One of the compelling sessions I attended at GHC15 was the Beyond Growth Hacking Session.
Panelists were: Miri Pruzan(Facebook),Winnie Wang(Airbnb), Erica Lockheimer(Linkedin), Molly Long(Uber), Vivian Qu(Pinterest).
(Notice that I have used the colors of the panelists’ shirts in the photo to highlight their names accordingly. I’ll use the same colors in further mentions of their names).
Hosted by the Five Panelists above, the session started on quite an agile note!
I think presenting this in dialogue format will make it more interesting for you. So, here goes…
In the last few years there has been an increase in people googling the phrase “Growth Hacking” – Miri Pruzan
Q: What does a growth engineer do?
Winnie: – Increasing a company’s top line metric. A growth engineer uses A/B testing to find the best user experience and subsequently helps increase top line metric.
Erica: Strategic areas in Growth Hacking include: Registration on boarding.
- Onboard Users, help them complete their profile.
- Help them create a network
- Talk to your users.
Q: Why do we even need a growth team?
Miri – At Facebook, One of the issues the growth team is focused on is what users actually do on the site.
They track users’ behavior, e.g drop off pages.
Molly – Product team focuses on making the product (pie) more delicious e.g at Uber(reducing ETA-Estimated Time of Arrival ) while the growth concentrates on making this product(pie) bigger.
Vivian – Three years ago at Pinterest we saw our growth metrics go down and thought we should create a growth team. More than about seeing your metrics go down, a growth team owns the services and frameworks that allows you to sequence when the product education goes out.
Also, We run a ton of A/B experiments, they can have a central repo where people who take over can go to catch up.
Growth teams are necessary for things like trying out projects e.g the invite-only Pinterest we tried in the Philippines- you need a growth team to own that and monitor the numbers and see if the project is a success.
Q: What metrics are we trying to move?
Miri : At Facebook we’re very interested in user retention.
Vivian: We originally tracked monthly active users. about 1 yr ago, we stepped back to make sure we were monitoring the right metric to track.
We chose a metric that most aligned with the product experience. You have to make sure to track the right metric which is aligned with the product experience because this metric will be optimized and you do not want to end up tracking the wrong thing.
Erica: – It’s really about sign up, but we started changing it with something we call quality sign up. Quality profile information, Reachable info, Incremental connections are things used to measure this. Also get them engaged within 28 days. We also measured invites from friends and engagement.
Winnie – At Airbnb , the most important metric was the number of nights booked.
We shifted our focus to intention to book. We also shifted to using registered users as top line metric.
Of course, metric used will depend on the context.
Molly– From supply side, number of driver signups per week. Another key metric we are starting to look at is viral factors.(Exponential growth)
Erica– The thing that is so important to growth teams is monitoring. It is very important to have the right monitoring in place.
Miri– It is not just about the metrics going down, it should be possible to explain why metrics go up.
Miri– Sometimes you take a hit at metrics for long term effect. E.g Forcing Facebook messenger on people. made people mad, but we were trying to provide a better messaging experience. A year later we see it as a success.
Q: What is the difference between growth hacking & growth engineering?
Molly– I see hacking as more about experimentation & iteration. Doing many things and finding out which works.
while growth engineering is all about owning these growth practices.
Vivian– I disagree with you(Molly) on growth hacking. It has a negative connotation . It sounds like being deceptive and growing users by selling a product in a way that it is not. The second negative connotation is like writing bad code to get your numbers up, but I find that the technical debt incurred is a lot and not worth it.
Molly– I think there are little tricks you can do to get your numbers up e.g at Uber we changed a “share with friends” text to “free rides” on the share panel and it increased click through rate by 100%
Winnie: growth in general is all about empathizing with the user.
Miri– (We should rename Growth hacking to growth experimentation) *Laughter*
Q: What are different approaches taken to grow products?
Molly: In paid marketing for instance you don’t want to be too dependent on cash to grow or to be in the negative. At uber we are banking on viral factor. e.g Referrals. which will increase exponential growth. and reduce your dependence on cash when this is optimized.
Erica: Focus on loops, making smart suggestions to users which leads to smart networks. Focus on nurturing. This helps the viral factor.
A Q&A for the audience followed after the main discussion and here are some questions covered.
Audience: How do you best communicate with teams that are important to growth?
Erica:We have tools that track experiments and how they impact growth.
Vivian: Data engineering team works closely with growth team to tackle these things.
Molly: We work heavily with city and operations team.
Audience: How do you use push notifications to improve user engagement?
Erica:Build intelligence around it to make it relevant to users, and timing.
Miri: Having a cross functional product helps. Then on some platforms for instance, you do not push all notifications.
Audience:How has rebranding & redesigning (e.g of logos) impacted your metrics?
Winnie – It shot up metrics quite a bit. But you have to be really careful as the growth in metrics is usually short term and after a while the numbers drop.
Erica: It’s a scary thing because sometimes you can’t predict the outcome but you have to monitor and iterate accordingly
Audience: How long do you run experiment to determine success or failure?
Erica: It depends. You have to understand your product to determine that.
Winnie: It depends also on how your experiment is set up, some experiments take longer to converge results.
Vivian also added that one has to be very careful which experiment is run.