Casey Winters is ex-Consumer Marketing Director at Grubhub and ex-Growth Lead at Pinterest. He’s advised startups from Airbnb to Reddit and is currently Growth Advisor in Residence at Greylock Partners, a VC firm supporting the likes of Facebook, DropBox, Medium and Pandora.
We spoke to Casey about his experiences with growth at Pinterest and Grubhub, and asked for some advice for startup founders today…
Driving acquisition with unique content
Casey was the first marketing hire at Grubhub – an online takeaway ordering platform founded in 2004.
Starting in Grubhub’s hometown of Chicago, Casey started exploring ways to drive organic growth. People were already searching for different types of takeaway in their area and Grubhub needed a way to make the most of that.
“The way a marketplace like Grubhub grows,” explains Casey, “is that you try it in one city, find a replicable way to grow and then try to expand that to other cities and see if it works.”
“We came up with all of the neighbourhoods, zip codes, colleges and cuisine combinations and created unique pages for all of them” Casey explains. “We would aggregate all of the existing restaurants that met the criteria on those pages so you could see all the restaurants that delivered that type of food in that neighbourhood.”
“One half of the SEO equation is about being relevant. The other half of the SEO equation is about being an authority”
Once people were ordering from the restaurants, “we asked them write reviews from those restaurants” says Casey, “which gave us more unique content.”
What Grubhub ended up with was “two sources of scalable unique content: one from the restaurants that gives you their menu data, and the other from the users giving you their review data.”
Casey says that solved “one half of the SEO equation is about being relevant. The other half of the SEO equation is about being an authority.”
The Grubhub team got the word out with local PR in every city they launched in, offering $10 discounts which could be claimed on purpose made landing pages. “Then for 30-60 days or however long we had the promo running for, we would redirect that page to one of the main SEO pages we created which would give us a lot of local links to make us authoritative in that city.”
In 2013, Casey joined Pinterest, which, while very different to Grubhub, also relied heavily on unique content for growth…
“We were working with a very high class collection of content, so we didn’t have to try to write a blog and make it somewhat interesting so that Google might be convinced to rank it”
“Pinterest has all of this content being added and described and curated” explains Casey. “People are creating boards for topics they care about like woodworking, healthy recipes, vacations etc. Some of those pages started to rank on Google. But the problem with that is that it’s only one person’s opinion on that topic. You may not have very good taste in woodworking projects or you may be a total beginner, and that stuff might not be useful to other people.”
While the quality of content is subjective, Pinterest has data on the globally most popular pins across all topics. “So we created pages for all the things that were getting searched for on Pinterest” explains Casey, “and instead of relying on users’ boards, we would create a topic page that looks exactly like a board with all the top repinned content on it.”
These pages increased traffic from Google by 20% and were higher converting than the board pages “because they had objectively better content” says Casey.
Now they just needed to figure out what pages they should create. “We looked at what people searched for on Pinterest and then we went to the Google Adwords keyword planner and said are people searching for this on Google and then if they are, how good is the content that we are servicing on Pinterest?”
The beauty of this type of content, says Casey is the quality. “We were working with a very high class collection of content, so we didn’t have to try to write a blog and make it somewhat interesting so that Google might be convinced to rank it. Users had already told us what’s cool and we just had to organise that in such a way so that search engines understand that it’s cool and relevant.”
Of course, being image based made this somewhat more difficult for Pinterest. “Google, by default, has no idea what these images are about” says Casey. “We had to work a lot to give more descriptive text to every pin that was on Pinterest.”
For example, with around 70 billion pins, it was inevitable that there were a lot of duplicates. “You and I have probably pinned a couple of the same things before” explains Casey, “but even though we’ve pinned the same thing, we might have described it differently. We decided to aggregate descriptions underneath the same images so that Google can see multiple people’s opinion on the same pin, instead of thinking it’s two different pieces of content.”
Converting organic traffic
Unique, useful content is one of the most powerful ways to acquire users, but what about converting them? Talking about Pinterest, Casey says “people were looking at all the pretty pictures and then leaving. We had to figure out how to get those people into Pinterest so they could really learn what it’s about and stick around in the long term.”
Casey recounts one experiment whereby Pinterest gated their content, allowing you to view 25 pins or so before being required to sign up. “This was a very simple project for us” says Casey. “It took two days to implement but the first day we saw a 50% improvement in conversion rate.”
A huge proportion of Pinterest’s traffic was mobile, so they had to “figure out if it was better to ask people to sign up on the mobile website or to download the app. We ran an experiment and what we saw was if you ask people to download the app, conversion rate dropped dramatically. It was a much heavier weight action. But the people that did download the app, activated at 3 times the rate than people who sign up on the mobile web.”
Once they learned that, the next question was how to get more people to download the app. ”We did the full page interstitial that everyone did for a while” says Casey, “then we changed that into a full page header. We started showing a preview of the content that you would see on the header. If you scrolled past that and clicked on any of the content that would deeplink you into the app or send you into the App Store on Google Play. When you open the app for the first time it remembers what you were looking at on the mobile web so you can get right back to it.”
The key to finding growth strategies that work for your business is experimenting, but it’s not as simple as just giving a few things a go.
While Casey has a few guidelines on what makes a good growth idea, he says “by far the most important one is how many people are going to see the experiment. There are so many growth experiments that only a thousand people see, or a 100 people see a day and it just doesn’t matter if you improve something by 1000% it’s not going to move the needle.”
“When you do an experiment you need to understand what the problem is, how you might solve it and how would you measure success?”
The first thing to consider, says Casey, is what you want to experiment on. “We have a cross functional team that works on growth including engineers, designers, PMs and analysts and we expect everyone to come up with ideas. It’s not the PM saying we’re going to do this, we’re doing that, everyone has to come up with ideas. We try to force people to come up with ideas on a weekly basis. A lot of those ideas are probably going to be pretty bad, especially when you’re new on a team, because you don’t understand growth super well, but you can practice and build that up over time.”
Another thing to consider, says Casey, is what does success mean? “When you do an experiment you need to understand what the problem is, how you might solve it and how would you measure success?”
Success should be the very first thing you decide on says Casey. “Then you pick projects that you feel like have the best chance of moving you toward that goal. You don’t do a whole lot of experimenting to get, say, 10% of the way to your absolute goal, because you just don’t know what’s going to have the biggest impact… You generate a lot of ideas and then you prioritise these ideas based on their potential.”
“One of the things that we focus on” says Casey, “is doing the minimal amount of engineering and design work… just enough to validate if it’s a good direction to go down.”
The future of growth
Casey believes that the main issue facing startups right now is the “lack of scalable growth channels… Attention has been aggregated on a few platforms.”
Google is a “fairly open platform” says Casey. “There’s a very convenient ad programme and you can rank organically. The problem with ranking organically is that it’s getting harder for startups because the incumbents have gotten pretty decent and Google is starting to take a lot more of that real estate for itself.”
Google Local for example, ensures that Google results for local terms appear first. “They are starting to do their own Google version of everything” says Casey, “and that is difficult for startups who are trying to focus on an area that is strategically being utilised by Google.”
“What I’ve seen over the last decade, is that new features never drive retention”
Casey says that the organic opportunity on Facebook has “mostly disappeared. They’ve figured out how to extract a rent from anything you want to do on Facebook. Obviously ads have worked well and page posts still work pretty well, but all of that still pays Facebook.”
“The other challenge that growth teams are facing is that there is a really high bar especially if you’re an app” says Casey. “There’s millions of apps and even if you do happen to get the distribution, then retaining users is still becoming more and more difficult.”
“The biggest thing that I’ve been talking about lately is when early stage founders try to build new features to increase retention” says Casey. “What I’ve seen over the last decade, is that new features never drive retention. Retention usually means reducing friction in the core product and adding features just increases friction.
Casey’s advice? “Look at the people that want to try your product out, and ask ‘are they understanding it?’ If not, what are the reasons they are not understanding it and how do we remove those things?”
Photo via greylock.com