How to pick a name for your startup: a step-by-step guide

Naming a startup feels like it should be quick. It rarely is. Here's a framework that actually works, based on how most founders get unstuck.

Step 1: Decide what the name needs to do

Before brainstorming, get clear on constraints:

• Does it need to describe what you do, or can it be abstract?
• Does the .com need to be available, or are you fine with a modern extension (.io, .co, .ai)?
• Is this a name your grandmother needs to spell correctly over the phone, or is a slightly unusual spelling fine?

Skipping this step is the #1 reason naming sessions go in circles — people are silently optimizing for different things.

Step 2: Brainstorm wide, then wide again

Good names rarely show up in the first ten ideas. Push for volume first, judgment later:

• Real words used in an unexpected way (Slack, Stripe, Notion)
• Invented words that sound like real ones (Spotify, Zapier)
• Combinations of two relevant words (Snapchat, Mailchimp)
• Names built from Latin/Greek roots (Google, from "googol")

Don't filter as you go. A name that sounds silly alone might sound perfect next to five others.

Step 3: Stress-test the finalists

For your top 5-10, check:

• Does a quick search show anyone already using it in your space?
• Say it out loud — does it sound natural, or does it need explaining every time?
• Does it work as a one-word verb or brand mention ("just Slack me")?
• Is the domain realistically gettable, even with a small tweak?

Step 4: Get real votes, not just opinions in a group chat

This is where most teams stumble. Someone shares a shortlist, three people reply with heart emojis, one person goes quiet, and the founder ends up picking based on vibes rather than an actual majority.

Structured voting fixes this: everyone votes on the same shortlist, ties get resolved with a clear re-vote round, and you get an actual answer instead of a slow fade into "let's just go with the first one."

If you're naming something with co-founders, investors, or an early team, it's worth running this as a real process rather than a Slack thread that dies after two days.