Most naming processes fail for a boring reason: the name gets evaluated in the wrong format. A word on a whiteboard, in a spreadsheet, or in a Slack thread is not the same object as a name on a business card, a storefront sign, or an app icon. Agencies that skip straight from brainstorm to client presentation are effectively asking clients to judge a logo they haven't seen yet.
Below is a working checklist for evaluating brand names the way they'll actually be used — not just how they read. Visit the live site
## 1. Separate generation from evaluation
Generating candidates and judging them are different cognitive tasks, and doing both at once produces weaker shortlists. A generation pass should aim for volume and range: descriptive names, invented words, compound names, foreign-language borrowings. Judgment comes after, against fixed criteria — not gut feel mid-brainstorm.
## 2. Test pronounceability out loud, not on the page
A name that reads fine can still fail in a phone call, a radio spot, or a founder introducing themselves at a conference. Say every finalist out loud to someone who hasn't seen it written down. If they mishear it or ask for the spelling, that's a data point, not a nitpick.
## 3. Check the domain before you get attached
This sounds obvious, but it's the single most common reason a client falls in love with a name and then can't use it. The fix isn't checking domains *after* shortlisting — it's checking them *during* generation, so the candidate pool that reaches the client is already viable. Tools that surface domain availability alongside the name itself (rather than as a separate step) save a full review cycle here.
## 4. Evaluate the name as a visual object, not just a string
This is the step most naming processes skip entirely, and it's the one that matters most once a client has to actually use the name. A name is going to appear as a wordmark, an app icon, a favicon, a sign above a door. Two names that look identical in a spreadsheet can behave completely differently in a logotype — letterforms, word length, and even the shapes of individual letters affect how a name reads at a glance versus up close.
In practice, this means shortlisting names alongside a rough visual treatment — even a placeholder typeface and color — rather than as plain text. Some naming tools now generate this automatically: [OpusName](https://opusname.com), for instance, renders every suggestion as a brand card with its own palette and typography rather than a bare word list, which makes this step usable at generation time instead of requiring a separate design pass later.
## 5. Run a trademark screen before the shortlist goes to the client
A full clearance search comes later and usually involves counsel, but a basic knockout search (USPTO TESS in the US, equivalent registries elsewhere, plus a general web search for existing use in your category) will eliminate names that would otherwise waste everyone's time in round two.
## 6. Pressure-test the shortlist against three unrelated people
Not focus-group scale — three people outside the project who have zero context. Ask what they think the company does based on the name alone. If the guesses are wildly off from the actual business, that's worth knowing before the client sees the list, not after launch.
---
None of this replaces strategic judgment about brand positioning — it's the operational layer underneath it. But skipping any one of these steps tends to surface later as a client complaint, a blocked domain, or a rebrand eighteen months in. Building them into the process up front is cheaper than fixing them after.