People hate to register, some follow-up thoughts
This blog post could also have been titled: how I've lost hours listening to a vocal minority of my users for close to no results at all.
To summarize I argued that people, when it came to non critical data (like a sport prediction website) much prefered the old fashioned secret URL generated for each user and optionally sent via email, to the process of registering and especially clicking on the deadful "sign-in with [social media]" buttons.
And boy was I right... plenty of users came to the site, plenty of nice things were said about the log-in system and everything was looking good. As a reminder, nospronos.com started as a side project with my friends who wanted to do friendly predictions about the worldcup with no money involved.
Anyway, after the first few days of the worldcup, I started getting lots requests for new features, which was good, I really love feedback, I even ask for it pretty often. One of the most demanded feature was the possibility to create a real account on top of the secret URL. I found this weird given that I first "sold" the website on the opposite concept and knew it was going to be a good chunck of coding to get it going nicely alonside the secret URL scheme, but I went for it anyway and... *tada*... a few days later I unveiled the possiblity of transforming each secret URL into a real login/password account, with a homepage, a proper username and all that.
There is around 7000 users who made predictions about the worlcup on the site and around 2000 unique visits a day. I must have had around 30 feedbacks asking for real accounts so I though that given that most people never actually take time to send feedback, at least a few hundred accounts were going to be created...
And boy was I wrong... 30 feedback directly asking for accounts = 30 accounts created, even if you stick the option right in the middle of the landing page of most users.
Same story happened again with Wimbledon, a few users sent me messages asking to replicate the friendly prediction league system of the worldcup for Wimbledon (starts today), so I did and it was a bit of a pain ! And again out of the 20~25 requests, I ended up a few days later with 20~25 people making Wimbledon predictions out of the 1000s visitors per day !
There are a few take home messages I think:
Users came to your site for a reason or a receipe that you did right somehow, so stick with it.
Users have an extremely high inertia and it's not because they are somewhere on your site, that you will get them somewhere else on your site
Feature requests have to be normalized by the total number of daily users on the site -> the users who remain silent are not too lazy to ask for something more, they just don't want anything.
Spending time developing something that people love is an amazing feeling, spending time coding something that people end up ignoring is excruciatingly tiring, so rushing into a new feature is a bad idea.
PS: When I don't post on HN or create websites, I'm a finishing PhD student in statistics/computer science/bioinformatics @Cambridge (UK) looking for the next step in his life, anybody recruiting somewhere in the world? Contact me