Database Reference
In-Depth Information
The third piece of what I do is having conversations with my peers to under-
stand where the data science world is moving and communicating what
MailChimp is doing to the rest of the world. It's about engaging in the emerg-
ing, global data science conversation. I think MailChimp is doing some things
that are cutting edge, and I want to tell people about them. Furthermore,
I want find out what everybody else is doing. This falls more under questions
like: Where are the tools moving? What types of data are becoming available?
What kinds of practices are being done? I want to understand the metaphors
and analogies other people are using to think about their work. For instance,
at the Strata Conference, I attend talks on healthcare, oceanography, defense,
etc. Where the speaker's world is not my world, but so many of the problems
are the same; these metaphors really help me.
Gutierrez:
How do you describe your work to someone who is familiar
with the math behind it?
Foreman:
Other than perhaps geeking out on the implementation of some of
our data science products, I wouldn't change that description much. Technical
folks (perhaps even more than non-technical folks) need to hear the emphasis
on communication as part of a data scientist's daily routine.
Gutierrez:
How does MailChimp think about spam?
Foreman:
First, let's talk about email. Email is a messaging standard that
exists that no one really controls. Other messaging systems, like Facebook
Messenger, which is operated by Facebook, are under the control of an entity
that gets to decide what happens on its system. Email, on the other hand, is
powerful because anyone anywhere can send you an email at any time about
anything—without having a larger entity dictate the terms of the exchange.
Anyone can set up a mail server and suddenly have the capability to send and
receive email. When someone receives emails that they do not want, we call it
spam. This has always been an issue with email and so people have dealt with
spam for a long, long time.
Fighting spam is especially important to MailChimp for two big reasons. One
is that we don't want our users to send spam and two is that big email com-
panies like Gmail, AOL, Hotmail, etc. will block our IP addresses if their users
report spam coming from our IP addresses. Potentially having our IP addresses
blocked is a big issue. We send emails for seven million users, but we don't
have seven million IP addresses. Outside of a few users who have dedicated
IP addresses, most of our users send with a pool of other users over one IP
address. If we allow a bad user to send really evil stuff, then the email receivers
will block that IP address, which is bad for the whole pool of users using that
particular IP address. So we do a great deal of work combating spam.
