wikiHow

Type a question into the internet - "how do I unclog a drain," "how do I write a resignation letter," "how do I survive a bear encounter" - and somewhere in the results is a page with calm, numbered steps and a slightly surreal hand-drawn illustration. That page is almost certainly wikiHow. Twenty years on, it is still doing the unglamorous, oddly comforting work of answering the question everyone is too embarrassed to ask out loud.

wikiHow is a wiki, which means the instructions were written by strangers. It is also expert-reviewed, which means a strangers-on-the-internet operation somehow grew a quality-control department. It runs ads but lets logged-in readers skip them. It turns a profit but treats profit as the side dish. For a company that explains everything, wikiHow is itself a little hard to explain.

The Problem They Saw

The internet knew a lot. It just couldn't teach.

In the early 2000s the web was excellent at telling you what a thing was and terrible at showing you how to do it. How-to content existed, but it was scattered, thin, written to chase ad clicks, and rarely good enough to actually follow. Jack Herrick lived this problem from the inside: he and a partner had bought eHow, a how-to site, in 2004. After running it, he concluded the underlying business model would never let it become the thorough, high-quality guide he wanted. So he and his partner sold eHow to Demand Media and Herrick started over.

The thing he could not stop thinking about was Wikipedia. If a volunteer community could build the world's reference desk, surely the same model could build the world's instruction manual. The "what" had its wiki. The "how" did not.

The Founder's Bet

Launch day was a tribute, not a coincidence.

wikiHow went live on January 15, 2005 - the same calendar date Wikipedia had launched four years earlier. The date was chosen on purpose, a small bow to the project that inspired it. The bet underneath was bigger and stranger: that you could build a genuinely useful public resource and a sustainable business without choosing between them, and without anyone's venture money.

Herrick called it a "hybrid organization" - a for-profit company run for a social mission. In practice that meant a string of decisions most founders never make. wikiHow declined outside investment. It turned down acquisition offers from larger companies, repeatedly, over years. For long stretches the whole operation ran out of a house in downtown Palo Alto with a core team you could count on two hands - a team that, at its peak, was drawing more monthly traffic than the Los Angeles Times and the Washington Post combined.

The Product

Anyone can write it. Then everyone fixes it.

The mechanics are deceptively simple. Any visitor can start an article. Once it exists, others edit, correct, illustrate and improve it. Layered on top is a review process - subject-matter experts and trained community editors check articles for accuracy - plus those famously plain step illustrations that turned a utilitarian feature into a beloved internet aesthetic. The result is a library of more than 235,000 guides in 19 languages, maintained by 2.5 million-plus registered users and tens of thousands of active contributors.

The money side is just as deliberate. Most revenue comes from on-page advertising, but logged-in readers see far fewer ads, which inverts the usual incentive to bury readers in them. In 2020 wikiHow added wikiHow Pro, a subscription that strips ads and unlocks extras like custom PDF downloads and expert-built email courses. There are mobile apps for people who want the library in their pocket.

The Proof

The numbers behave like a tide.

wikiHow's reach grew for a decade, peaked, and then began absorbing the same shock hitting every how-to publisher: search engines that increasingly answer the question on the results page itself. Google's AI Overviews, launched in 2024, and the rise of AI chat assistants pulled clicks away from the open web. wikiHow's monthly visitors - over 100 million at the 2016 peak - sat around 35.5 million in early 2025. The chart below is less a victory lap than an honest weather report.

Strip away the traffic curve and the credentials still stack up: a Webby, a Co-Creation award from The Guardian and Nesta, a place on Gizmodo's list of sites that shaped the internet, a Forbes "Best Small Companies" nod, and content trusted enough that One Laptop per Child and the United Nations both came knocking. Not bad for a site whose house style is a stick figure calmly demonstrating the impossible.

The Mission

Teach everyone, everywhere, how to do anything.

That is the whole point, and it has not moved in 20 years. The 19 languages are not a growth-hack; they are the mission spelled out - because "anyone, anywhere" is meaningless if it only works in English. The refusal of donations, the refusal of buyouts, the choice to show logged-in readers fewer ads: each is a small act of protecting the thing from the incentives that usually eat things like it.

Why It Matters Tomorrow

In an age of instant answers, someone still has to be right.

AI assistants are very good at producing how-to answers. They are also, frequently, confidently wrong - and they learned much of what they know from places like wikiHow. That is the quiet tension of the next decade: the machines that summarize human instructions still depend on humans writing good ones. A library of expert-reviewed, community-corrected guides is not a relic in that world. It is the source material - and the fact-check.

So picture the original scene again. Someone, somewhere, is about to type "how do I…" into a glowing rectangle. Twenty years ago the answer would have been scattered, thin, and probably wrong. Today there is a decent chance it lands on a page with calm numbered steps, a quietly absurd illustration, and a small footnote saying an expert checked it. That page didn't have to exist. A dozen people in a house decided it should.

Watch & Explore

See it in motion