Complete guide: Voice over Casting

01 — Casting and Auditioning Are Not the Same Process

Most agencies treat VO casting and VO auditioning as interchangeable. Brief in, voices out, someone picks one. In practice, they are structurally different decisions with different time requirements, different risk profiles, and different quality gates — and conflating them is where the process breaks first.

Casting: You Already Know What You're Looking For

Casting is a matching problem. The brief specifies tone, dialect, regional register, brand voice, and project type. The question being answered is: does a voice already exist in a known roster that fits this brief precisely?

Done well, casting is fast. The talent database already contains the work history, regional credits, and audio profile you need. Match against the brief, present the shortlist, move to record. The sourcing work is done; the decision is about fit.

The risk in casting sits in the database itself. Most commercial talent rosters skew toward dominant broadcast variants — Castilian Spanish rather than Mexican Spanish, Modern Standard Arabic rather than Gulf Arabic, Mandarin as spoken in Beijing rather than Cantonese or regional variants. When the brief genuinely requires a voice of the market rather than for the market, a database that doesn't contain that voice cannot return it. Casting becomes a false positive: you get a shortlist, but it's the wrong shortlist.

Auditioning: You Need Proof Before You Commit

Auditioning is a quality verification process. The brief exists, but the right voice for this specific script, this specific tone, this specific brand read cannot be confirmed without hearing it against the actual copy. Custom auditions against real script lines are not a nice-to-have. They are a quality gate.

The risk in auditioning is two-fold. First, the quality of the brief that goes out — a vague brief returns demos, not auditions. Second, the review process itself. Sequential file review (one attachment, then another, then a third in a separate email) causes stakeholder anchoring on the most recently heard voice. The last audition listened to has an unfair advantage because it's the freshest. Side-by-side playback — all candidates in one place, clearly labelled, reviewable in sequence or at random — produces faster decisions and more confident ones. The standard email process makes better decisions structurally harder to reach.

These are different problems. Treating them as the same process means the casting workflow cannot adapt to which problem it's actually solving.

02 — How the Process Works

Casting and auditioning have different inputs, different steps, and different deliverables. Here's how both run end-to-end.

03 — Where the Standard Workflow Breaks

The standard agency VO casting workflow runs on email. Email is the wrong tool for a multi-stakeholder creative decision, and the failure modes are predictable enough that most PMs have already lived all of them.

The Single Source of Truth Doesn't Exist

Five stakeholders. Five inboxes. The brief went out on Monday. The creative director replied to the email instead of clicking the shared link. The account lead sent feedback as a voice note in Slack. The client-side contact hasn't responded to either. The PM is consolidating inputs across three channels, trying to determine whether the creative director and the client are aligned or whether they reviewed the same three files and reached different conclusions about which one.

They didn't. The creative director reviewed takes 2, 4, and 6. The account lead reviewed takes 1, 3, and 5. Nobody noticed until the debrief.

This is not an unusual failure. This is the default outcome when a creative decision with multiple reviewers runs on email with no version control and no centralised review environment. A casting decision that should resolve in 24 hours routinely takes three to five days once review cycles complete. The VO window on a post-picture-lock production may be five to seven days total. When casting alone consumes half of it, the schedule doesn't survive.

The WeTransfer Problem

Links expire. Files arrive inconsistently named. The third option that everyone agreed was strongest is now in a second WeTransfer that one reviewer received and three didn't. The PM re-sends. Waits. Re-sends again. Chases. Consolidates.

None of this requires strategic judgment. All of it requires PM time. PM coordination and communication tasks already consume the majority of a working week on complex creative projects. VO casting via unstructured email represents an outsize share of that — fragmented, multi-stakeholder, no centralised tool — and none of the overhead produces creative value. It is purely administrative work generated by the absence of process infrastructure.

The Dialect Gap You Don't Find Until Re-Record

Standard rosters skew to dominant broadcast variants. The brief says "Spanish — Latin America." The talent database returns options with professional credits in Spanish. The shortlist goes out, the client approves the second option, recording completes.

Three weeks later, the client-side brand team in Mexico City flags that the voice reads as distinctly Argentinian. The talent was not misrepresented. The brief was not precise enough, and the database was not deep enough to surface the distinction unprompted.

04 — The Re-Record Is Not a Line Item

When a miscasting surfaces after recording, the instinct is to treat it as a budget line: re-record cost, re-edit cost, re-master cost, re-delivery cost. That is the visible cost. The timeline cost is larger.

A VO re-record is a pipeline restart. Re-brief. Re-source. Re-record. Re-edit. Re-master. Re-deliver. On a single-market campaign, that sequence takes a minimum of several working days from discovery to broadcast-ready asset. On a 10-market campaign, a single miscasting at the casting stage can cascade into 40 or more re-deliveries — because the wrong voice went into production across every language version.

A re-record required 10 days from air date is not a budget conversation. It is a potential air date miss.

Most VO re-records are caused by casting mismatch, not performance failure. Wrong tone. Wrong dialect. Brand incongruence that was visible in the original brief but not caught at the audition stage because the audition was demo-based, not script-based. Script-based auditioning is a quality gate. It surfaces pronunciation flags, dialect issues, and brand register mismatches in the audition round, before the booking is made. The standard email-and-demo-reel process doesn't have this gate. It discovers those problems at re-record.

The financial cost of a re-record is real. The 2.1 hours per PM per week already lost to VO coordination overhead — chasing availability, re-sending expired links, consolidating conflicting reviews — does not include the time absorbed by a re-record cycle. That is additional, and it lands at the worst possible moment in the production schedule.

05 — Multi-Market Casting Multiplies Every Failure Point

A 10-market campaign via email is not a 10-market version of a single-market campaign. It is 10 separate casting processes running sequentially — because no parallel infrastructure exists in an email workflow.

The Timeline Multiplier

The timeline consequence is direct: 10 markets at the single-market cadence equals 10 times the single-market casting window. The coordination overhead scales with every market added — ten briefing threads, ten file sets, ten sets of stakeholder inputs from regional brand leads reviewing in different time zones, on different schedules, with different versions of the brief.

Failure Mode 1: The Dialect Gap

The talent who genuinely owns the dialect — professional recording credits, available within the campaign window, capable of a brand brief rather than a standard commercial read — is not always in a database. Mexican Spanish, Gulf Arabic, Cantonese, regional French, minority-market variants across Central and Eastern Europe: finding the right voice for these markets requires active sourcing.

Network contacts, local agency partners, social media scouting, regional talent communities. Most global agencies do not maintain this in-house. A marketplace returns what is in the database. A partner hunts when the database doesn't contain the answer.

Failure Mode 2: Stakeholder Alignment Breaks Down

Separate files, separate reviewers, separate feedback channels, conflicting conclusions that require another review round. This is not a cultural coordination problem. It is an infrastructure problem. When 10 markets are each running this process in parallel, the PM discovers at delivery that three markets approved take 2, four markets approved take 4, two markets haven't responded, and one market wants to start over.

VO Casting Should Return a Decision, Not a Coordination Problem

The PM's job in a VO casting process should be: review the shortlist, approve the voice, move to record. That is a 30-minute task. It is not a three-day coordination sprint across email threads, expired links, and misaligned stakeholder feedback.

The email workflow makes the PM the coordination layer. A structured casting process removes that role. The difference is not speed — although 20% of VoiceArchive's jobs deliver within 24 hours, and active sourcing begins immediately when the database doesn't contain the answer. The difference is where the PM's attention goes. On a campaign with 10 markets, eight other workstreams, and a fixed air date, VO casting should not be the thing consuming two hours of PM coordination every week.

It should be the easy part.