I worked for a Groupon clone for a few months in the past, I can guarantee that email confirmation is an important step since many users will mispell their emails and won't be able to log in again and get really mad at customer support.
This article is like all the other "I don't understand why others do what they do so I'm going to say it's stupid".
It sounds like in the proposed architecture, an email confirmation is still needed, but not right away. That is, a user can immediately start using the product without verifying their email address, but will need to do so eventually.
The million dollar question is when to require confirmation. The proposed answer (based on functionality/privileges) seems destined for failure. What if a user spends hours using your product, then comes across a "needs-confirmation" feature, and can't move forward because they mistyped their email on signup.? Or leaves and comes back a day later only to realize they can't login?
Maybe the solution is to give them a taste of the product "for free", then require email confirmation after X number of minutes (30? 1 hour?).
The trick is balancing not punishing people who enter everything correctly while ensuring people can recover from a (pretty common) error.
Thanks for your feedback. We did think about all the points you mention and are very aware of the issue of typos. I also think you might want to consider the kind of business you are in before implementing such a flow. A Groupon clone clearly doesn't attract the same users as a B2B SasS.
This article's goal was to share an experiment. So far, it works for our users. The day we crawl under the requests of people making typos, we will figure something out to solve that. Even then, there are ways to fix the email on an account and make sure the user gains access back to her work.
As to giving a taste for free, this is in our roadmap and will be the subject of another article when we get to implement it.
At low volume and for the right use case it totally makes sense, I wish you wrote that instead of the blanket statement you used in the article. Other than that, it would have been a LOT more useful if you published the stats for the previous designs and see how things have changed.
To be honest, if I knew it would drive that many comments and reactions (see Reddit), I would have named it differently. Our goal wasn't to make a blanket statement at all but to share our experience.
We have a post about the data in mind. We are still working on the signup and onboarding flow and are waiting to have all the learnings to share them.
I'd love to see it in action but my first reaction is that it would confuse users even more than needing to confirm their email.
But love that someone is trying this out and trying new things to make the onboarding better.