You can spot a bad voice over before you record it. The signs are there in the script, in the brief, in the casting process, in the translation. By the time you're in the session listening to takes that don't work, you've already lost. The damage was done upstream.
I've been doing this for over 20 years. Coca-Cola, Nike, Google, Ford, Netflix, Amazon β the list keeps growing. And the pattern I've seen across hundreds of projects is consistent: the voice overs that fail were predictable failures. The warning signs were visible to anyone paying attention. The problem is that most people don't know what to look for.
The script tells you everything
A bad voice over starts with a bad script. This is true in any language, but especially true in Spanish because of the 30% rule. Spanish runs approximately 30% longer than English. According to research from the localization industry, this expansion rate is one of the highest among major commercial languages. When someone hands you a Spanish script that was translated word-for-word from English with zero adjustments for timing, you're looking at a future problem.
The delivery will be rushed. The talent will be forced to speed through phrases that should breathe. The audience will feel the compression even if they can't name it.
Look at your script before you book the session. Read it out loud. Time it against your video. If you're already running long on paper, you're going to run long in the booth. And then everyone will blame the talent for sounding unnatural when the talent was set up to fail. Why Spanish scripts translated from English always need editing β this is a topic I've written about extensively because it keeps happening.
A brief with no reference audio
Here's a quality-control checkpoint that most people skip: does your brief include reference audio?
When a client sends me a brief that says "warm, conversational, professional" with nothing else, I know we're heading toward misalignment. Those words mean different things to different people. "Warm" to a pharmaceutical brand and "warm" to a tech startup are two entirely different sounds. Without a reference track, the talent is guessing. Sometimes they guess right. Often they don't.
The fix is simple. Find an existing voice over β any voice over, any language β that captures the tone you want. Send it. Now everyone has the same target. The session goes faster. The first take is more likely to be close. And you can spot potential problems before you record because you can compare your brief against your reference and ask: does this make sense?
When the casting pool is too big
Have you ever opened your inbox to find 300 auditions for a Spanish voice over casting? Congratulations. You now have a quality problem disguised as a quantity solution.
Platforms like Voices.com and Voice123 promise efficiency through volume. A Voices.com case study claims they deliver thousands of auditions within 24 hours for most projects. That sounds impressive until you realize you need one voice, and sifting through thousands of mediocre options doesn't get you closer to finding it. (I've seen clients spend more time listening to bad auditions than they would have spent working directly with a professional who could deliver three variants in an hour.)
The casting pool tells you something. If you're receiving auditions from people who clearly didn't read the brief, who submitted character reels for a corporate narration, who claim to do "neutral Spanish" but have thick regional accents β you're using a broken system. The algorithm doesn't filter for quality. It filters for activity. The people who game the algorithm rise to the top regardless of skill.
Mass casting makes it harder to spot bad voice over, not easier. You're drowning in options, most of which shouldn't exist.
The accent mismatch nobody caught
A Fortune 500 brand I worked with last year (I won't name them, but you've seen their ads) came to me after a disaster. They'd recorded a pan-Latino campaign with a talent whose accent was distinctly Caribbean. The creative director loved the energy. The problem? Their target audience was Mexican, Central American, and South American. According to Pew Research Center, Mexican-origin Hispanics make up about 60% of the US Latino population. The Caribbean accent didn't land. It created distance instead of connection.
This happens constantly. Someone in the room likes how a voice sounds without understanding that regional accents carry regional associations. Latin American rivalries are real. A Colombian audience might find an Argentine accent charming or grating depending on context. A Mexican audience might find a Puerto Rican accent hard to follow. And none of this is visible to a non-native speaker making the casting decision.
The quality-control checkpoint: before you finalize your casting, have a native Spanish speaker from your target demographic listen to the auditions. Not your bilingual marketing coordinator who learned Spanish in college. A native speaker who grew up in the region you're targeting. If that person isn't available, default to neutral Spanish. It works everywhere because it belongs to no single region.
The translation nobody reviewed
I had a session recently where the script included a phrase that was grammatically correct but culturally absurd. The translator β clearly not a native speaker β had chosen a word that no Latin American would ever use in that context. It was technically Spanish. It was functionally nonsense.
The talent caught it. I caught it. But the client had already approved the script, the project was on deadline, and we had to do takes with the original wording for legal compliance while flagging the issue. The client never would have known there was a problem if I hadn't said something.
Before you record, have your translation reviewed by a native speaker who isn't the original translator. Fresh eyes catch what the translator missed. This is especially true when the translator is working from English and trying to preserve phrasing that doesn't transfer. And it's even more true when the client doesn't speak Spanish and can't evaluate the translation themselves β you need a native speaker involved in the process at multiple checkpoints, not just one.
The "neutral" that isn't
Many Americans who learn Spanish believe they speak neutral Spanish because they're not from any Spanish-speaking country. The logic: "I have no regional accent because I'm from no region."
Completely false.
What they learned is a broken version of their teacher's accent, plus their own American foreign accent layered on top. Every native Spanish speaker can hear it immediately. It's not neutral. It's non-native. And the foreign accent is extremely recognizable β there's a Brazilian foreign accent, a German one, a French one, an American one. Each has specific phonetic characteristics.
But here's how this becomes a quality-control issue: sometimes clients hire heritage speakers or non-native bilinguals for Spanish voice over because they sound "neutral" to non-native ears. The casting decision is made by someone who can't detect the accent. The result goes to market. And the target audience hears something that sounds off, even if they can't articulate why. Viggo Mortensen, Anya Taylor-Joy, and Alexis Bledel speak better Spanish than Danny Trejo, Jennifer Lopez, and Selena Gomez β because the first group are Argentine natives who grew up speaking Spanish, while the second group have Latino names but barely speak a word.
The checkpoint: if you don't speak Spanish natively, don't make the final casting decision alone.
What the demo actually reveals
A voice over demo should sound like the talent on their worst day. That's the standard I use for my own demo, and it's the standard you should use to evaluate others.
If a demo is heavily produced with music beds, effects, and editing that makes every transition seamless, you're hearing a production β not a voice. When you hire that talent and they show up without the production polish, you get surprised. And not in a good way.
Look for demos with dry reads, solo voice, minimal production. Listen for consistency across multiple spots. Does the voice sound the same, or does each clip feel like a different person? If the demo is too perfect, too polished, too varied β that's a warning sign. The talent may have been over-directed, over-produced, or may simply not be able to replicate what you heard.
The direction that guarantees failure
When a brief says "I want a Colombian accent" with no strategic rationale, that's a red flag.
Usually it means one of two things: the client wants "not Mexican" and doesn't know what the alternatives are, or someone on the team has a Colombian friend whose voice they like. Neither is a strategy. Both lead to arbitrary accent choices that don't serve the actual audience.
The quality control question to ask: why this accent? If the answer is anything other than "our target audience is specifically Colombian" or "we're localizing for Colombia specifically," you're probably making a mistake. And the mistake will be invisible until the campaign launches and the audience doesn't connect.
Before you book the session
Run through this checklist. It takes ten minutes. It can save you thousands in re-records and missed opportunities.
Is your Spanish script timed against your video, or are you hoping it will fit? Has a native speaker reviewed the translation for naturalness, not just accuracy? Does your brief include reference audio so everyone shares the same target? Is your casting pool manageable, or are you drowning in unqualified auditions? Has a native speaker from your target demographic approved the accent choice? Does the talent's demo sound like raw voice, or like a produced reel that can't be replicated?
If you can't answer yes to all of these, you're not ready to record. And if you record anyway, you're hoping the session will fix problems that should have been fixed in pre-production. Sometimes you get lucky. Usually you don't.
Need a Spanish voice over for your next project? Get in touch and I'll get back to you within the hour.



