How to Build a Billion-Dollar Company with SEO

Hamburg, what's up! Great to be here. Hamburg is my adopted home from back when I lived in Germany, before I moved to the States, and I'm thrilled to be here and happy to finally be at live events again after two years.

Quick intro: I'm Director of SEO at Shopify, I advise a number of companies, and I'm an angel investor. I write a weekly newsletter about tech and SEO, and I have a podcast. But enough about me, what are we here for today?

What are SEO unicorns?

Billion-dollar companies, or "unique roles in SEO," take your pick. Quick show of hands, who here does SEO? Quite a few, good, otherwise I could have skipped this talk.

An SEO unicorn has to meet one of two criteria: either SEO is its main channel, its main customer-acquisition channel, or it has a valuation, a market cap, of over a billion dollars. I brought examples from four categories, SaaS companies and publishers on one side, marketplaces and ecommerce companies on the other.

If you plot them by valuation, Canva leads on the left at $40 billion, and on the other side you have companies like G2, where I worked for many years. If you instead plot organic traffic against valuation, you quickly see that the companies with the most traffic don't necessarily have the highest valuation, simply because different business models monetize traffic differently: traffic is worth something different for an ecommerce company than for a SaaS company. Looking at the traffic mix, some companies get more than half of their total traffic from organic search, while others, like a shoe retailer or Canva, get under 10%, because they also get a huge amount of direct traffic through the brand, though in absolute terms that is still a lot of SEO traffic.

Why is SEO interesting?

Why put yourself through this at all? Simple: eight billion people search Google every day for something, roughly one search for every person on earth, so there is enormous volume. Compared with other platforms, search engagement sits in the middle, Twitter sees 500 million tweets a day and YouTube gets 30 billion views a day, but the user intent behind a Google search is very different from scrolling Twitter or watching YouTube.

SEO is also one of the four channels that scale extremely well for customer acquisition, alongside things like sales and product-led growth. The magic is in the marginal cost: like product quality, SEO has low marginal cost, you set it up once and it runs, whereas Google Ads or a sales team only scale by putting in more money. Add scalability and SEO stands out as a high-yield, highly scalable channel.

SEO has two problems, though. Google uses more and more machine-learning algorithms, which makes reverse-engineering harder and more complex. And Google answers more and more questions itself in the results. For a Valentine's Day bouquet search in Chicago, I marked in blue how little real estate the organic results still get, and for some queries it keeps shrinking. Movie reviews are another example, Google shows so much directly that there is not much organic search left.

Integrators vs aggregators

These unicorn SEO companies fall into two groups, integrators and aggregators, a concept coined by Ben Thompson, who spoke here at OMR a few years ago. SaaS companies and publishers tend to be integrators; marketplaces and ecommerce companies tend to be aggregators. They scale through very different growth levers, and both rest on the pillars of product marketing, you do not need a content team if you do not have a product that actually pulls the market.

The main difference: integrators have to generate their own content, blog articles, landing pages, and generation tools, while aggregators can use user-generated content to scale, or hold inventory in the ecommerce sense. Their challenges differ too: integrators must differentiate with a strong brand and high recognition, while aggregators mostly need more inventory or UGC. In SEO terms, integrators drive results more through content and links, while aggregators live on technical SEO and the number and quality of the subpages they provide. And the ultimate driver differs: for integrators it is a strong brand, for aggregators it is the product.

A few examples. Tripadvisor has 80,000 pages in its tourism subdirectory generating over 6 million users, showing restaurant and hotel results for specific cities and ranking for the city names themselves, for example a search for Cancun. On the other side, an integrator like Bankrate has 800 pages in its loan directory, about 1% of Tripadvisor's 80,000, yet still generates a third of Tripadvisor's traffic, over 2 million monthly visitors, but they have to create the content themselves.

Look at the structure and you see integrators get most of their traffic through the blog and the tools they build, all self-created content, while an aggregator like Zillow has a giant subdirectory for all its listings plus its own page types for each state, like California, and metro, like Phoenix, Arizona, scaling with page types integrators simply do not have. Traffic per subpage tells the same story: integrators generate fewer subpages but more traffic each, aggregators generate more subpages but less traffic each.

What can we learn: the top five

Now the core question, what can we take home? Here are the top five things these companies do extremely well.

1. User intent. What does a user actually want to achieve with a search? Take "near me" queries. Search volume for real estate near me has gone through the roof over the last six years, so it makes sense to ride the trend: Zillow ranks for over 17,000 of these keywords, generates over 2 million monthly users, and delivers great UX, detecting where I am by IP, showing results on a map, sorting by the newest listings, and doing it fast. Tripadvisor does something similar for restaurants and pulls over 13 million monthly users from a single restaurant page. Ask yourself how users expect to reach you, what experience they expect, and what the result should be, a list, a single answer, or a blog article.

2. Expert content. Who writes your content matters more and more in SEO, especially in finance and health, where expert opinions carry weight. Bankrate cleanly credits authors beneath the article headline and works with experts regularly, and some companies collaborate with doctors and specialists to create content. When I worked at G2, we did the same, over 30 writers who each dug deep into a topic and established themselves as a kind of expert in the industry. Ask which experts could create or contribute to your content, whether you are an expert in your area, and how important expertise is for your topic.

3. Internal linking. Used more by aggregators, and it scales well. On Zillow's homepage, with modules collapsed on the left and expanded on the right, you can see how aggressively they scale internal linking, to each state, city, and beyond, and they do it on product pages too. On a property page in Chicago, I get internal links to 13 apartments in Chicago, two-bedroom apartments, one-bedroom apartments, very thoughtfully done. Tripadvisor links to similar page types, restaurants near me broken out by cuisine, Chinese, Italian, and so on. An integrator can do it well too, linking at the end of its articles to the apps it mentions, giving those pages more weight in the site structure. As an aggregator, ask which link modules you can build that link deeper into the site; as an integrator, ask which of your pages need internal links, often landing pages for competitive or comparative keywords, and which links actually get clicked, because too few people look at the data on which nav or footer links get used.

4. Lead-generation tools. Especially important for integrators, tools they build themselves. Bankrate has a built-in loan calculator, and that single page generates a million visits a month while solving a real problem for its audience, finding the right loan. Credit Karma does something similar with a general credit calculator, ranking for over 3,000 keywords and pulling over 100,000 monthly visitors. Ask whether you can solve a core problem for your audience with a tool, whether tools fit the user journey, and what information users are searching for.

5. Optimize the result itself. As Google gives more direct answers, it matters more how you win attention in the results. Tripadvisor runs a highly optimized snippet, with expandable questions and current signals like "made in Italy" at the end of the title and "ten best restaurants in Chicago" implying curation. Zillow does it well too, showing a regularly updated listing count to signal they have the freshest inventory, delivering value before the click. Featured snippets, the direct answers at the very top, get a lot of clicks, and to win them you shape your content accordingly: at G2 we rebuilt our "best ERP systems" category pages to put exactly that answer in a paragraph at the top. Another company used an unordered list as the top answer and wrote its blog article that way, which works very well. The top-tier example is Shein, with expandable FAQs, an image carousel that grabs attention, and three value props in one title, "top brands, low prices, free shipping." You can use structured data to earn these enhanced results. Ask how you can win attention in search before users even click, whether you can use schema, and which direct answers Google shows for your area.

The winning formula

Put it together and the winning formula for unicorn SEO in 2022 is additive content that matches user intent, click-through optimized, powered by technical SEO, and behind a strong brand. It is a complex topic, and there is plenty I could not cover, backlink generation, feature testing, index management, and more, but I will come back next year for those. Thank you.

Q&A: the one tool

Asked which tools he uses, Kevin points to Shopify (as a joke), the Google suite (Docs, Gmail) for daily work and internal communication, and a lot of tools built in-house, like a custom wiki. The one tool he could not do without, the obvious answer for an SEO: Ahrefs, for the daily data and content his work depends on.