How I built Genius.com with Tom Lehman

We learned the hard way what power looks like when Google put its own lyrics box above us and even lifted our watermarked text, we called it out, they would not stop, and that fight taught me how the world really works.

Welcome to Before It Clicked, I’m Sunny Becky, and today we’ve got Tom Lehman, co‑founder of Genius, the go‑to place for lyrics and their meaning, from artist notes to fan annotations, a site that scaled to hundreds of millions of visits and was acquired in 2021.

Set the scene at acquisition so founders can picture the endpoint before we rewind.

Genius is a lyrics and annotation platform, and by 2021 we were around one hundred people with a big ads business across programmatic, sponsorships, artist content, and some great integrations like Apple; we sold to MediaLab, and it’s still a strong site today.

Let’s go back to 2009 and how you got here.

After graduating, I worked at D. E. Shaw while hacking nights on ideas in the Web 2 era, first a Facebook‑based debt tracker called Flif, then a DTC sheets play, and Genius showed up as the anti‑idea we could not ignore because it was just more compelling to build.

Were those first products real launches while you were still employed?

We shipped both and even sold sheets to strangers via search ads, then I cut down to part‑time so we could pour more hours into what felt most alive.

How did you decide when to quit or keep going on each idea?

It is brutally hard because a new, better idea tends to displace the old one slowly, not through clean disproof, so my rule of thumb became stick with things longer than your mood suggests, even as you fight the urge to endlessly build features in the dark.

Why spend so much energy on a passion project when the drop‑shipping business could make money?

The sheets business barely paid and our hearts were pulled toward building a consumer product that touched daily life, and timing mattered because lyrics later consolidated under giants so you could not start Genius today the way we did then.

What sparked the idea and how did v1 come together?

I loved having friends decode rap references for me, saw how understanding changed the way a line hits, and hacked a prototype in a day with clickable lines and pop‑up explanations, launched it, heard crickets, and kept refueling off tiny wins like a New York Magazine nod while pushing through a first YC rejection before getting in.

Were you thinking business model early or just following curiosity?

It began as an artful experiment, then we realized owning lyrics at scale plus better ads and partnerships could be a big business, which became the YC pitch.

How did you fund the grind, and did you timebox the journey?

We lived cheap, I had a part‑time paycheck and some savings, and I did not set a clock on it because I was committed to the non‑traditional path with no real plan B.

Walk us through day two after v1.

We hand‑built a small set of deeply annotated songs, blasted a big email to everyone we knew, pitched bloggers, and a few months later flipped on UGC so people could sign up, contribute, and give us feedback that pulled us out of our echo chamber.

Did you worry about music rights and the infamous do not build in music warning?

We knew publishers control lyric rights and that investors were wary, so we stayed quiet until we had traction and could sort licensing, but the nastier surprise was Google becoming the de facto competitor by taking our text and owning the search surface.

On speed versus analysis, where did you land?

We overthought too much, from naming to legal, and I now preach worse is better because shipping the rough thing to learn from real users beats polishing in your head.

What did 2010 look like as you tried to make it work?

We tuned roles, permissions, and points to motivate contributions and realized the spear tip was fast, accurate lyrics at release so we could win each song’s search, then bring people back with explanations.

What finally unlocked growth?

Some mix of SEO improvements and credibility led us to win more searches, and traction arrived like a mystery because you are riding the opaque Google algorithm.

Did copycats chase the annotation experience?

Not really on annotations, though we had tons of lyrics competitors including Google, which made it easy at YC to show how outdated the old sites were.

Once in YC with a million users, was it standard scale‑up blocking and tackling?

Mostly yes, we talked to users, shipped community features like chat and forums, did the unscalable work to get every new song up instantly, and dabbled beyond music while keeping the main graph up and to the right.

If you could advise your 2009 self, what would you change?

Guard your focus, avoid side quests that do not move the core metric, and be far more structured in hiring and firing because messy people decisions ate a ton of time.

You are building again; give listeners the quick plug.

I co‑founded Fastit, an Ethereum layer 2 focused on being unstoppable with no admin off switch and Ethereum ordering, born from that Google fight and a desire to build infrastructure outside the control of giant platforms.

Founders will hear resilience in this, from tinkering to traction to institutional scale.

One last thing, watch out for AI‑induced analysis paralysis because unlike humans it will debate with you forever, which can keep you from shipping and learning.