side project in 2016. 

Using my background in psychology, I was determined to remove as many psychological barriers and cognitive biases as possible. For example, I addressed the Paradox of Choice by replacing the standard swiping functionality with a button that would instantly match you with someone compatible nearby. I believed at the time that this particular decision would also help mitigate the “chicken and the egg problem” because users would not see how many (or few) others were in their area as I grew the user base. 

I knew I wanted to make Playdate centered around activities from the start. I had made most of my friends through common interests and mutual activities, such as playing poker or drinking together. There was also a McGill study that demonstrated that strangers would exhibit friendship-levels of empathy after playing a video game for 15 minutes together. I figured if two strangers did something that they were both passionate about, this would increase the chance of them getting along. And even if the friendship didn’t stick, neither person would feel like they wasted their time because they were doing something they would want to do anyway.

Playdate was always activity-focused, but the big question remaining was which activity to start with. I considered “Magic: The Gathering” because it’s a popular game that requires 2 or more players to play, and Magic players are sometimes known to have trouble making friends. Embarrassingly, I have played against myself in the past because I had nobody else to play with. Ultimately, I decided against Magic and went with cannabis consumption. This was because cannabis is a social lubricant that can help people feel less anxious hanging out with new people and the legalization revolution had begun across the world. I also hypothesized that stoners would be more willing to try an app like this and would be more amenable to giving feedback. In hindsight, this choice of first vertical would be one of the reasons for Playdate’s downfall.

‍

How did you build it?

I started working on Playdate as a side project in early 2016. I had assessed the options for mobile development and decided on doing it natively at first, a decision I would later regret. Swift was relatively easier to learn than Java, and I decided to supplement development with contractors. I hired a few different ones through Upwork and had modest success. As a developer myself, it was straight-forward to vet the offshore teams before starting contracts with them.

As I worked on this, I realized that I could not juggle my full-time job as a developer with this side project in the timeframe I wanted. So I quit my job in August 2017 to pursue Playdate fulltime and found a cofounder in early 2018.

Around April 2017, we soft-launched Playdate to the App and Play Stores. After we got initial traction, we garnered as much feedback as we could from hundreds of users to help design the next version. I ultimately decided to rewrite the codebase using React Native for the following reasons: React Native was in a much more stable place than when I had last looked into it on project inception; it had garnered much more of a community and available online resources. Facebook had also changed the license from their own restrictive one to the MIT License, making its use feasible. Moreover, React Native was certainly appealing because JavaScript is my most familiar language and our designer-turned-engineer was similarly comfortable with JS. Finally, there were some “nice-to-have” features like hot reloading sims, but perhaps the most important impacts were that the overhaul provided a way to rebuild our client-side foundation to erase significant technical debt and helped build team morale with a tangible win. 

Playdate was free and ad-less, and would always be that way. We planned to make money by providing coupons from venues when users had an in-person playdate, but this was to be post-MVP and after building up a user base. During the lifetime of Playdate, we focused exclusively on growing as quickly as possible.

Here are some pics that show the evolution of the app:

‍

It was around this time I used a friend-of-a-friend designer to create the logo (best $20 spent ever). Lesson learned, a good designer can create logos and designs faster and better than I can. Awesome specialists are worth their weight in gold. 

What were your marketing strategies to grow your business?

We knew marketing would have to be done well because we had the classic “chicken and egg problem.” We overcame that by doing things that didn’t scale (initially legally giving away branded bags of cannabis for downloads, co founders automatically matching with users on their first playdate, etc) and by creating content for inbound marketing. We also had a theory that because the app sent curated matches, our users wouldn't see or feel a significant loss in UX based on low user density. This would allow our user base to steadily grow and provide better matches over time, as long as we could retain a good number of users.

We explored several marketing avenues, beginning with a grassroots campaign. We created hundreds of branded bags of pot and gave them out to legal adults throughout campuses and popular smoke spots in exchange for app downloads. This was technically legal because we weren’t selling it and it was impossible to construe this as sales because our app was completely free. This also helped us seed a small user base in a small space. This ended up burning us in the end (pun intended) because these users didn't stick around, causing terrible retention and poor matches.

SEO was difficult because we couldn't get many good inbound links (I spent a few hours writing a Medium article but it went nowhere), but we had our SEO bases covered on our website. ASO was much easier; by identifying lower-competition, higher-traffic keywords like "meet people fast", we were able to become #1 in the App Store for a search for "playdate", "meet people fast", and "meet stoners." 

The thing that worked best for us was our inbound marketing. Facebook and Twitter were pretty worthless, and we wrote one blog post which had a really low ROI. Instagram ended up being how we got most of our users. We created an Instagram profile with daily memes that had mild success. We grew to thousands of users by providing people with content they wanted and infrequently converting them into app users. We also paid influencers to advertise us, which was a total crapshoot. Some influencers that we didn't think would do well charged $100 and we got many hundreds of followers. Others that we thought would do much better, like Cheech and Chong's official profile, barely moved the needle.

‍

Which were the causes of Playdate failure?

At the end of the day, Playdate failed because we ran out of money, like all failed companies. The causes were, however, distinctly ours. First, we chose cannabis consumption as our first vertical for reasons outlined above. But this was a mistake. Our stoner user base turned out to not be more likely to give us feedback; it was actually quite the opposite. Many users didn’t feel comfortable attaching their online presence to marijuana, some were paranoid about law enforcement, and the rest were doing other high things that didn’t include giving feedback or writing emails. We offered free cannabis and free pipes to our power users in exchange for feedback, and we still didn’t hear back from a single user. Even though legalization was in full swing, the stigma was still very much there. Moreover, illegal states were tricky to navigate as they were murky legal areas.

Another major problem was the “chicken and egg problem,” which we had thought we could mitigate with our matching mechanism. Although users didn’t have the chance to run out of users to swipe or see how few other users were nearby, they still ran out of compatible nearby potential playdates far too quickly. Speaking of, we couldn’t provide unlimited playdates so we only allowed three a day. This was because users would exhaust their pool of potentials too fast and we wanted to make each playdate valuable in order to encourage people to actually go on their playdates. All of this tied our hands during a time we should have been freer to explore different methods and ideas.

Another thing that burned us was our organic growth. We simply didn’t grow fast enough to make a social networking app take off. If we were able to geofence app downloads more stringently (we could only do it by country), we might have been able to maintain user density. Initially, we thought we could drive growth in one location by going hard on our grassroots campaigns to give out free cannabis for app downloads. But this didn’t work because these users would not stick around, causing terrible metrics and bad matches with inactive users. Furthermore, we didn’t think to geofence the app through our API or other means, which led to random users from all over the country finding our app through our SEO and ASO methods. This would normally be a great thing except user density was more important than growth, an insight we realized too late.

The final straw was when I realized we were “the walking dead” or “a zombie,” a company that wasn’t dying exactly but wasn’t growing fast enough to gain momentum or be interesting to investors. We officially ended things on February 22, 2019, after sending farewell emails to all our users.

Which were your biggest mistakes and challenges you had to overcome?

One challenge we had was Android QA. The team was covered on iPhones and we had three Android phones. The problem was that there is such a variety of Android phones that we were constantly battling bugs that we couldn’t reproduce in simulators or on the phones we had. We were mostly unsuccessful in managing this problem but we occasionally got lucky. In one case, a user emailed us complaining that our app didn’t work for her and she was going to leave a bad review (by the way, almost all of our inbound user emails were either appreciative and helpful or mean-spirited and hurtful). The initial email was vague and emotional, honestly like 99% of the emails. I responded immediately and made sure she knew she was being heard. I connected with her a bit and as our conversation continued over days, I was able to uncover all the technical details necessary to fix all the bugs she was experiencing. While performing a root cause analysis, I also got to know her as a person. She was very appreciative, and she ended up not leaving a negative rating. This was just one of many times that I’ve been able to successfully interact with disgruntled users. Applying empathy and mindful listening to diffuse difficult situations while applying principles like the 5 Whys and the MomTest has been successful in turning detractors into advocates.

Another interesting challenge with making a friendship-only app was that people genuinely did not seem to only want platonic relationships. About half of our users wanted to use the app to date new people. We could’ve banned or prevented signups from these romance-seeking users, but this would’ve meant (1) ignoring what our user base was telling us they wanted, (2) taking the time to prune our existing user base and throttling signups, (3) missing a good opportunity to learn how to achieve better product-market fit, and (4) potentially lowering the quality of our playdates by decreasing user density. On the other hand, we could’ve let them stay, but that would mean (1) compromising our values and turning Playdate into what amounted to another dating app with a twist, (2) converting from a friendship-only app to a friendship-and-relationship app likely too early (it’s easier to go from friendship to both over relationship to both, but once you make that leap it’s irreversible), and (3) making about half the playdates incompatible because one person would want just a friend while the other would want more. We ended up finding a compromise by adding a new feature called “looking for.” It was basically a user-selected flag for “strictly friends” or “maybe more.” This essentially split our user base in two but allowed us to continue to serve both types of users and gather data and feedback about both.

Also, don’t hire any interns for a while. It is decidedly negative expected value.

‍

Which were your expenses? Did you achieve some revenue? In the end, how much money did you lose?

Our business model couldn’t monetize until we signed venues to pay for coupons, but this was a bit of a catch-22 because we couldn’t get venues to sign up without more users. I burned through about $30-40k when all was said and done over the course of 2 years. I had saved enough money for about 2 years of personal runway, but at the end of the day, I wish I had been about 20-50% more conservative with my living expense estimates and didn’t rely on raising a seed or going through an accelerator.

If you had to start over, what would you do differently?

I would’ve begun development in React Native instead of Swift and Java for faster dev cycles. We could’ve iterated a lot more if we had used React Native from the start. Without any animations or UI performance concerns, Playdate would’ve been much better had we started with React Native.

I would’ve also grown the team more slowly, and being more discriminating with friends who wanted to help part-time. Counterintuitively, one lesson I learned in people management is that it is far more stressful to manage someone who unreliably does their work than someone who never does their work or always does it poorly. 

The quality of our users was more important than I’d thought from the start. Doing things like growing organically but not virally and playing with campaigns like giving out free stuff for downloads meant that too large a proportion of our users were fake or inactive. It was particularly difficult to separate users we should’ve disabled from those who were waiting for new users to sign up in their area.

I also put too much weight on special Silicon Valley events like “pitch competitions” wherein you pitch your companies to a crowd of entrepreneurs and investors. This turned out to largely be a waste of time. The time would’ve been better spent on getting user feedback, development, and chasing smart money through my network. We also applied to some accelerators, none of which accepted us. Got some good photos out of it though (notice the original logo!), and I could still see something like this being the positive expected value.

One thing that we were torn internally about and spent far too much time on was our name. Playdate usually evoked a significantly positive or negative response in people, and despite my best efforts, the team kept bringing up the name without good suggestions of replacements. “Playdate” is already a fairly common word so it had unwanted associations and usernames/domains all had to be “MakePlayDate.” In the end, it didn’t matter and it was a curiously emotional matter for people.

‍

Which are your favorite entrepreneurial resources?

I love reading books, and can’t support it enough. I highly recommend reading the classic startup books from The Lean Startup to The Hard Thing About Hard Things to Zero to One. A lesser known book that really helped me rethink how I thought about users and getting actionable feedback was The Mom Test.

When I listen to podcasts, I listen to How I Built This by NPR.

Startup School’s videos are free and pretty awesome too.

‍

Where can we go to learn more?

The old site for Playdate is still live here. If you want to check out our first and only unsuccessful blog post, it’s here.

Our Instagram and all the old memes that helped us get thousands of users live here.

You can find me here. I currently work at a small startup called Retain.ai (we’re hiring engineers and a designer!).