from Hacker News

Ask HN: How would you solve the spam problem at Gmail?

by zatkin on 12/5/19, 6:02 PM with 7 comments

Imagine for a moment that you are on the GMail team and you need to solve the problem of deciding whether an email intended for one of your customers (@gmail.com user) is going to go to the Inbox or Spam folder.

The world is full of actors - both good and bad - who are hungry to get their message into Inbox.

So how would you ensure only the good actors land in the Inbox?

  • by pacificenigma on 12/8/19, 1:57 AM

    You start with SPF, DKIM, DMARC and IP reputation.

    But then I'd use the moat GMail enjoys as a primary mail destination by statistically looking at the frequency an incoming email's metadata and content similarity has recently arrived at GMail as a whole, then moving suspicious arrivals into a progressive exposure pathway that tentatively delivers a small percent to known currently-active users (eg mobile GMail app open on screen and unlocked) and see what percentage are flagged as spam, archived or deleted. Then use that real time feedback to vary the delivery flow of remaining messages to inbox vs spam.

  • by gt565k on 12/5/19, 6:12 PM

    First and foremost you want to make sure SPF, DKIM, and DMARC are valid from the sender's mail server

    https://www.smartertools.com/blog/2019/04/09-understanding-s...

    Then, email filtering companies use things like IP Address of the sender, keywords, and NLP, and other tools to determine if the email has spam and assign it some kind of score to it.

  • by quickthrower2 on 12/5/19, 7:40 PM

    You can use the + feature to generate different email addresses and then chuck anything coming to the root address. That way when you get spam on a particular address you just block that one address.
  • by buboard on 12/5/19, 7:00 PM

    Apart from a preapproved list of contacts, I would ask for a fee to deliver each message and only high bidders get to pass. Which is more or less what they re doing
  • by gesman on 12/5/19, 8:01 PM

    Perfect task for AI. Especially considering that Google has more resources for it than bad actor.
  • by joegahona on 12/7/19, 8:14 PM

    This sounds like a product-manager-interview question.
  • by gtirloni on 12/6/19, 2:30 AM

    Interview question? ;)