Three quarters after Yummly launched, its Chief Growth Officer Ethan Smith watched the food discovery platform log its 10 millionth visit in a month. However, like most enduring recipes, it didn’t become a crowd favorite by tossing together ingredients in a fit of inspiration. Yummly’s SEO strategy, which helped the startup top the charts as the best global recipe app for iPhone, iPad and Android, was years in the making, going back to Smith’s first company.
Smith started his career as a user experience designer and researcher. At Wize, he managed product, design and marketing, until Nextag acquired the product review startup. Since 2010, he’s held various roles at Yummly, including its Head of Product and VP of Growth, helping it tally over 20 million monthly visits. Beyond Yummly, he actively advises startups such as Wanelo, Vinted and Thumbtack on strategies to improve growth and SEO.
In this exclusive interview, Smith shares why a specific inflection point with SEO has led to costly misconceptions and missteps. He outlines how to recognize and sidestep them as well as identifies key tenets for high-performance SEO. For any company that has lost confidence in or hit the wall with SEO, Smith’s tactics can help retrofit and reenergize your strategy.
One reason the concept of product-market fit resonates is because the balance and reliance between the two elements — product and market — is front and center in its name. But when many startups think about SEO, minds often jump to sly maneuvers and hacking a way to growth. The truth is that optimization is only as strong as one’s understanding of the search engine part of the equation. When the rules of search change, so must the strategy. Over the last five years, Smith has seen startups neglect a shift that has changed a big part of SEO.
“A new era for SEO began in early 2011, when Google launched Panda, its change to its search results ranking algorithm. What was once a very simple algorithm was revamped to penalize low-quality, thin sites. Gone were the days of a formula based mainly on keyword density and pagerank,” says Smith. “In the SEO of the past, one could repeat and conceal a bunch of keywords in a tiny font with camouflaged text to increase word density and pagerank. A SEO manager’s job was to find holes like this and exploit them until Google would slap your wrists and patch the gaps. Because the penalties were short-lived, people would go back to finding and exploiting another loophole as soon as they could.”
Together with Google Penguin, which penalizes artificial boosts in page rankings through backlinks, Panda brought in somewhat of a new world order with SEO. “While most of the workings of the Penguin filter is public, there was — and still is — more unknowns with Panda. For example, if your site is penalized by Panda, it could take years to get released from it and it’s not obvious how that happens,” says Smith. “Here’s what I’ve seen occur over and over: Founder starts a company. He’s not an SEO expert, but hears that it’s important. He asks around, and is told to launch 50 million pages. The startup starts with 100,000 and sees traffic go up. Then it launches 1 million more pages. More traffic comes. Then it launches 50 million pages. Traffic soars. Then a few months later, traffic tanks. The founder frantically tries to get the traffic back. Months go by, but nothing works.”
It’s not earth-shattering to learn that Google tweaks its algorithm to reward good content and user experience, but the hard part is figuring out how to act to actually reap those benefits. “What’s required is not only a change in behavior, but a switch of mentality. Nearly every get-traffic-quick scheme pulls the carpet out from under you eventually,” says Smith. “You have to think for the long haul; the growth curve you want is more gradual and steady. It won’t look good in the early days, but it will lead to significantly more long term traffic if you’re consistent.”
Here’s where to start to embark on the steady curve:
If improving your content quality and traffic is your goal, there are levers to pull, but don’t assume you can choose which ones. “Most people instinctively nominate themselves as authorities of what’s good when it comes to content and user experience. With SEO, that’s not what actually matters,” says Smith. “What’s important is what Google has defined as good content and good engagement. Note that I didn’t say what is objectively good, but what is subjectively defined as good by Google. Google has a set of guidelines — about 150 pages of them — and has hired thousands of quality raters to rate pages and sites based on their guidelines.”
The closer your page resembles those labeled as good content by Google’s rubric, the better your “content quality” is and the more your page will rise in the rankings. “It’s like any sport or board game. Some rules may not make sense, but you have to abide by them to win over time,” says Smith. “For example, scraping and summarizing content can be arguably a very useful thing for users, but Google has decided it’s not. For Yummly, recipes that have a lot of reviews or that have a ‘cook mode’ are ranked highly because Google decided that these make a recipe high quality.”
When it comes to Google’s guide to content quality, it’s big theme is EAT, which stands for expertise, authority and trustworthiness. “The guide goes into detail about what standard of expertise is expected for various topics. For example, medical advice should come from people or organizations with the proper accreditation, but forums for spouses of those who died from an ailment are also considered experts. The differentiating factor is what life experience is necessary to credibly give others value. The guide offers a good guiding question: ‘What kind of expertise is required for the page to achieve its purpose well?’”
Smith works with companies to help establish high EAT scores for their particular category. “In order to get a high EAT score in shopping, Wanelo shows reviews and photos for each product as well as return policy and shipping information. To show that Wanelo is a highly reputable company, it highlights press coverage in the New York Times, Fortune and other credible media publications on its press page,” he says. “For local services, Thumbtack highlights how many years a practitioner has been in business, whether they are licensed, how many jobs they have completed and, most importantly, lots of 5 star reviews.”
With SEO, you can be righteous or right. To be the latter, follow Google’s guidelines, not your gut.
Beyond the few hundred pages of SEO guidelines from Google, Smith highlights a few underappreciated tips to consider as you tweak your search engine optimization strategy.
One of the underappreciated aspects of SEO has less to do with what you do and more with what you don’t. It’s not the type of sexy growth hacking that gets written about, but it’s as critical. “Too many people continue with growth strategies that worked before Panda, but which now penalize them. Don’t spend energy trying to get out of Panda’s penalties; direct your energy toward good hygiene and maintenance. This strategy involves striking a balance between testing new SEO strategies and tracking your pages to detect issues early to prevent future penalties.”
Here are a few tips from Smith on keeping good SEO hygiene:
Schedule a weekly open-ended crawl of your site. A lot of times, companies’ sites will have pages cluttering Google’s index that they don’t want indexed and don’t know about. “What most people don’t understand is that Google will index whichever page they want. It's not about what you put in your sitemap. It's about any page Google can find it will index unless you tell them not to,” says Smith. “Establish an open-ended crawl of your site weekly to look for pages that shouldn't be there. Detecting those is supremely important to improving SEO — and part of good hygiene. Use tools such as Screaming Frog to crawl your site and QA."
Track your number of pages, how many are getting crawled, how many are indexed and how many are getting traffic. "Take note of any big differences in those figures. So, the alarm should sound if you have 1,000 pages, but only 200 are getting crawled. Or 1,000 pages crawled, but only 200 are getting indexed. Or 1,000 are indexed, but only 10 are getting traffic," says Smith. "You want all your pages to be crawled on a weekly basis and 95% indexation rate with the majority getting traffic.”
Eliminate incidental indexed pages. It bears repeating: it doesn’t matter if you think your pages are important; what’s key is what Google deems important. Google will tell you whether your page is important by crawling it often, indexing it and sending it traffic. Google doesn’t like sites with lots of thin pages. Once a crawl catches these “thin” pages, get rid of them. Don’t be surprised if you find "thin" pages indexed unintentionally or unknowingly.
“Depending on their site architecture, startups can accidentally create and index tons of pages without knowing its impact on SEO,” says Smith. “For example, Yummly users have profile pages for each of its 18 million users. These millions of pages are core to the Yummly feature set and thus are very useful. However, very few people search Google looking for profile pages — they search for a chicken casserole recipe. The issue is that Google will classify these profile pages as too thin and penalize us. We avoid this by being careful about which pages we allow to be indexed.”
Once you’ve found “thin" pages, there are a few ways forward. “The best and fastest way to eliminate them is to remove an entire directory by disallowing it in the robots.txt file or using the directory removal tool in Google Webmaster Tools. You can also remove individual pages either with a noindex tag or by returning a 404/410 error code,” says Smith. “But stay vigilant, as there are other ways 'thin' pages crop up. Many sites have HTTPS and HTTP versions of every page; that means that every single page gets indexed twice. Pages can be indexed with and without a “www.” At Yummly, we noticed our DEV servers — like DEV1.yummly.com, DEV2.yummly.com and so on — created duplicate versions of our site. If we didn’t catch it, we’d accidentally have 10 versions of our site getting indexed and causing duplicate content issues.”
Flatten your internal link architecture. Pagerank sculpting is dead. What matters now is a sufficiency of links and pagerank. The more you can flatten your internal architecture, the more each of your pages attain sufficient pagerank, and the more your traffic will increase. “Most internal links are skewed toward just a few pages. You might have internal links that are the most recent posts. If you have a thousand pages, the most recent posts are likely ten pages. Then you're not linking to the other 990, which signals to Google that these pages are not important and so they don't get traffic,” says Smith. “Instead, link across all your pages so that Google has many paths to find all of the pages on your site, not just a small percentage of them. So, if you’re showcasing an easy-to-use digital camera, don’t just link to other cameras. Link to other easy-to-use categories, like headphones or TVs. Cross-linking vertically and horizontally creates a more tightly-connected, flat link architecture.”
The growth strategy that took Smith three years to find and fine-tune at Wize started to work at Yummly after only three months. Given that long term growth in SEO takes time, startups can’t afford a steep learning curve when just getting started. “Worry about innovation when it comes to your core product, not your SEO. It’s not well publicized but most successful growth teams and companies take the most effective strategies from competitors, apply them in-house, then improve upon them,” says Smith. “SEO is no exception. Most of what works is not obvious at the onset, so startups spend an enormous amount of time reinventing the wheel.”
Smith has found success with research — not just secondary, but primary. “We look for the most successful SEO companies. We deeply analyze their site and strategy to understand what’s working. Then we try to connect with the person in charge of their SEO to ask them directly what’s working and what’s not,” he says. “Don't just research and reach out to who you subjectively think might be interesting. We use Sistrix and SimilarWeb to evaluate competitor traffic and engagement. For Yummly, we care about certain growth indicators — such as overall traffic, high SERP rank, and a steady growth trajectory — to identify which sites have the most effective SEO strategy. Then we find people who work at those companies, reach out and ask if they’d be open to a call. I never have open-ended chats over coffee. I do my homework and bring a list of pointed questions about specific parts of their site.”
Ninety percent of the people that Smith reached out to share techniques with him. “They might not say everything but they’ll answer if I have specific questions. Taken together with other conversations, this helps keep our strategy current. Of course, I offer them ideas on what has worked for us, too, so that the exchange is mutually beneficial,” he says. “It sounds simple, but many just research from afar. We like to both reach out to other growth experts as well as test common attributes across sites — like Yelp, Zillow and Houzz — that correlate with traffic. In fact, don’t just look at competitors, but those outside your category. We’ve actually learned a lot from TripAdvisor’s page structure.”
At Yummly, Smith used both types of research — along with his personal experience — to successfully revamp its category pages. “We launched category pages at my last company where we would find keywords and create pages that target those keywords. So we did the same at Yummly, basing them on structured taxonomies like shopping categories and filters. So there’d be a page for digital cameras and a page for every filter, such as digital camera between $50 and $100,” he says. “It worked, but not well. By talking with peers and pattern-matching with other sites, we realized that our structured taxonomies were not using the words that people use to search in Google form. We then found the phrases that people are actually searching for on Google, such as ‘digital cameras for wildlife photography,’ and curated category pages for those terms. That worked much better because the page exactly matched an actual query that people type into a search box.”
According to Smith, most big companies became very successful by being extremely aggressive with their growth. “Airbnb used Craigslist early on to tap into a broader market. Pinterests emails all your Facebook friends telling them you followed them when you didn’t. Linkedin makes it easy to accidentally send invitations to your entire contact list. The most successful companies are also extremely aggressive — with email marketing, paid advertising and referrals — but not SEO,” says Smith. “In the past, SEO success favored the most aggressive companies. After Google Panda or Penguin, success favors companies those that take a steady long term approach to growth. Be aggressive with SEO in the sense that you want to deploy resources to work on it in the first place, but don’t do too much too fast that’s not in your actual strategy.”
In the age of mobile, most will claim that websites aren’t important anymore or that SEO is dead. “The truth is that, when used as an acquisition vehicle, SEO is more effective now than it ever has been. In fact, other than gaming apps, the most successful apps, like Pinterest, Yelp, and Houzz, use web SEO as their main app acquisition channel,” says Smith. “Most SEO visitors come to a site, look at a couple pages, then leave. It’s easy to dismiss these passerby users as not valuable. The goal is to get SEO visitors to come back. Whether it’s by getting someone to register, use an engaging feature or download the app. When viewed as an acquisition channel, SEO can be one of the most impactful channels to drive long term lifetime value.”
With SEO, you can never be 100% confident. But once you know that, you’re more likely to be good at it.
The first mistake you might make — especially as a startup — is thinking that you need a team or person dedicated to SEO from the get-go. “Your SEO strategy and roadmap might come from someone you have in-house, but it’s as likely that it will — and probably should — come from a consultant or advisor,” says Smith. “At Wize, we turned the corner when our consultant, Leo Haryono, gave me his initial recipe, which I took and built from. I know I could have eventually come to the same result, but calling him in helped us shortcut that learning process.”
An SEO hire or consultant should join in once you’ve validated some traction and traffic. “For most consumer companies, that'll be around 100,000 visits per month. The role of the SEO manager will not only be to double-down on what’s started to work, but to build really good content, make engagement better and build PR/marketing backlinks through the methods I’ve mentioned,” says Smith. “At this point, the SEO person needs to become a stakeholder in product decisions and should advise on the parameters that the content creators or Product should consider in their work. Make sure this person can point to an SEO engine they’ve built at another company and explain how exactly she grew traffic with tools that don’t cost a fortune. A lot can be done with Google or Bing Webmaster tools as well as free or low-tier versions of Majestic SEO or Sistrix. If she can’t clearly communicate those two elements, take pause and move onto other referrals.”
It’s at this point that a startup should double-down with SEO, but through smartly building out its existing teams and product. “You’ve got a team and a dedicated SEO person or consultant. Now it’s about syncing and scaling the teams in an intentional way. If reviews are a lever for growth for your business, build in hooks into your core product to encourage users to write them. Airbnb has done a great job here in creating hooks that feed SEO into the core product,” says Smith. “Nerdwallet is another great example. By scaling into an army of really great finance writers, they have been able to have great SEO. Its focus was not around weird tricks and hidden text, but about writing high-quality articles on topics that people sought out.”
There has been a sea change with SEO. With the introduction of Google Panda and Penguin, the game has changed over the last five years. The problem is that the methods of many people tackling SEO have not changed along with it. Through good hygiene and understanding of Google’s guidelines, startups can avoid falling into traps that could penalize their traffic and pagerank for years to come. Schedule a weekly crawl of your site to identify and eliminate “thin” pages on your site. Take proactive measures to flatten your link architecture. Start SEO from Day 1, but don’t be too aggressive too fast — either with your strategy or your investment in an SEO hire. The name of the game is slow and steady wins the race. Instead invest in good content and frame SEO as an acquisition tool. Have faith in in in the long term; quick fixes lead to quick declines.
“With SEO, even if you find a way to identify every variable, you won’t be able to control each one. You might change a title tag but there are a thousand other factors that you're not controlling for that influence results. The challenges with causation and correlation abound,” says Smith. “If there’s one thing I hope startups take away about SEO is that it’s a marathon, not a sprint. Quick-twitch muscles won’t do any good here — in fact, they’ll likely get you into trouble. Endurance and awareness are the attributes that get rewarded with SEO. I wish people would swap growth hacking for growth harvesting. That’s the type of approach that truly works.”
Photography by Bonnie Rae Mills.