Of course, there are limitations to the power of the content model. As Winters points out, neither Pinterest nor Grubhub employees were primarily responsible for driving the content value or volume, as the demand and supply sides of their products took care of it. \u201cIf you try to create everything yourself, it\u2019s just going to be hard to compete with a company like Pinterest which now has a 100 billion pins,\u201d he says.<\/p>\n
But for Pinterest, the content was just the right mix, and the introduction of this loop unleashed a second wave of growth, which was a much-needed follow-on after the first wave had flattened out following Facebook\u2019s Open Graph platform rules change. The company has since found that this new wave of users still drive significant revenue and have been more active, most likely because they saw the value of Pinterest for discovering content, not just as a platform that has their friends on it.<\/p>\n For the startup seeking to create a content loop, here\u2019s the the 5-step process that Winters suggests to assess the opportunity and get started:<\/p>\n Find ways to get content from your users.<\/strong> Content is unsurprisingly the key ingredient of any content loop, so identify the kind of content that will drive your model, and be valuable for users and non-users alike. \u201cAsk yourself: \u2018Do I have some sort of asset that’s being created that I can lean into? Is there a natural way the product can create this content or not?\u2019\u201d says Winters. \u201cIt\u2019s okay if there isn\u2019t content being created today. There\u2019s a lot of low-hanging fruit to go after. You\u2019d be surprised how powerful a simple review is. Or building specific pages for geographic areas. It doesn\u2019t have to be rocket science.\u201d<\/p>\n Give users an incentive (and a mechanism) to share that content.<\/strong> Early on, make the process of sharing of content is as frictionless as possible. Winters points to YouTube and Musical.ly<\/a> as stellar examples of seamless sharing. YouTube allowed users to embed their videos on any website or app, driving users to sign up and encourage their friends to follow their content. And after noticing teens filming themselves lip syncing to songs online, Musical.ly made it easier to film and share videos inside and outside of the app. \u201cThink about your users and where they\u2019ll be sharing. For example, in different countries it may be WhatsApp, so you have to support that,\u201d says Winters. \u201cAnd if your users aren\u2019t motivated to share, see if you<\/em> can share it in a way that makes sense. At Pinterest we distributed popular boards to Google search and at Grubhub we distributed the information we got from restaurants to customers.\u201d<\/p>\n Find where your community lives and double down.<\/strong> After putting the content and sharing tools in place, startups need to determine which communities will drive enough viewership of their content and then find a way to maximize them. \u201cYou’re trying to figure out who the audience is. What is their intent? Where do they live online?,\u201d says Winters. \u201cIf your audience intends to be seen as thought leaders, maybe they\u2019ll go to LinkedIn<\/a>. If they\u2019re engineers, they may go to Hacker News<\/a>. At Pinterest, we found that our audience was going to Google, so we built up a team of engineers focused on SEO.\u201d<\/p>\n Trace traffic back to the source to tweak your product.<\/strong> When looking at users streaming in, entrepreneurs should break the traffic down by source to match user intent with the value of the product. If users aren\u2019t converting from a particular source, it may be because they are a bad fit for the product. But with small changes, they may become a good fit. For example, when Winters\u2019 content loop campaign at Pinterest initially started bringing in more Google search users, his team found that they were converting at lower rates. That was because Pinterest\u2019s onboarding flow focused on surfacing content from their friends that were already on the network, which didn\u2019t match up with their intent. \u201cIf users searched for Chukka boots on Google, landed on a Pinterest board and signed up to see more, their feed focused on content from their Facebook friends. But that typically had nothing to do with Chukka boots \u2014 not exactly the ideal experience,\u201d says Winters. \u201cSo we changed the onboarding flow to show topic recommendations instead. And we saw an increase in activation rates. This also helped with another problem we had, which was a growing gender imbalance. We didn’t think there was anything fundamentally female leaning about the Pinterest platform, it\u2019s just that our early go-to-market community happened to be midwestern moms,\u201d he says. \u201cSo when men signed up, 90% of their feed was content targeted to females. But when we switched to a topics approach instead, men were able to immediately connect with the content they cared about.\u201d<\/p>\n Convert, activate and add some friction to feed the loop.<\/strong> Converting those who see the content is a huge component of the loop. \u201cYou need to make it sticky. If people came from Google to Pinterest and didn\u2019t sign up, that’s fine, but at the very least they should leave with a better understanding of how Pinterest can help them in the long run. So we were constantly experimenting with ways to unlock that,\u201d says Winters. One of those experiments involved introducing a roadblock. \u201cIf you came to Pinterest from a Google search, we would show you a bunch of related pins, but after you started scrolling, it would actually block you from scrolling more until you signed up. It worked very effectively. We threw that thing up as a two day hacky experiment, but kept it,\u201d says Winters. \u201cBut if you\u2019ve created value, you can feel comfortable introducing some friction in order to showcase more because it\u2019s the higher quality users who are going to withstand the friction in order to participate.”<\/p>\n If your growth team has mastered the basics, there are a few extra tips that Winters has for startups seeking more advanced techniques with their content loops. Here, he shares three recommendations that can help companies chart course for the deeper waters of content-driven growth.<\/p>\n Experiment to accelerate learning, not to find a silver bullet.<\/strong><\/p>\n The loop won\u2019t be effective if visitors aren\u2019t converted into users that actively engage with a product. And in Winters\u2019 experience, that takes frequent \u2014 and failed \u2014 experiments.<\/p>\n For example, Winters and his team ran an experiment with Google users that skipped the new topic picker during onboarding entirely, instead building a content feed based on the user\u2019s original Google search query. The thinking was that specificity would increase activation rates, but in fact the opposite happened. The query-based feed approach helped users find what they were looking for, but it didn\u2019t encourage them to return to see new content. \u201cThey had a great first experience and found precisely what they wanted \u2014 they just never came back,\u201d says Winters.<\/p>\n Iterating to surface actionable insights is key to fine-tuning the cogs of the content loop, so founding teams shouldn\u2019t hesitate to dive in and start tinkering, even if they are resource constrained. \u201cThe amount of experiments you run is limited by how long it takes to learn, not by how big your engineering team is. Experimenting on search engine traffic could take about a month to glean insights and measuring activation rates can also be a slow process. Sometimes there\u2019s no way to speed up the learning,\u201d says Winters. \u201cOther things such as conversion optimization could be tested in a couple of days. And it\u2019s fine if you\u2019re working with a smaller number of users, because in the early days of a startup you\u2019re just looking for big changes. When it comes to experiments, don\u2019t shoot for victory or failure. Instead, ask: \u2018What did I actually learn from this that helps me build the next thing?\u2019\u201d<\/p>\n Vary strategy and metrics for each segment of the content loop.<\/strong><\/p>\n Once a content loop is set in motion, the next step is to measure how it\u2019s working (or not). And according to Winters, the key metrics to pay attention to will vary depending on the stage of the content loop. \u201cWhen you\u2019re focusing on the start of the loop, you\u2019ll want to increase traffic and then conversion rates for new users,\u201d says Winters. \u201cLater in the loop, your focus should be on increasing activation rates to ensure that these new users won\u2019t drop off.\u201d<\/p>\n Content isn\u2019t just about getting people to your product. You have to stop people from landing and leaving.<\/p>\n For the founder that\u2019s not sure where to start, Winters advises spending your time on the slowest moving part of the loop, zeroing in on what\u2019s dragging and concentrating firepower there. \u201cAt Pinterest, the content loop strategy initially was all about boosting top-of-the-funnel traffic<\/strong>, because our original problem was that we weren\u2019t getting enough new people. But once we got more traffic in from Google, we realized people weren\u2019t signing up, so we honed in on boosting our conversion rates<\/strong> specifically from Google. Then it became about driving sharing behavior<\/strong> from these new users, so we were looking at how many new pages were generated that Google could index and the traffic that those new pages received,\u201d he says. \u201cBut to take a step back, the key metric isn\u2019t signups or traffic. The ultimate goal for a business like Pinterest is just engaged users<\/strong>. How many people are you actually providing value to? We measured that by looking at the repin rate because it also drove the quality of our content loop.\u201d<\/p>\n Lean into SEO to build authority and ignite long-term growth.<\/strong><\/p>\n For many teams just getting started, SEO can be an intimidating tactic<\/a>. It seems competitive, as though you\u2019ll be drowned out by more established competitors with bigger purse strings. But for Winters, it was an essential building block of his scrappy growth strategies.<\/p>\n \u201cMany entrepreneurs think SEO is scary black magic that\u2019s tough to decode. But it\u2019s actually more understandable and sustainable than other platforms. You need to set the foundation for good SEO in the future immediately, because the sooner you start the better it’s going to be,\u201d he says. \u201cI always tell my clients that there is a kindle strategy<\/strong> and a fire strategy.<\/strong> The kindle is whatever helps you grab those first groups of users so you can validate product\/market fit. But then the fire is how you actually scale<\/em>. Too often people confuse the two. While SEO is a fantastic fire strategy, it’s hard for it to be a kindle strategy, so you need to keep that in mind.\u201d<\/p>\n For Winters, it comes down to remembering that SEO is a different way of thinking. \u201cWhen you\u2019re creating content and looking at SEO, you have two clients essentially: the user and the bot. It\u2019s a common misconception that you\u2019re building features or writing content purely for SEO, when in reality you should be building a feature because it’s good for the user, but then it also has a by-product of being good for the search engine, which in turn makes it good for new users. For example, we built reviews at Grubhub so that people had a better understanding of which restaurant they order from, but it also worked out as relevant unique content that helped us rank on search engines as well.\u201d<\/p>\n More specifically, Winters notes that Google\u2019s algorithms care about two things:<\/p>\n Relevance<\/strong>. Does the content match with the problem the user is trying to solve when they search? To improve on this, Winters recommends including enough text on your pages to explain what your content is about. \u201cPinterest\u2019s pages used to be all images. Google had no idea what any given content was for, so it didn\u2019t rank it,\u201d he says. \u201cWe got increasingly better at aggregating all of the descriptive content across Pinterest into a focused area on these pages, both in the metadata and on the pages themselves.\u201d<\/p>\n Trust<\/strong>. To build trust with Google search, you\u2019ll need to build up your domain\u2019s authority by increasing the quantity and quality of external links that point to your pages. Solving for trust is the trickier end of the equation, as it largely comes from media mentions. And if you’re an early stage startup, typically you’re competing with an incumbent that could have decades worth of links pointing to it. \u201cWhen Grubhub launched in new cities, no one was using us and we had no authority. And Seamless had actually been on the web since 1999<\/a>. So we had to think of a scalable strategy to get links quickly,\u201d says Winters.<\/p>\n Here are his two top tactics for startups looking to punch above their weight on SEO:<\/p>\n Reach out to influencers.<\/strong> \u201cWhen we were launching into a new city at Grubhub, I went to a lot of local bloggers and newspaper sites and offered them a promo code for their readers in exchange for a write up. So that helped with our authority pretty quickly,\u201d Winters says. \u201cSince then Google has changed their guidelines, so now I recommend seeding content to receptive writers and browsing sites where they learn about new products, such as Product Hunt<\/a> or Techmeme<\/a>.\u201d<\/p>\n Pitch your data.<\/strong> \u201cStartups often don\u2019t realize how the data they\u2019re collecting might be useful for getting press,\u201d Winters says. \u201cHow many people are doing this weird trend on your app? That should be a scalable story every month. At Pinterest, we would curate the most popular trends around DIY or weddings on a regular basis and release these \u2018what\u2019s hot on Pinterest\u2019 trend stories.\u201d Winters suggests scaling out data-driven stories by geography. For example, a jobs site could write about job trends in different locales and distribute them to local papers. Or take the case of Hipcamp<\/a>. The campsite-booking site created content about the solar eclipse and distributed these stories to areas in the path of totality, where viewers and die-hard eclipse chasers could see the full show. Local papers in these areas, as well as national publications, picked up their content and linked to it.<\/p>\n Funding rounds aren\u2019t even a shoo-in for press these days. There\u2019s something intriguing in every data set. Find it and pitch novel trends to get links back to your site.<\/p>\n Winters is bullish on content loops, not only because of their results throughout his career, but because of the issues he\u2019s seen companies bump up against when using other models.<\/p>\n<\/figure>\n
THE DIY GUIDE TO GETTING A CONTENT LOOP IN MOTION<\/h1>\n
HOW TO DELVE EVEN DEEPER INTO CONTENT LOOPS<\/h1>\n
THE REST OF THE LOOPS: OTHER APPROACHES TO GROWTH<\/h1>\n