Tweet about this on TwitterShare on FacebookShare on LinkedInShare on Google+


April Underwood, a founding partner of investment collective #Angels, shares more from her diverse experience in business development, marketing, and product management at companies like Google, Travelocity, and Twitter.

In our 2nd episode of Product Management Confessions, April gives advice on managing expectations, partnerships, and data, specifically:

  • Creating check-ins to evaluate both short-term metrics and long-term consequences after a release
  • Making sure to set expectations upfront when embarking on new partnerships
  • Being transparent about data and setting guardrails about what it can and should be used for

April also introduces us to #Angels and the team of talented women who comprise the investment group. Hear more in our final interview with April next week!

If you missed the first episode, watch it now:

Part 1: Former Director of Product at Twitter on Goal-Setting & Managing Expectations

Create Your Wizeline Account

April Underwood

Full transcript:

April Underwood: Sometimes you're working on a new feature, and as soon as that team is done, you've already gotten them subscribed to work on the next thing. I certainly haven't been perfect in this. I don't know that anyone is, but it's very important to leave the product manager, the engineers, the designer that have worked on a feature enough time to understand how their users react to a new feature, and then make modifications as necessary. Baking that into your road map, the expectation that you will learn things after launch and you will make adjustments is a good best practice.

Matt Pasienski: Just as a guide to others, obviously it's different for each business. How long did you feel was long enough to come in and do those check-ins on something that you'd released, and how did the designers and product managers, even the engineers come back and reevaluate, was that feature we developed successful in the long run, and how long is a long run?

April Underwood: I think that a lot of times, even getting use in one to two months after a launch feels like enough. You've got some user data. You can look at both metrics. Are people not coming back to this part of the product anymore? Are they actually using this feature? If you're on the monetization side, you can look at how much revenue is actually being generated, or the revenue metrics associated with it.

I think the reality is that usually, those time horizons really ought to be longer. I think there can be second order effects. You can launch any feature. You can introduce a new ad type into an experience. That's not going to drive users away right away. It would have to be pretty egregious. If you were to start showing ads on a consumer property, I think back to when we started showing ads on Google Finance. It would have been shocking if just suddenly, people had stopped coming to the product overnight. You have to track this over a longer time horizon and see, does it make the product less competitive? Does it inhibit the ability to offer new consumer features, because that page real estate is being taken up with an ad.

Those sorts of things, again, you have to have some gut instincts about what the real business in product and company priorities are to be able to have some confidence to continue to evaluate those things over a longer time horizon, and potentially poll things, not necessarily just because the data is not supporting it, but because it hasn't turned out to be a big enough win. That's when it's harder, putting it in a gray area.

Matt Pasienski: Yeah, when there is that going in, you're like, this is going to hurt consumer experience but it has a potential for a lot of revenue. If the revenue isn't there, then you have to come fall back. That's really interesting. One of the things that you've been really heavily involved with, I think in a lot of points in your career is the data ecosystem. You have companies sharing data with other companies. Whether it's a big player like Twitter sharing data with maybe smaller companies, or you work with, how do you pronounce it, Gnip?

April Underwood: Gnip.

Matt Pasienski: Gnip?

April Underwood: Yup, that's right.

Matt Pasienski: See, that's what I was worried about.

You have Gnip and Firehose and all these things, I think within the time you were there, you went from having partnership, sharing Twitter's data with Google to not having it, and to now turning it back on again and sharing a lot of the real-time Twitter updates with search aggregators like Google. How do you set the correct expectations for all of these partners when you're in something that's so ephemeral as sharing Twitter's data, or things like that?

April Underwood: I think what's most important is that your partnership strategy, and then the actual partner products that you're offering, whether it's an API or whether it's a series of widgets, whether it's a platform like Gnip that gives you a variety of tools for managing a dataset you're licensing, that product and partnership strategy really are best off if they're established in concert up front. That can sometimes not be the case. You can certainly have scenarios where you may actually have products that take off and start as developer products, and then they become useful and strategic partnerships, and they may take off ahead off a really clear and externally-communicated partnership strategy about why you're offering it.

You can also go in the other direction. Back at Travelocity, I was a product manager for our platform, which was a way for us to power travel websites. Powering the travel tab at Yahoo and AOL and American Express, and a variety of other sites. That was an instance where the partnership opportunity exceeded the product. There were lots of partners that were interested in having Travelocity power their travel tab, but because I've been a product manager and actually even been an engineer within Travelocity, I knew at the time that our product was not actually built with that in mind. There was a need to catch the product up to match the partnership strategy.

The ideal state is you establish those things together. Since that's oftentimes not the case, partners and developer ecosystems really value clarity and they value understanding the same things that the platform itself believes and knows.

Matt Pasienski: You say like when you started, even though in API that you're offering to other people, you treat it like a product and you're looking, what does this accomplish for the specific users we have in mind. What kind of situations does it get out of control? You build a product that can do a lot of things. Sometimes, people use it for things that you don't expect. What do you do in a situation where someone starts using Travelocity's data, Twitter's data, Google's data in a way that is antithetical to the business interests of your own company? How do you handle that situation?

April Underwood: If it's antithetical to the values, or it harms users in any way, I think we just have to put that in the category of something where it's the responsibility of the platform to take action upon that. There are instances where sometimes, intentionally or otherwise, data will be used in some ways that could potentially be harmful. That is something that I think all the companies I worked for have moved quickly to address whenever those things present themselves. I think the next category falls more into the business interests area. I think that those things are always evolving, so it's very hard to be perfect in expectation-setting with the ecosystem, with partners, and with users about what data is being shared for what purpose, and how it can be used.

I think what's most valuable is having as much transparency as possible in your partnerships. I think that's something that we did well in a lot of relationships that I was part of at Twitter, which is really just being up front with our partner about what we were looking for out of that partnership. Were we hoping that it would help drive user growth? Were we hoping that it would help drive more users to see tweets? Just being really up front about these basic things, like what are we hoping to get out of this?

Matt Pasienski: You drive them into a little bit better rails. It's like, as long as we stay in here, we're all going to be fine. If they go off-track, at least they're not so surprised.

April Underwood: Right, and it should be bi-directional. This is what we're hoping for out of this. The partner tells us what they're looking for. I think it's clear in the vibrant ecosystem that exists in the business that Gnip built that there's a ton of opportunity to be there. There's actually such a giant opportunity, particularly with Twitter data, that you can see that there are hundreds of companies that are licensing data directly through Gnip, and solving a lot of business problems with those. Those are really interesting, because they I think have historically gotten a little bit less attention when people think of the things that are happening on the Twitter platform. There's a lot of focus on the consumer experience around it. The Twitter data can be used to predict a lot of things that help businesses be better at what they do. That ranges from manufacturing, to media, to data related to the upcoming elections. There's an unlimited number of applications of the data. A lot of that stuff happens, and people don't even necessarily know it. It certainly isn't as much what is talked about when people think about the platform.

Matt Pasienski: You can get Twitter predictions in your Bloomberg console now, things like that.

April Underwood: Yeah, that's absolutely one of the partners.

Matt Pasienski: I want to make sure we get a chance to talk about you and several other Twitter executives have formed this thing. Help me pronounce it. Is it hashtag Angels, or is it like Angels, or-

April Underwood: I always like to do the hashtag hand signal. We just call ourselves Angels, but certainly, hashtag Angels works as well. It's really exciting. We're brand new. We just announced the fact that we're effectively working together as an investment collective is the way I would think of it. Each of us was already doing some amount of individual Angel investing. We've had this really phenomenal shared experience at Twitter and being on the leadership team and in a variety of different functions at Twitter. We kind of realized sitting around the table that we have such a diverse functional perspective. We've got a legal perspective, media perspective, business development. I'm more on the product and engineering side, corporate development, that we want one another's advice as we're thinking about the deals that we're investing in. We also want to be able to offer the expertise of this network that we have to the companies that we back.

We already were doing this to some extent informally, so we decided to tell the world that we were doing it and help founders understand that we have something to offer that is somewhat unique. It's been really great to see the reception. We feel like we definitely tapped into something, given the amount of interest that we've seen coming in.

Posted by on Saturday, June 20, 2015.


Leave a Reply