Anna Smith - Data Scientists at Work - page 204

questions like: Do they notice when we launch anything? Do they notice when
we take away some functionality? Luckily, we have friends and fans that talk
to us and are very vocal about their thoughts. While sometimes painful, it's
great to hear feedback like, “That was crap. You guys need to put that back up
there.” But, overall, it's really easy to get excited about something when you
like the service or you enjoy it. And I like the idea that it's very empowering
to women. It's all about democratizing fashion and making it more accessible.
It's not about celebrating the material part of it, but it's about you wanting to
look good and have a great night. And everyone should be able to do that.
Gutierrez: How do you pick projects to work on?
Smith: Interest and ability to persuade others that it's a good project. A great
deal of my work here has been in support for other people's projects. For
instance, one thing I've worked on is research into the recommendations sys-
tem. They built the recommendation system and it's been running. Now I am
doing the research into how it's actually working and if it's actually working.
Many of the projects end up being formulated this way. I think of an idea or a
different hypothesis or assumption than what we are currently doing, and I go
and test it. Then I present the data and we discuss the findings. From there we
can figure out where to go next. I've always been told that it's better to ask
for forgiveness than for permission. So I think that's a lot of what data science
is about. We need to have the freedom to explore on our own. I don't know
if it's 80 percent free time, but maybe 40 percent, definitely. We need to have
monkey time to get involved in something, get really excited about it, and then
still make sure you get your other work done and deliver on time.
Gutierrez: How do you measure your success?
Smith: I tend to measure success at work on project-based metrics. Even
though that sounds straightforward, I think that's very hard because it's not
just whether you finish something. It's also about whether all along the way
you were making sure that what was coming out was representative of what
was going in. It's also about whether you did the project well and whether you
were able to finish something that you could communicate to other people.
The success metric here is much more about convergence. The questions we
measure ourselves against are much more in line with the business. Which
leads to questions like: Did we get more people by buying or renting? Did we
get more people coming through and using it? Did they actually use it? Were
they having a positive experience? Those are more of the metrics that we
look at here, which is very different from the ways I'm used to thinking about
it, where I'm like, “Look! It's awesome!”