Learnings from Building reread.io
tl;dr Built reread.io over a weekend. Briefly tried to growth-hack/market it. Only marginal success. No revenue model. Now Open Sourced.
The Beginning
I have a long-time problem with my to-reads or read-it-laters.
I seldom read them.
Because I forget. Because I procrastinate.
These to-reads end up sitting in my Pocket, hardly seeing daylight again, although I know most of these articles would be beneficial to me.
I also have a long-time problem with bookmarks in general.
I never read them again.
They are great articles. I thought I would refer to them again in the future. However, I don’t. I probably don’t even remember I bookmarked them.
These bookmarks are hidden out of sight (again in Pocket) and forgotten. Basically, I missed the opportunity to reread these interesting articles and gain new insights from them again as suggested by Paul Graham.
So I sat down one weekend and created reread.io:
reread.io is a service which sends you an email containing your forgotten Pocket links (with many config options).
Marketing
We already all know:
if you build it, they may not come..
Startups don’t succeed the day they launch. A tremendous amount of effort has to go to marketing and spread the word about the product too.
Well, reread.io is not a product that I am betting on (just a weekend hack!), but it would be cool to see some traction. And maybe hockey-stick growth?
Given that Pocket has 17 million users as stated on their website:
And that reread.io actually seems to be like one of the Top Feature Requests (search for “Reminders to Read/watch) by Pocket users, which Pocket don’t plan on working on:
I was pretty positive about reread.io because even if only 0.01% of 17 million gave it a try, that’s 1,700 users.
But how do I reach out to the 17 million users?
Let’s start with my own personal network, so I started tweeting about it. Those tweets were retweeted and liked but the network effect soon tapered. I don’t have a wide enough network.
Let’s try “pushing” it to people who like or are associated to Pocket then. Hence I did a search for “Pocket” on Twitter and sent out a few tweets:
I also pinged Nate Weiner, Founder of Pocket, who seemed to like the idea since he also retweeted it! That didn’t result in hockey-stick growth though..
Of course, I also tried to put it up on Hacker News, but the stats was dismal.
Somehow, I haven’t had much luck with any of my posts on Hacker News. Am I posting at the wrong timing (I usually pick morning, SF Time)?
At that point, the other logical thing to try would be Product Hunt right? Unfortunately, I wasn’t part of the “circle”, so I couldn’t post.
Hence I turned to Quora next to ask for some advice.
Yuval Shoshan replied and he not only suggested that I could try posting it on Product Hunt, he also sent me an invite. Thank you!
And he also mentioned PocketRocket which has 228 likes on Product Hunt. Given that reread.io is similar to PocketRocket and that people liked the PocketRocket idea, I believed that reread.io should do well too.
Up it went on Product Hunt. The result (13 likes) was astonishingly funny. Probably that’s why Product Hunt doesn’t even want to list me as the maker of reread.io too (or maybe Product Hunt is just broken).
Although reread.io is pretty similar to PocketRocket in idea and concept,
it actually started with more configuration options than PocketRocket,
and I have also just pushed a few new features exclusive to reread.io.
Read the “New Features” article to find out more.
All the numbers above looked terrible. 10 points on HN, 13 likes on PH..
Sorry to disappoint if you were hoping to learn some growth-hacking tips but reread.io didn’t achieve hockey-stick growth. But it does have users!
Current Stats
Let’s first look at how Google Analytics looks like:
The first spike, at 220 sessions, had traffic coming from Twitter, HN and PH. The second spike, at 150 sessions, interestingly was a result of a comment that I made on this Medium post by Tiago Forte.
From the stats, Bounce Rate is pretty high which could mean my copy sucks. Also, reread.io by nature is not a sticky app. You just have to config it once and let it do its work every day/week, which explains the low “Return”.
In any case, these are essentially vanity metrics.
How about Active Users and other stats?
As of 5 Jan 2015, reread.io has Active Users of 86%,
and Email Open Rate of 56% and Click Rate of 19%.
According to MailChimp’s Email Marketing Benchmarks, the average stats across various industry stands at 22% for Open Rate and 3% for Click Rate. This means that reread.io is doing very well on these fronts!
Previously, I did mention that reread.io does not have hockey-stick growth, and so far, reread.io has managed to attract a market share of…
0.001% of 17 million users
That’s about 170 users after about 2 months since lauching on 15 Nov.
Alright, enough with the percentages, let’s look at the absolute numbers:
As you can see, the actual numbers are really, really tiny. 200 users? C’mon!
But I can still see growth, although I have already stopped doing any active growth-hacking marketing.
More importantly, amongst these 164 active users, some of them really do appreciate reread.io and they actually bothered to tweet or send me emails about feature suggestions or bug reports. Thank you!
Passionate Users
Occasionally, I’ll see someone tweet about their fondness for reread.io:
Or submit a Feature (not a Bug!) report:
And to top it off, a “Take my money!” request:
Contacted By Pocket
Given that it did caught the attention of Pocket (since I tweeted at them), they have contacted me to talk about reread.io.
Nothing happened (yet?).
In reality, reread.io is a weekend project (and another for the new features). Pocket can probably get their engineers to replicate reread.io in a week or so with scale in mind for 17 million users.
What I have just done is a MVP to validate the hypothesis that people would take to the idea of a “Timehop for my Bookmarks”. Yes. 86% of them.
What’s Next?
Features
Over the last few weeks, I have added four new features to reread.io based on suggestions by users. Read the article for the details.
There might be one or two more features that I might add, but we’ll see.
Revenue Model
Honestly, I don’t have one. It was a scratch-my-itch project remember?
reread.io is currently hosted on Heroku and free dynos are enough for it. But it has already outgrown the Free — 10k rows Hobby Postgres database and has been migrated to the Basic — 10M rows tier (at $9 per month). That’s the one cost I am incurring right now though.
So what’s good as a Revenue Model? Subscription? Advertisement?
I should probably post this as a question to the users, but would love to hear of any other suggestions!
Open Source
Finally, although this is a weekend project, I did not scrimp on code quality. In fact, the app is now open source on GitHub and you can take a look at it. This a Ruby on Rails app developed TDD-style using RSpecs for testing. Have a feature that you would like to add? Pull Requests are welcome!
This is also an alternative to having a Revenue Model: Host it yourself. :)
That’s all! Hope you found this Retrospective insightful and entertaining. Give reread.io a try now!
Do you have a tech-product idea but lack the technical expertise to realise it?
At Jolly Good Code, we specialise in rapid development of high quality MVPs. Let us help you turn your idea into reality and ship it. Contact us.
