Funeral home video under the FTC Funeral Rule

Almost every piece of video advice starts from the same premise: earn attention in the first three seconds, keep the cut rate high, put a bed of music underneath, end on a call to action. Apply that to a funeral home and you produce something that feels like it is trying to sell a car to someone whose father died on Tuesday.

This is the one vertical where the standard playbook is not merely suboptimal, it is the failure mode. Slow, static, unscored video wins here, and it wins for a reason that is easy to state: persuasion technique reads as predation when the audience is grieving. Every choice that would raise engagement elsewhere lowers trust here.

What the Funeral Rule actually requires you to be careful about

The FTC's Funeral Rule governs how funeral providers disclose prices. The parts that touch marketing:

Providers must give consumers a General Price List — an itemised written list — at the beginning of any in-person discussion of arrangements, goods, or prices.
Providers must give price information over the telephone when someone asks, without requiring them to come in or give their name first.
Separate casket and outer burial container price lists must be offered before those goods are shown.
Consumers can buy only the items they want rather than a required package, and a provider cannot refuse to handle, or charge a handling fee for, a casket bought elsewhere.
Providers cannot misrepresent legal, cemetery or crematory requirements — the classic example being telling a family that embalming is required by law when in most ordinary circumstances it is not.

Whether and how those disclosure obligations extend to a provider's website has been under active review, and it is not a question to answer from a blog post. Treat the online question as live, check the current rule text with counsel, and in the meantime build creative that would be fine under the stricter reading. That means: never let a video imply a package is mandatory, never let it state or imply a requirement that is not one, and never quote a price on screen without pointing at where the full itemised list lives.

The practical version of that last point is a production rule, not a legal one: no number reaches the screen unless a human typed it. Image and video models render text as visual pattern rather than as characters, and they garble digits routinely. In most verticals a mangled digit is an aesthetic annoyance. On a card describing a basic services fee, it is a compliance incident. Every figure, every disclaimer, every phone number goes on as a text overlay you control. If a generated shot contains incidental legible text, regenerate it or soft-blur the region so it reads as depth of field.

Slow, static, unscored — and how to actually get that

The tonal target is a room, not a montage. Long takes, no camera movement, no music, room tone only.

Getting a genuinely locked-off frame out of a video model takes an explicit instruction. Models bias hard toward adding movement, and omitting camera language does not give you a static shot — it gives you an unrequested drift. Write the camera is entirely motionless into every prompt. If you want a move at all, keep it to one slow reveal and name the mechanism rather than the effect, because "closer" resolves as a dolly or a zoom at random. A parameter-based camera control setting is the only reliable way to get the same move twice.

On audio: room tone, and nothing else. If you generate footage with native audio, exclude reverb explicitly — it arrives uninvited and it is the most artificial-sounding thing in a quiet room. A voiceover should sit forward and dry. If the silence feels like it needs filling, the script is too long. The general case for ducking music under a voiceover is elsewhere; this is the exception where the right level for the music is off.

Length rules invert too. Three to six minutes is reasonable for the pre-need explainer, because this is a website and Google Business Profile asset watched by someone who went looking for it, not a feed asset competing for a scroll.

The pre-need explainer structure

Pre-need is the format worth building properly, because it is the one people research calmly rather than in crisis, and because it is where the confusion is genuinely highest. Six beats:

State what the video is, in the first sentence. "This is a short explanation of what pre-planning involves." No hook, no question, no tension. The absence of a hook is itself the signal.
Separate the two decisions. Recording your preferences and paying for them in advance are different choices, and a lot of people believe they are one thing. Say plainly that you can do the first without the second.
Be specific about what pre-payment does and does not lock. Which items are guaranteed and which are not, what happens if the family moves out of the area, what happens if the funeral home changes ownership, and how the funds are held. This is the section that earns the call, and it is the section most competitors skip.
Point at the price list. Say the itemised General Price List exists, that it is available on request, and exactly how to get it — by phone, by email, in person. Do not put a partial price on screen with no route to the full list.
Say what happens next in concrete terms. Who they will meet, roughly how long it takes, what to bring, whether anything is signed on the day.
Close on a name and a phone number, held on screen for a full five seconds. Not a booking widget, not urgency language, not a countdown.

Cut each beat as a standalone thirty-to-sixty-second piece for the FAQ pages. One production, seven assets.

Imagery never to generate

Some of this is legal exposure and some of it is simply the fastest way to lose a family. Treat the list as hard.

A deceased person, a body, a body being prepared, or the inside of a casket. Never, in any style, in any context.
A generated grieving family. Manufacturing grief to sell grief services is the single most damaging thing on this list.
A generated person delivering a testimonial about your care. A synthetic family member saying you were wonderful during their loss is a fabricated endorsement and treated as one — see the fake review rule and AI testimonials.
Clergy, religious ritual, or denominational specificity you cannot verify. Models produce confident nonsense here, and getting a rite subtly wrong is worse than not depicting it.
Headstones or memorial plaques with legible names and dates. Garbled text plus the implication of a real grave.
Children at a graveside.
A hearse in motion, or a procession shot as cinema. Movement plus scale plus scoring is exactly the register to avoid.
Crematory or preparation-room equipment.
Rain at a graveside, an urn handed over in slow motion, a single flower on wet grass. Metaphor-as-emotional-leverage. It reads as a stock library, not as your firm.
A photorealistic version of your own building that is not actually your building. Families visit. Show the real one.

What to generate instead

Places, not people, and mostly empty. The arrangement room with the lights up and no one in it. The chapel from the back row. The exterior across four seasons. Step-free access, the car park, the door a family will actually walk through. A desk with paper and a pen, framed so no text is legible.

Every one of those answers a question a family is silently asking — how far is the walk, is there parking, will it feel institutional — and none requires a person in frame. The AI B-Roll Generator handles the seasonal and access shots you cannot schedule; the explainer video generator covers the structural sequences.

Caption everything, burned in. Much of this research happens on a phone, quietly, at work or late at night, with the sound off.

FAQ

Should the funeral director appear on camera?

Yes, and it should be the actual director, not a synthesised presenter. This is a trust purchase in a small geography, and the point of the video is that the family recognises the person who opens the door. Use generated footage for the rooms and the seasons; use the real person for anything that asks for belief.

Can we put prices in the video at all?

You can state that price information is available and how to get it, and many providers do quote a starting figure. If you do, type it as an overlay rather than letting a model render it, keep it accurate to the current price list, put the qualifying conditions on screen at the same time rather than in a caption, and re-export the video whenever the list changes. A stale price in an evergreen video is the version of this that causes problems.

Does this mean no social video at all?

No — it means a different register. Short pieces that answer a single practical question work well: what to do in the first twenty-four hours, what documents a family needs, the difference between burial and cremation costs at a category level, what a celebrant does. Informational, static, captioned, no music. What does not work is trend audio, fast cuts, or anything built to be shared.

Is generated footage of a place appropriate here at all?

For seasons, weather and access shots of a building that genuinely is yours, yes — you are recreating a real place under conditions you cannot schedule. It stops being appropriate the moment it depicts something untrue: an amenity you do not have, a chapel larger than yours, grounds that are not your grounds. The test is whether a family arriving in person would feel the video had told them the truth.