Smell and taste do not travel through a phone, so sound and numbers carry the ad. Three script shapes, twelve hooks and the roast date shot that settles it.
A coffee UGC ad has to sell two senses the camera cannot record. The substitutes are sound and numbers: the grinder, the bloom, the pour, the roast date read off the bag, the brew recipe in grams and seconds, and the cost per cup. Say the taste in plain words, not in tasting-note language.
A date is the one checkable fact in this category, and specialty buyers look for it in a shop before they look at the origin. Hold it still, read it aloud, and say how many days ago that was. A bag with no date on camera reads as a supermarket bag with a brand on it.
The grinder, the hiss at the start of a shot, the pour changing pitch as the cup fills, the ceramic on the counter. Sound is the only channel you have for smell. A music bed over the top removes the one sensory thing that was working, which is why the steam montage format sells nothing.
"Eighteen in, thirty-six out, twenty-seven seconds" or "thirty grams to five hundred, water just off the boil". Numbers are the credibility currency here, and they make the video useful to somebody who does not buy, which is how coffee content travels beyond the ad account.
"Notes of stone fruit and jasmine" is a phrase people repeat and do not understand. "It tastes like the sharp end of a green apple" lands. Pick one comparison, say whether it is sharp or heavy, and say whether you would drink it with milk. That is the whole tasting note.
A twenty-two dollar bag sounds expensive until you say it is about twenty-four cups. Then send the click to the same bag with the same grind option the creator used, because ordering whole bean when you have no grinder is the single most common bad first order in this category.
Beats with timecodes, then a written-out example. Swap the product name and the detail, keep the shape.
A single-origin bag sold to someone who already buys specialty coffee and checks dates.
Product: Ember Row, Kochere, Ethiopia, 340 g (fictional)
[Kitchen counter, morning light, bag held back-side to the lens, finger under a date stamp.] "Roasted the fourth. That is nine days ago. This is the only thing I check."
[Grinder runs, loud, no music. Cut to a V60. Water goes on and the bed swells.] "Listen to that. And watch it rise, that is the gas coming out."
[Scale in frame, timer running.] "Thirty grams, five hundred of water, two and a half minutes."
[Drinks. Eyebrows go up before she speaks.] "Sharp. Like the first bite of a green apple. I would not put milk in it."
[Holds the bag up.] "Twenty-two dollars, about twenty-four cups. Order it ground for filter if you have no grinder."
Spoken word count: 88. Runtime: about 27 seconds.
Converting a daily cafe buyer, where the pitch is arithmetic rather than flavor.
Product: Third Bell House Espresso, 250 g (fictional)
[Counter with a takeaway cup on the left and a ceramic cup on the right.] "Five dollars twenty. Or eighty cents. Same drink, roughly."
[Holds up five fingers, then a phone calculator.] "Five a week is twenty-six dollars. That is thirteen hundred a year on a habit."
[Pulls a shot. Real time. The grinder is loud, the machine hisses, milk steams.] "Four minutes including cleaning up. I am not going to pretend it is instant."
[Tastes the cafe one, then the home one.] "The cafe one is smoother. Mine is stronger. I prefer mine and that is a preference, not a fact."
[Holds the bag.] "Nineteen dollars, about thirty shots. Under seventy cents."
Spoken word count: 92. Runtime: about 29 seconds.
A roaster whose refund requests are really user error, where teaching the fix sells the bag.
Product: Marlow Lane, Colombia, 340 g (fictional)
[Sits at a table, drinks, pulls a face immediately.] "That is sour. Not bitter, sour. Like licking a battery."
[Holds a grinder up, points at the dial.] "It is not the beans. My grind is too coarse, so the water ran straight through."
[Turns the dial two clicks. Grinds again. Brews again. Same bag in frame the whole time.] "Same beans, same water, two clicks finer, ten seconds longer."
[Drinks. Nods once.] "That is it. Sweeter, heavier, no bite at the end."
[Taps the bag.] "If you own this grinder, start at eighteen. Bag is twenty dollars, roasted last Tuesday."
Spoken word count: 87. Runtime: about 28 seconds.
The opening line carries the ad. These are written to be said out loud by a person holding a phone.
The wrong face kills a good script. Match the person to the buyer, not to your taste.
A home barista in a small kitchen with a scale and a timer permanently in frame. Talks in ratios and seconds, adjusts one variable at a time. Credible to the specialty buyer who already owns a grinder and is deciding between roasters rather than deciding whether to care.
Makes it in a cafetiere with no gear at all, sitting at a dining table. Openly suspicious of expensive coffee and says the price out loud twice. Credible for converting the buyer who thinks specialty coffee is a personality rather than a difference in the cup.
A former barista now working from home, filming next to a takeaway cup he has stopped buying. Fast hands, does the weekly arithmetic on his phone on camera. Credible to the daily cafe spender, because he is the same person one year earlier.
You cannot. So every sensory claim gets swapped for something the camera and the microphone can actually record.
| What you cannot show | What you film instead | What fails |
|---|---|---|
| Smell | The grinder at full volume and the bloom rising in the brewer | "It smells incredible" |
| Taste | An unrehearsed first sip, then one everyday comparison | "Notes of jasmine and stone fruit" |
| Freshness | The roast date on the back of the bag, read aloud | "Always roasted fresh" |
| Body | The pour changing pitch, crema holding on the surface | "Rich and full-bodied" |
| Value | The per-cup number said out loud with the bag in hand | "Great value for money" |
The rule is that a viewer should be able to check the claim themselves, either with their ears or with a calculator.
The steam montage. Slow motion, warm grade, no person, no sound. It is the default coffee ad and it communicates nothing a buyer can act on.
Tasting-note language. Words borrowed from a cupping sheet make the viewer feel excluded. The ad is not for a judge, it is for someone deciding between two bags.
No date, no numbers. Without a roast date and a recipe, the ad is a mood, and moods do not beat the supermarket on price.
Music over the pour. A track chosen for vibe deletes the only sensory evidence the format has. Leave the room noise in, including the kettle.
Everything sensory stays on your camera. The generated half is the person talking to the lens: the roast date line, the per-cup math, the recommendation at the end.
On MakeViral's AI creator page a persona costs 1 credit on Z-Image or 2 credits on Nano Banana 2, and a talking lip-synced video on Kling Avatar 720p costs 2 credits per second of speech, from 3 to 15 seconds. A Z-Image persona plus a six-second read plus the 2-credit hook-and-demo stitch is 15 credits: about $1.50 on the $49.99 plan for 500 credits and about $1.00 on the $199.99 plan for 3,000.
Then cut to your own audio for the grinder and the pour, because a generated soundtrack of a grinder is not evidence of anything. Record that section with the phone close to the machine and do not add music over it. MakeViral does not post or schedule: you download the file and upload it yourself. See the AI product video generator.
Coffee gets brands into trouble in two places: health language, and claims about the bean that nobody in the chain can support.
A creator you generate has no palate and has tasted nothing, so a generated tasting note is a fabrication rather than a review. Keep the sip and the reaction on your own footage, and treat the generated read as your brand speaking. Labelling requirements for a generated creator vary between platforms and they move, so read the current rules for the channel you are buying on before this ad goes live.
A supplement ad meets a scepticism a coffee ad never meets. Three script shapes, twelve hooks and the label shots that answer it.
Read morePlaybookThe delivery moment is this category's native opening. Three script shapes, twelve hooks, and the scale and color shots that stop returns before they start.
Read morePlaybookThe room is the proof, and the room has to be real. Three script shapes, twelve hooks, and the scale, light and renter tests that decide the purchase.
Read moreAll UGC ad playbooks · AI influencer generator · AI UGC ad generator
Pick a creator, paste the script, and get a talking video. Two credits per second on Kling Avatar, two credits for the creator portrait.
Cancel anytime. 14-day refund window on unused credits.
