Nicely done, and excellent use case for pg_query :)
We essentially run what is called a “fake door test” — you show a button that seems as if it does the action, but when people click it instead shows an error message (and in this case suggests to use Twitter instead).
This gives you a good feedback if people were to even click a button, before building it all out.
Hope that answers your question :)
I’d never turn off learning, as in listening to whats happening and incorporating that into your thinking.
However, I think often startup founders and product leads try to take the next step too quickly (and ideating through all the variations), instead of focusing on the current task at hand.
There is more that is not mentioned in the article. You are right that Login itself is well-documented and easily accessible.
As said our Twitter integration was quite in-depth, and we’re adding more Facebook integration points every week on top of their platform :)
We ran the test on the login screen — so basically it would show both options for a few days, and if you clicked the FB button it would tell you that the option is not available yet (+ send a tracking event), and direct you to Twitter.
cc Chris Fohlin Jeremy Gollehon Kilim Choi who had the same question :)
Yep, we use Mailchimp for the Daily Newsletters, and Mandrill (also by the Mailchimp team) for delivery of other emails sent from the app.
Works pretty well, and in the rare circumstances where there have been issues, the Mailchimp team has been very helpful and responsive.
This might be because you are sending with a “@gmail.com” address through Mailchimp.
Gmail’s SPF records (for “gmail.com”) have a “soft fail” for everything thats not their own servers:
v=spf1 include:_netblocks.google.com include:_netblocks2.google.com include…