Last week I launched Penguit.com (a service that helps you keep in touch with your friends - try it out!).
This was the first time that I single-handedly took a product through the entire cycle of ideation, development and public launch. I've developed many projects in the past - however, most of them never reached the public launch stage (they were either private tools for myself or alphas that I only released to a small group). At AppNexus, I took lots of products from ideation to public launch; nevertheless, it was very different doing it on my own with a non-work related product.
I realize Penguit isn't the next Google. In fact, if Google is Michael Jordan, Penguit isn't even the next Brian Scalabrine. At the same time, this was exactly the sort of project I was looking for - something small that would :
- Provide me with a feature that I wanted (I suck at keeping in touch)
- Get me back into the groove of full stack coding
- Practice a public launch on a somewhat inconsequential project. If I'm going to fuck up a launch, I'd rather it be on Penguit than on a future project / startup that I actually care about
It took me roughly 2 days to build Penguit. I released it last Thursday (March 7th). Here are my take-a-ways :
Things I did poorly :
- Focused too much on a making a "perfect" UI : I went a bit overboard here. I spent countless hours changing positions of each object on a pixel by pixel basis. After a while, I started to get obsessed with this. It's definitely an easy trap to fall into and one that should be avoided.
- Posted my "Show HN" post at 11:30 PM : Foolishly, I released to HN pretty late at night. While I instantly got some feedback (b/w 11:30 and 12) nothing really came in afterwards. As the HN ranking formula factors in time since post, I was already fighting a losing battle by the next morning.
- Lack of analytics on usage : I used Google Analytics very naively. I didn't build a way to distinguish between non registered and registered users (they both use the same domain so they are both reported the same way). I also didn't build a better way to filter out my own visits to the site.
- Didn't segment users : I wish I'd segmented users into 2 groups : those that did and didn't sign up. Without this, I have no way to engage with the users that didn't actually register for the site - what a waste.
- Landing page had no pictures : When I first launched, my landing page didn't have the standard 3 pictures that explain everything. It was just text explanations and a sign up form.
- I don't explain what Penguit does well. More on this in the lessons learned section below.
- Little Things : I initially had no favicon or title to my page. This just made it look unprofessional.
Things I did well :
- Getting the product out quickly : At its essence, Penguit is supposed to automatically email users when its time for them to reach out to someone (i.e. Bob tells Penguit that he wants to talk to Tim 3 times a month - Penguit needs to email Bob 3 times a month reminding him to reach out to Tim). I launched Penguit without building either the e-mail system or the algorithms to determine when to e-mail someone. I figured I determine when to email and actually send the emails manually while I waited to see how much traction Penguit got. This allowed me to push Penguit out more quickly and get user feedback.
- Focused on a nice UI : I tried to make the Penguit UI nice and easy to use. There's obviously tons of room for improvement, but overall, it's pretty decent. I know I said this is also something I did poorly - and in fact, it is. Striking the proper balance here is very difficult but very important.
- Personality : I tried to give Penguit some personality. In fact, I even named the mascot "Penguit the Penguin" and had her be the voice of the company (see the FAQ section - Penguit the Penguin is answering the questions). I also made it so Penguit the Penguin shakes anytime you mess up a login/register or add/delete a reminder (I got positive feedback on this from many users). I was definitely inspired by Hipmunk here.
- Engaging all feedback : Whether Hacker News comments, direct e-mails or tweets, I made a concerted effort to be very proactive in responding to everyone that gave me feedback. Thanks to all of you!
Things I learned :
- Building an effective landing page is really hard : I always thought marketing was easy. After all, isn't it just saying what the product is and why it helps the user? I figured I'd be able to make my landing page in 20 minutes and launch. 5 hours later, I was no where. Defining what Penguit does was much harder than I originally envisioned In my head, I know exactly what it does and what its virtues are. Translating that vision into a concise explanation that would entice users to sign up proved to be much harder than I'd expected. Frankly, the landing page still doesn't do a good job of explaining what Penguit is (I'd love any suggestions / advice on this one)
- Backlogs are never ending : I currently have a backlog of ~20 features I want to add to Penguit. Pre-launch, I kept trying to do more and more (bulk reminder adding, more advanced reminder frequency inputs, tagging of reminders etc..). It's really hard to actually stop at one point and say "That's it. I'm launching." Having a very clear demarcation of what you want to include when you ship is really important - not just because it lets you know when you can release, but because it gives you some mental freedom from constantly questioning whether or not it's time to release.
- Most users will go to the site but not sign up : This one is pretty stupid of me. I figured that mostly everyone that actually went to Penguit would sign up. Boy was I wrong.
Open questions :
- How do get more sign-ups?
- I launched Penguit through 1) an announcement on HN 2) tweeted my followers a couple of timers 3) emailed my contacts. Now, ~5 days later, I'm no longer getting users to the site (and thus, no sign-ups). I can't say this is unexpected by me - in fact, it's exactly what I was expecting. At the same time, I'm unsure of what steps to take to increase adoption. Ideas that I have for getting users include 1) Have users follow Penguit when they sign up 2) Incorporate some sort of Tweet / Post ("I kept in touch with @penguit" style tweet) in the reminder emails 3) Advertising (not going down this route but possible would have if I segmented users) 4) Make a more useful product 5) A/B testing on landing page to figure out what converts best (However, this doesn't help as I'm currently not seeing much traffic to the site)
Results of launch :
- 35 users signed up (ok - actually 27 users not including family and friends).
- 140 reminders added to the system.
- Most reminders for one user : 27
- 224 unique visitors for a 12% signup rate.
- Top traffic sources : #1 Hacker News (73%) #2 Twitter (20%)
And that's my launch. I'm unsure of what my next steps will be / if I'll continue to add functionality to Penguit / if I'll even continue to support Penguit. Considering what my original goals were with this launch (see above), I'm pretty happy with the results. I learned a bunch about how to launch a product and expect to do better in the future.
Thanks to all of you that signed up for Penguit!