Do Ads Need a Person in Them? What Actually Matters
A straight answer on whether ads need a person in them, plus the real factors that decide it.
No, ads don't strictly need a person in them — plenty of ads perform well on product shots, text-on-screen, screen recordings, or motion graphics alone. But a person (real or AI-generated) tends to help because a face grabs attention faster and a voice delivering a script usually reads as more trustworthy than a slide of bullet points. Whether you need one depends on your product, your platform, and what you've actually tested, not a fixed rule.
When does a person help most?
A person on screen tends to help when the product benefits from a demonstration — skincare, apparel, gadgets, anything people want to see used or worn. It also helps in UGC-style formats, where the entire appeal is that it looks like a real customer talking through their experience. A presenter can front-load the hook in the first second or two, which matters more than almost any other single factor in the creative, since most viewers decide whether to keep watching almost immediately.
When can you skip the person entirely?
Text-driven or product-only ads can work well for lower-consideration purchases, B2B offers, or anywhere the value proposition is simple enough to state in on-screen text plus a voiceover. Some of the strongest ads on any platform are just a screen recording with captions, or a product spinning on a clean background with a price and offer overlaid. The deciding factor isn't the format — it's whether the creative earns attention in the first moment and communicates the offer clearly from there.
Does the platform decide this for you?
Platform mechanics don't force a person into the ad, but they do reward certain habits regardless of format: vertical 9:16 framing for Reels, TikTok, and Stories placements, keeping key text and the subject inside the safe zones so nothing gets cropped or covered by UI, and putting the hook in the opening frame rather than building up to it. Those rules apply whether or not a person appears on screen.
So how should you actually decide?
Treat it as a testing question, not a philosophy question. Run a person-led version against a product-only or text-led version for the same offer and see which one holds attention and converts better for your specific audience — results here depend heavily on margin, auction dynamics, and the specific creative, so a rule that works for one brand can flop for another. This is exactly where creative volume matters: the more variations you can get in front of an audience, the faster you find out what your customers actually respond to. Polaris connects to Claude as an MCP tool, so you can describe a person-led version and a faceless version in plain chat, attach your product photo, and get both rendered as UGC-style video ads without opening a separate app or briefing an editor. A live panel shows progress and the finished ad right inside the conversation, and you can batch several angles at once to compare side by side.
What should you track once ads are live?
Whichever version you run, watch hook rate to see if the opening actually stops the scroll, and tie results back to ROAS so you're judging creative on outcomes, not just watch time. A hook generator can help you draft opening lines for both a person-led and a product-only cut, and a ROAS calculator makes it easier to compare whether the person-led version actually earned its extra production effort.