Database Reference
In-Depth Information
Gutierrez:
Do you find it easy or hard to find and hire the right people?
Foreman:
I find it tough to find and hire the right people. It's actually a really
hard thing to do, because when we think about the university system as it is,
whether undergrad or grad school, you focus in on only one thing. You spe-
cialize. But data scientists are kind of like the new Renaissance folks, because
data science is inherently multidisciplinary.
This is what leads to the big joke of how a data scientist is someone who
knows more stats than a computer programmer and can program better
than a statistician. What is this joke saying? It's saying that a data scientist is
someone who knows a little bit about two things. But I'd say they know about
more than just two things. They also have to know to communicate. They also
need to know more than just basic statistics; they've got to know probability,
combinatorics, calculus, etc. Some visualization chops wouldn't hurt. They also
need to know how to push around data, use databases, and maybe even a little
OR. There are a lot of things they need to know. And so it becomes really
hard to find these people because they have to have touched a lot of disci-
plines and they have to be able to speak about their experience intelligently.
It's a tall order for any applicant.
It takes a long time to hire somebody, which is why I think people keep talking
about how there is not enough talent out there for data science right now.
I think that's true to a degree. I think that some of the degree programs that
are starting up are going to help. But even still, coming out of those degree
programs, for MailChimp we would look at how you articulate and commu-
nicate to us how you've used the data science chops across many disciplines
that this particular program taught you. That's something that's going to weed
out so many people. I wish more programs would focus on the communication
and collaboration aspect of being a data scientist in the workplace.
Gutierrez:
How do you deal with privacy and user expectations concerning
their data at MailChimp as you build data science products?
Foreman:
To start, MailChimp doesn't make its money through any means
that's at odds with our users or readers. We make our revenue via monthly
subscriptions from users with permission-based lists. That means that I don't
have to use our data in ways that trick, manipulate, or in any way suck more
money out of our customers (think ad placement). When I use data, it's in line
with what our customers want, which is fantastic.
Obviously, the in-house lawyer's a huge help. But the law isn't what it's really
about, is it? It's about user expectations. So it's great to have such an active
user base that stays in touch with us because they always make their prefer-
ences clear, more than any law would. Our users say, “This is what I think this
is acceptable and this is what I don't think is acceptable.” Which is another
great reason that we look to maximize user happiness rather than some cold
