That's sometimes true, but I think (like mentioned in the article) a combo of a knowledge base with clear ways to contact a human usually leaves a good taste in my mouth.

We use Django REST and an angular frontend. Django is hosted on heroku, and angular is hosted on netlify. Netlify makes it super easy to proxy API requests from the front-end to our API backend.

Maybe it's time to update that anti-fascist deck...

That's really interesting and definitely something worth trying. I wonder how the effectiveness of email changes with the goals for the user (play solitaire vs pay for a subscription, etc).

I think age plays also a role. I have met various young people who have't entered the "working world" yet, that only have an email address because they needed one for registration and to recieve some school related information. But it is percieved as something completely arcaic and they never use it for anything else nor check it regularly. All communication is via messangers or social media.

On the other hand also elderly are or could be moving in this directions. E.g. my mother getting an iPad and apps like Whatsapp have completly changed "computers" for her. From, being mostly a chore that you have to use, to something really useful for getting information and to stay in touch with relatives in other countries

He does mention, somewhat off-handedly, that the lack of email makes password recovery harder. (Presumably impossible without some other communications channel like a phone number.)

And you really need to be able to recover account access for a paid subscription. It's probably also reasonable to assume that if someone is going to give you a credit card number and address, they're probably OK with giving you an email.

I run a site that takes payments for a subscription, but then just stores a cookie on the users machine proving they've paid.

It will give them the cookie again if they re-visit from any IP they've previously used.

It also re-gives them the cookie if they try to pay again with the same credit card.

Support just tells people to try to resubscribe if their subscription has 'vanished' - but it seems to happen to very few customers.

This is pretty clever, but people might get double billed if they accidentally try to confirm their account with a different card than they used to sign up.

Public IP as an auth token seems like a horrible idea.

You're giving anyone on CGNat or even the same coffee shop access to your customers account.

In my case, customers don't have any data on the account - it's simply a bit saying 'has paid for premium?'. And if I end up giving premium to a few people who didn't pay it isn't an issue. The sign-up friction of needing an email address is greater.

How about you put the recovery code onto the invoice / credit card line?

I would suspect brand too makes a difference in whether people will give away their email address.

Our startup uses a combo of Google Analytics, Stripe, and our own database for important metrics. Any plans to be able to pull directly from databases (we use postgres)? Thanks, seems like a cool product.

That would be great, not sure where that fits in our roadmap, but I'd love to be able to do it.

Just wanted to say that even though you said you're not a native English speaker, I never would have known if you hadn't mentioned it. Very well-written and clearly organized post.

I take the article to be written in the mindset he had at the time. When he made that quote you mentioned, he thought as he was heads-down building the product that it would be useful to those people.

He then goes on to describe the evidence specifically why doctor's DIDN'T want it. All the author was trying to do with that quote was explain what he was thinking at the time.

Couldn't agree more. The first interview is usually with a non-technical person (like HR), so at the first round candidates are often rejected for things that have no real bearing on job performance.

I've had multiple calls where the interviewer obviously didn't look at my resume beforehand and rejected me for not having enough years of work history or previous experience of a specific technology. Both are fair reasons but it's insulting and representative of how little they respect engineers' time.

We do the same, first non-technical interview, and we reject most candidates due to the cultural fit issues. It’s too expensive to involve dev resources into the process first. In our process: if you are friendly, sound smart and enthusiastic, you get the tech challenge as the next step.

Fair enough. In that case, it still takes time but I imagine the opportunity cost isn't quite as high.

That's honestly something he haven't considered. We generally tend to believe that ideas aren't worth much on their own, so we're happy to talk about our company/experience with anyone.

At our size, the downsides to protecting most of our ideas are outweighed by the value we get from practicing our pitch.