TTS for radio ads and marketing

Updated July 2026 · 6 min read

By Zohaib Akeel · Cosette Team · July 5, 2026

Radio and Spotify ads compress storytelling into fifteen to thirty seconds — every word fights noise, commute distraction, and the skip button. TTS lets agencies iterate copy Friday afternoon and ship Monday flight without voice talent availability drama, when license terms permit broadcast use.

This guide covers ad script structure, loudness standards, call-to-action clarity, and Hindi/Urdu market variants for South Asian campaigns. Draft scripts in docs, generate finals in Cosette, and LUFS-normalize before handoff.

Ad lengths and word budgets

• 15 s ≈ 35–40 words
• 30 s ≈ 75–85 words
• 60 s ≈ 150–160 words

Read with a stopwatch — padding kills recall.

Hook and CTA placement

Brand name in first five seconds. Phone number or URL twice if mandated — once early, once at tail. Use spoken URLs ("cosette AI dot com") for clarity.

Voice casting for ads

Upbeat for retail; trustworthy for finance; urgent sparingly for limited offers. Female and male tests across demographics — no universal winner.

Loudness and delivery specs

Radio often targets −16 to −12 LUFS depending on station; streaming ads near −14. Leave headroom — stations apply processing.

Export WAV masters; MP3 only for previews.

Hindi and Urdu ad copy

Shorter words beat formal prose on air. Test Roman brand names in Urdu sentences before studio deadline.

Hindi IVR shares clarity rules.

Music and SFX beds

License commercial beds; duck under VO. Avoid lyric beds that compete with words.

Compliance and claims

Legal review superscripts — TTS does not remove advertising law. Archive approved script PDF with audio ISRC if applicable.

See commercial license guide.

Revision cycles with clients

Version scripts v1–v5 with date stamps; regenerate only changed lines in TTS to save retake fees versus human talent.

Business and client delivery

Client work needs signed-off scripts before generate — charge revision rounds in SOW. Export WAV or high-bitrate MP3 masters; keep 48 kHz archives even if delivery is video.

For product marketing, sync feature names on screen with spoken words within two seconds — desync feels amateur on demo videos. Version voiceover filenames with semver matching product release tags.

Verify commercial redistribution in your TTS license before paid campaigns; internal drafts may be allowed where public ads are not.

Key takeaways for ad voiceover

Brand name in first five seconds; CTA twice if required. Match loudness to station spec sheet. Legal review for regulated categories — TTS does not remove compliance obligation.

Sample 30-second ad structure

Brand (5 s) → problem (8 s) → offer (10 s) → CTA twice (7 s). Seventy-five to eighty-five words. Read with stopwatch before generate.

Broadcast delivery specs

Confirm station LUFS and file format — WAV 48 kHz common. Leave headroom; stations apply limiting.

Writing ad copy that survives TTS

Radio sentences need breath every eight to twelve words. Put the brand name in the first five seconds and repeat the phone number or URL twice if the station spec requires it. Avoid tongue-twisters — alliteration that looks cute on paper fails on air.

For Hindi and Urdu spots, test on a small Bluetooth speaker at arm’s length. Sibilance and clipped endings disappear on studio headphones but fail on car radios. Leave 0.3 dB headroom; stations apply their own limiters.

Client approval without endless retakes

Lock script v1 with word count and estimated duration before generate. TTS makes line edits cheap — charge scope for copy changes after approval, not for voice talent rebooking. Deliver WAV 48 kHz plus MP3 320 kbps; include loudness report if the client is a regional FM network with strict QC.

Regulated categories and disclaimer timing

Finance, healthcare, and alcohol ads need legal-approved disclaimer text with exact wording and duration. TTS reads disclaimers quickly unless you insert commas and slow speed to 0.9× for that block alone. Never paraphrase compliance copy to sound “more natural.”

Archive approved script PDF with station contract ID — regulators and clients trace ads by audio fingerprint and script hash months later.

Local FM versus digital streaming specs

FM stations may require mono 44.1 kHz while digital pods accept stereo 48 kHz — confirm before batch export. Hindi and Urdu spots for regional markets sometimes need formal register; write separate scripts rather than speed-tweaking one casual read.

Seasonal campaign versioning

Holiday ads need new scripts, not only date swaps in old TTS files — listeners notice recycled audio. Version filenames with campaign code and language. Batch Hindi and Urdu variants in one session so loudness matches across rotations.

Spot rotation fatigue

Stations rotate the same thirty-second spot dozens of times daily — scripts must stay clear on repetition, not only first listen. Avoid clever wordplay that confuses on the twentieth play. Refresh creative quarterly even when offer unchanged; TTS makes swaps cheap.

Traffic sponsorship reads

Traffic and weather sponsorships often need ten-second TTS tags — write word counts tight and test on car Bluetooth. Repeat station call letters at spec-required frequency without rushing.

Closing production checklist

Before station delivery, timer-read the script twice, confirm brand and CTA in the first five seconds, match LUFS to station spec, and archive legal-approved PDF. Test on car Bluetooth at conversational volume. Hindi and Urdu spots need consonant clarity on cheap speakers — studio headphones lie. Filename with client, campaign, language, and version prevents wrong spot airing. TTS makes revision cheap; stations still reject wrong duration — count words against slot length every time.

One habit to keep

Document voice ID, script version, and export date in every project folder before upload. Future you — and any freelancer — ship faster when settings are not guesswork. That habit prevents most inconsistent TTS output across a series.

Frequently asked questions