On Wednesday 30th January 2019 we finally launched Leave Me Alone – a service to easily unsubscribe from spam emails with a focus on privacy.
If you just want to skip to the stats you can find them here.
The best time to launch on Product Hunt is 00:01 San Francisco time as this is when the new Product Hunt day starts. Launching at this time gives your product the most amount of exposure and time to get votes, and therefore the biggest chance to bag one of the all important 1st, 2nd or 3rd product of the day badges.
We launched Leave Me Alone from Huancacho, Peru which meant we would go live at 3am. After just 2 hours restless sleep (too much excitement obviously) our alarms jolted us awake. Laptops at the ready, Product Hunt, Twitter, and analytics open we waited and then BOOM we were launched!
00:18 – incident 1
We performed a routine check of the landing page and noticed our stat suggested we join 0 users who have unsubscribed from a total 0 spam emails. Seeing as we had over 500 users, something was definitely wrong!
“OH MY GOD HOLY COW WHAT HAPPENED”
We had deployed a dodgy config file and the app wasn’t connecting to the live database! Luckily we caught this early and it was quickly fixed and deployed. Phew!
Easy like Wednesday morning.
By the end of the first hour we had an incredible 100 votes. Things started picking up even more in hour 2. We hit 24 active users online and were seeing 6-10 people running scans on their inboxes at the same time.
The server held up as we frantically replied to comments on Product Hunt and Twitter. We’ve never had this much traffic on any of our sites – it was so exciting!
In hour 3 we got our first 5 sales and were had reached 223 votes – more than any of our previous launches have ended with! We were neck and neck with another product, and even though they had more votes than us we were still #1 for now.
We watched the sun rise and posted to Indie Hackers. Unfortunately, we also noticed a bug with our Product Hunt banner on mobile which caused the entire header navigation to bunch up and disappear. I powered through and set about fixing this while James decided it was all too much too early and had a nap – I was definitely not pleased and neither was he when I woke him up to deploy my fix!
Then the internet in our hostel died! We tried to use our mobile data but nothing was loading and I started to panic.
It was nearly 8am, so we grabbed our laptops and sprinted out in search of internet and breakfast. We had already scoped out the cafes with the best internet (and food) in our little surf town so went straight to our favourite one – Chocolate Cafe Peru if you’re ever there and in need of a good place to work!
Obviously they didn’t open on time so we were camped outside using their wifi, dying of hunger, and desperate for a coffee trying to deploy the fix for mobile.
07:00 – incident 2
I had long since fixed the styling bug with the PH banner but we couldn’t find a window in the traffic to deploy. On any normal day this would have been a good problem to have, but not on launch day! As the bug was front end we could deploy the fix by building the code locally and copying it over without restarting the server.
This worked brilliantly until we realised that we had built with our local Stripe API keys and we had broken the payments.
“AAAAH OMG. I AM TOO HUNGRY TO FIX THIS. COFFEE REQUIRED”
This time it was more serious, we had tens of active users and people running scans. Luckily it was a quick copy and paste fix and the crisis was averted within 5 minutes. Phew again!
Sales rolling in
By hour 8 the sales were starting to roll in hard and fast. We were receiving Digital Ocean bandwidth alerts every few minutes and the next few hours were spent refreshing Product Hunt, watching Google Analytics, managing hundreds of notifications on both Twitter accounts, and celebrating every Stripe payment notification!
The Hacker News Effect
We planned to post to Hacker News at 9am SF time but completely forgot! We posted to a Show HN at around 09:45 and shared with a few people in the maker community to try and build some early traction. It’s notoriously unpredictable whether a post will make it to the front page of Hacker News so we didn’t expect any success. You can see the post here.
Then it started climbing. More and more users were landing on the site. Every time the number increased I thought that would be the peak and took a screenshot. We hit the front page and watched it climb to 70 users active right now and the server was still hanging in there!
James was excellent at replying to the comments. He was very diplomatic and actually enjoyed it! I focused my efforts on PH and Twitter, I’m not so good at dealing with difficult comments!
11:00 – incident 3
The Hacker News effect was exhilarating and exhausting. Then disaster struck! We had our first outage and the server was down for around 15 minutes.
“THIS. IS. BAD.”
We aren’t sure what happened but our database crashed and had trouble restarting. This time James came to the rescue and performed some lightning fast devops to repair Mongo and bring us back online while I just twiddled my thumbs and asked repeatedly if he’s fixed it yet.
If you didn’t have an outage, can you even say you launched?
Halfway to a thousand
Thankfully that marked the end of our incidents as we reached 500 upvotes and 1,000 Leave Me Alone users. The traffic remained at a consistent 60-70 active users for the next few hours and the first support requests started coming in.
There were no big problems and most of the issues were scans not finishing or taking a long time to run because the server was under such a heavy load.
Hacker News Demotion
It was now launch hour 14 and 5pm in Peru. The 2 hours sleep was starting to take its toll and we were both feeling slightly delirious.
We were also removed from the front page of Hacker News and demoted from position 28 to 238 instantly. Rumour has it that a post will be demoted if the number of comments is too similar to the number of votes. It may be coincidence but we had 48 votes and 47 comments when were removed. You can see our ranking progress here http://hnrankings.info/19037399 (thanks Leandro for sharing the link).
The removal from the front page of Hacker News reduced our traffic back to a manageable amount. We were slightly disappointed, but honestly also quite relieved.
It was now 6.30pm and we had eaten all 3 of our meals sat in the same cafe. It was time for a leg stretch and a technology break. One of the benefits of launching from a surf town was that we could grab a couple of local beers and watch the sunset on the beach!
Launch hour 20 was now upon us and traffic was still steady. We stayed up in our hostel and dealt with a few more support requests but we were both absolutely wiped out. I wish we could have seen the full 24 hours through but we are definitely too old to pull stunts like that now!
We ended our day on 821 votes and #1 product of the day. We hoped we would wake up to find we stayed there!
🥇 🎉 NUMBER ONE!! 🎉 🥇
We woke up and grabbed our phones to find that we finished as #1 product of the day and had almost 1,000 votes!!
The launch went better than we could have ever imagined. Not only did we manage to sustain a huge amount of traffic and make an astonishing number of sales, but we also got a lot of incredible feedback and support for our product.
The majority of comments, tweets, and messages were positive and we are so very happy and grateful to everyone single one of our users for trying Leave Me Alone.
The day after launch the traffic was still high and we were super happy to reach the mega milestone of a thousand upvotes. We were featured in the Product Hunt daily newsletter which generated a load more traffic and we made almost as many sales as on the actual launch day!
Then Leave Me Alone was featured in the Product Hunt weekly newsletter 5 days later as the top hunt of the week! Once again this drove more traffic and we saw another spike in sales!
We hope we can keep the post launch momentum going. If we can make $50 a day (about 10 6 month scan sales after tax) we would reach our goal of ramen profitability, which would be insane!
All the data
There are lots of improvements and features planned for the coming months! Our first focus will be performance and user experience improvements.
Then we’ll start on the big stuff; adding support for more email providers besides Gmail and rolling out an enterprise monthly subscription pricing plan for employers to clear their entire company of spam emails!
During the launch we had several requests from users to delete their accounts and their data. In the interest of privacy and transparency we have released a new profile page where you can see all of your previous scans and invoices, clear your Leave Me Alone browser data, and deactivate your account which will delete ALL of your data and revoke your API keys.
We will continue to build Leave Me Alone in the open your feedback is more important than ever. We hope that you will continue to follow our journey!
Thank you EVERYONE who supported our launch. We’re so grateful for all the love, the tough questions, the bug reports, and for keeping us sane during a dizzying 72 hours.
Why we think this launch was successful
In the past two years we have built and launched a handful of products but none of them have been very successful. We wanted our launch of Leave Me Alone to be different, so we took a different approach to the whole process leading up to, and including the launch.
Validation is key
We built the first prototype of Leave Me Alone in 7 days and asked people if they’d like to join the closed beta. We initially reached out on Twitter and in the maker community at the end of November 2018. The app was basic but the feedback for the concept was overwhelmingly positive.
Clearly there was a need for this product – beta users had validated our idea. This motivated us to continue building but we were careful not to include too many unnecessary features.
Although we kept coming up with great ideas for new things to add, we made sure that Leave Me Alone performed its core functionality really well – unsubscribing users from spam emails.
Everything else ended up on the “next version” task list. We wanted to keep the first version lean.
Lesson 1: Don’t waste time building a product without validating a need for it first.
Build in the open
Leave Me Alone is proud to be an Open Startup, which means that we share all of our metrics including users, revenue, and traffic.
This also means that we build completely in the open by continually sharing our progress on Twitter, asking users for input on important decisions, requesting feedback, and being honest about the highs and the lows. You can read more about my thoughts on what it means to be an Open Startup here.
This has proven to be an incredibly good way to develop the service. Without our users there would be no Leave Me Alone so it makes perfect sense that they should be the ones directing the product.
Some people have voiced concerns about managing feedback as the product grows – here’s one specific comment on dev.to. Right now too much feedback would be a good problem to have. Ask us how we feel about this again in a year!
Lesson 2: Ask users what they want instead of wasting time guessing and building redundant features – users love giving feedback.
Release with users
The beta users helped us find a few critical bugs we hadn’t anticipated, mostly from malformed email headers or dodgy unsubscribe links. Once we had ironed these out we considered launching but decided to soft launch first and try to get more people using the app to ensure it was robust enough for a public launch.
We posted again a couple of weeks later to Twitter and maker communities announcing we were out of beta but not quite ready to be hunted yet.
Our long list of features were prioritised and we focused on improving the existing functionality. Inspired by this blog post by Maxime the founder of PixelMe, we created a public roadmap where users could vote for what they wanted us to work on next. This was a great way for us to determine what was most important to the users, and avoid implementing unnecessary features.
We started to get more interest, and a steady stream of users. Just before Christmas we received a huge boost when two customers loved Leave Me Alone so much that they purchased gift scans for their entire companies. Thank you so much Steph and Kelly!!
We set ourselves a target to launch before the end of January, worked really hard on both the app and social media promotion, and managed to reach some incredible milestones even before our official release!
Lesson 3: Build a user base and hype on social media before launching
Over and out
We have read hundreds of articles on how to pull off a successful Product Hunt launch. Some of the tips are definitely useful, but it’s important to note that there are no magic tricks or trade secrets. A large community following can help with getting that #1 spot, but for long term success or growth the product has to be good and people have to use it!
Peace and all the love in the world!
– Danielle & James.