Database Reference
In-Depth Information
For us, it's about making a better product, which means that we have to have
lots of little ways of assessing whether that's happening. Some questions we
look at are: Are there places where there are errors? Is there feedback we're
constantly getting where people really want something that they don't have?
And is there feedback that we're getting where they're super, super happy
and they're over the moon that they can do something? So it is that bigger
picture—which I think is a really important piece of data science—that we
look at.
I think there's been a lot of focus on data science as kind of the optimization
piece. And I think we're definitely much more on the kind of exploration and
big-picture piece. So we focus on: Are we even telling the right stories? Are
we even looking at the right data? You can't really optimize for that, so how
we measure and view success is a bit more product-y and product design-y,
where it's hard to prove apart from seeing the output of people really wanting
to use the software and actually using it.
Gutierrez:
How do you view and measure your own success?
Heineike:
I'm very driven by wanting to build something that I think is useful
and important. When I can log into the product and use it to learn something
new, that's personally very satisfying. When I see our users using the tools to
explore important global issues, it is really satisfying. In my day to day, I value
being continually challenged and to be learning something new. I want that to
translate into real improvements in what we've made. I want to see others
empowered by what I've made.
Gutierrez:
Where do you get ideas for things to study and analyze?
Heineike:
It comes from a few different places. Some of it comes from
end-use cases, where people are saying, “Oh, I really want to answer this
question.” And then we might be thinking, “We can let you answer that if we
built an additional feature.” Sometimes it might be that they're saying, “Oh, this
thing doesn't really work the way I want it to,” and then we might be thinking,
“We could make it better if we improve the tokenization.”
And then sometimes it's we're looking at the data and we're thinking, “No one
else has probably thought of this, but just by doing this and this and this, we
could make something that people would want.” So there's this interesting
balance of piecing those things together and choosing what to do. I think some
of the choosing what to do comes from us as individuals, and some of that's
coming very much from talking to other people in the business and figuring
out what they want, and from the product management process. We have a
very long list of things we would love to do.
