Joost de Valk on Building Yoast, SEO Fundamentals and the AI Content Shift

Joost de Valk began with a modest plan for a paid video SEO plugin. Its first month brought in more than the $100,000 he had hoped to make in a year. In this 2023 interview, he explains how that side project became Yoast and what founders should understand about search, open source and AI.

De Valk was an SEO consultant who built WordPress plugins to fix problems he encountered in his work. He brought separate tools together into Yoast SEO, aiming to help sites get their technical foundations right and give writers feedback on content.

A large user base created a support burden before there was much of a business. His wife Marieke urged him to find a way to earn money from the work. The paid video plugin changed his expectations, although his consulting income initially made the decision to focus on the company difficult.

De Valk says Yoast did not take outside capital because the business generated enough cash to finance the growth they wanted. That choice gave the founders freedom to pursue things they found interesting rather than maximize revenue at every decision.

At the time of the sale, de Valk says Yoast had about €15 million in annual revenue and was used on roughly 11 million sites. The episode description says Newfold acquired the business in 2021 and later puts the number of sites at 13 million; these figures refer to different points in time.

Freemium and open source require a difficult balance: the free product must attract users and be useful, while the paid product needs a reason to buy. A company also has to support a large community even though only a fraction pays.

Open source brings a special constraint. Once a feature has been released, other developers can fork it; removing it from the free product can alienate users and invite a competing version. De Valk cautions founders against putting too much into free too early, while stressing that no single split works for every product.

The code alone is not the entire business. De Valk points to community, reputation and brand as defensible assets: users may arrive for the software and stay for the people. A competitor can copy code more easily than it can copy the relationships around it.

Branding matters in developer-led businesses too. De Valk credits Marieke with bringing a perspective beyond his own: left to himself, he might have optimized for SEO experts and developers requesting highly specific features, rather than ordinary users.

Harry challenges that advice: power users love the product and recommend it to others. De Valk agrees there is a fine line. Yoast tried to satisfy experts by fixing problems automatically and cleanly, while keeping the interface simple enough for everyone else.

That simplicity is hard to maintain in an SEO plugin with many possible settings. Most people simply turn the plugin on and expect it to work, so defaults matter enormously. De Valk describes the WordPress principle as “decisions, not options”: choose a sensible default when it serves most users, while allowing developers to change behavior in code.

One product mistake was assuming people used features merely because Yoast had built them. Limited usage data in a distributed WordPress plugin made this harder to see. When the company obtained better feedback, it discovered that some features received much less attention than expected.

Harry asks whether tighter company budgets threaten freemium. De Valk is concerned about businesses with an inflated cost base and small contract values, particularly if they depend on distant enterprise deals. A low-cost product still needs economics that work when customers scrutinize spending.

Evaluating open-source investments takes more than counting GitHub activity. De Valk listens to the people building and using a project and looks for a community capable of making a difference. Code can be copied; sustained commitment and a clear purpose are harder to reproduce.

For a founder thinking about SEO, de Valk starts with the words potential customers actually use. Understand what people are searching for and answer those questions. He recommends choosing a technically search-friendly platform and acknowledges his own WordPress background and commercial ties.

Content need not mean a huge block of text on every page. Even a transcript can be presented in a way readers can navigate. Search engines still need text they can interpret, and the business needs information that genuinely answers the visitor’s question.

The right time to think about SEO is early, particularly when choosing the platform and creating a knowledge base. Founders need not immediately hire a large SEO team, but avoiding technical problems and publishing useful answers can make later work much easier.

A recurring mistake is writing in the company’s internal language instead of customers’ search terms. Another is treating SEO as a substitute for a memorable brand. Search captures demand that already exists; people also have to know why they should remember or seek out the business.

SEO is not a task for one isolated specialist. De Valk says the work touches product and engineering, which control the site, and communications, which shape what is published. The specialist must be able to work with the teams that can actually make changes.

Results depend on the starting point. A new company may need time to establish recognition and authority, while an established brand that has neglected its site may see improvement sooner after publishing useful, well-structured content. Neither outcome is a guarantee.

When Harry asks about budget, de Valk uses a share of existing cost-per-click spending as a possible starting point—around 10% in his example. He does not present it as a universal formula. SEO can influence other channels and be difficult to attribute cleanly, so a company must evaluate the investment in context.

The best SEO teams understand the product and the customer’s use cases, rather than mechanically inserting keywords. De Valk suggests developing stories around what a product lets people do, which can produce useful pages matching real searches.

AI can help with outlines, research, excerpts and making text easier to read. De Valk sees practical uses for writers, but cautions that generated material can contain invented facts. A person must verify factual claims and contribute original insight.

He expects AI to make large volumes of content cheaper to produce. That does not mean all of it will be useful or distinctive: creativity, expertise and a real point of view remain ways to make material worth reading and sharing.

De Valk recognizes that Yoast’s own traffic-light guidance is an imperfect proxy for quality. People can over-optimize to please a tool rather than write something useful. Good writing and judgment are harder than simply meeting a checklist.

He expects search engines to put more emphasis on signals of expertise, authority and trust as generated content proliferates. This is his prediction about how search may evolve, not a confirmed ranking rule or a promise of results.

Harry raises the question of whether AI companies should disclose the sources used in training. De Valk argues that creators’ work has value and discusses attribution and licensing as possible approaches. The interview does not establish a settled technical or legal solution.

De Valk worries that automation could enable firms to do the same work with fewer people and concentrate gains among those who own the technology. He raises the risk of greater inequality and questions whether regulators will keep pace; these are concerns about possible outcomes, not established forecasts.

On compensation for training data, he considers ways to recognize creators and the potential role of open models. The discussion is exploratory: practical attribution, consent and payment remain difficult questions.

De Valk expects AI to change software development as well. He speculates that in five years a large share—he suggests 60–80%—of code could be generated. In his view, understanding user experience and the underlying business problem will still require careful human thought.

After the sale, de Valk and Marieke invest through Emilia Capital. He describes typical angel checks of roughly $100,000–$200,000 and an ambition to make 20–40 investments over the next five years. Those are his plans and ticket sizes in the interview, not a guarantee of future activity.

De Valk values active relationships with founders and discusses helping with company building, SEO and product decisions. His advice also reflects his own position as a founder, WordPress contributor and investor; those ties give him experience as well as a stake in the ecosystem.

In the quick-fire round, he returns to the importance of clear documentation, strong brands and giving customers answers they can find. Growth is not merely a set of tricks: what the company builds and how it communicates have to work together.

Asked about platforms including Webflow, de Valk notes that his WordPress background colors his view. He closes by emphasizing the importance of accessible, search-friendly technical foundations and acknowledges that competing platforms have their own strengths.

Harry thanks de Valk for the conversation, and de Valk thanks him in return.