How to turn product photos into short video ads for Facebook and Instagram Reels

Start from the photo a buyer already believes

Video cannot rescue a weak photo. If the product colour is wrong, the background is cluttered, or the product is small in the frame, animating it only makes the flaw move. Pick your best listing photo instead: the product large and sharp, the background simple, the light even, and nothing in the frame that is not included in the order. A clean studio shot works better as a first frame than a busy lifestyle scene, because the motion tool has less to invent and less to get wrong. If your only photo is a supplier image with a watermark or another logo, replace it before you start. An ad that carries someone else's mark is a trust problem before it is a creative problem.

Crop that photo to vertical before you animate it. Reels are 9:16, and a horizontal photo cropped after the fact will cut the product in half or push it under the app buttons. Export a 1080 by 1920 version with the product centred and clear space above and below it, then use that vertical frame as your first frame. The same discipline applies to the product studio workflow in our single-photo demo video guide, which covers choosing a first frame in more detail. Here the goal is narrower: one photo, one ad, ready for Reels.

The 15-second structure that fits Reels

Meta allows far longer Reels than anyone should run as a cold ad, so ignore the ceiling and build to fifteen seconds. Fifteen seconds is long enough to show a result and short enough to finish before the swipe. Every second has a job, and if a second has no job, cut it.

Write the words before you generate the motion. A fifteen-second ad needs at most three short lines: the hook, one benefit, and the call to action. Anything longer will be read by nobody, because Reels viewers read the image first and the caption second. Put the benefit in the image itself where possible, for example a hand holding the product next to a phone for scale, rather than writing a paragraph about size. If you cannot say the benefit in six words, you have two benefits, and you should pick one.

Specs, safe zones and the one-motion rule

Export vertical video at 1080 by 1920, in MP4, and keep all text, logos and the product's key detail inside the safe zone. For Reels ads, Meta's published guidance keeps roughly the top 14 percent, the bottom 35 percent and 6 percent on each side clear of text and logos, because the profile, caption and action buttons sit on top of the video there. On a 1080 by 1920 frame that is about 269 pixels at the top, 672 pixels at the bottom and 65 pixels on each side. Check the final export on your own phone before you upload it, because a headline that looks fine on a desktop preview is often hidden behind the caption on a phone. Ads Manager also shows the live limits and safe-zone preview at upload, so treat its warning as final, not as a suggestion.

For the motion itself, describe exactly one movement and nothing else. Image-to-video tools fail when they are asked to do three things at once, because each extra instruction is another chance for the product to change shape, colour or size mid-clip. Keep the camera still, keep the product still except for the one motion, and name the lighting so it does not drift.

Generate three or four variants of each motion, keep the one where the product matches the photo at the start and at the end, and discard the rest. A clip where the bottle changes shade halfway through is worse than a still image, because motion makes the change impossible to miss. Finish with captions burned in or supplied as a file, because most feed video is watched with the sound off. Our auto-captions guide compares accuracy and burn-in versus SRT, and our subtitles and voiceover workflow covers timing and the final quality check.

Test three versions, then stop and learn

One ad is a guess. Three ads are a test, if only one thing changes between them. Keep the product, the price and the call to action identical, and change only the hook: version one opens on the result, version two opens on the problem, version three opens on the close-up. Run them to the same audience for the same number of days, then keep the winner and make the next test about the benefit line, not about a new product. Sellers who change the hook, the price and the audience at once learn nothing, because they cannot tell which change did the work.

Judge the test by the cheapest honest signal first. If viewers do not watch past three seconds, the hook failed, and no price change will fix it. If they watch but do not click, the benefit or the offer failed, and the video itself may be fine. If they click but do not buy, the landing page or the listing failed, and making a fourth video is the wrong next step. Write down the winner and the loser in one line each, and reuse the winning structure for the next product instead of starting from a blank page every time. A small seller's advantage is speed, and speed comes from a structure that repeats.

Cost stays small if you keep the test small. In Movapix a video costs about 2.4 cents per successful generation, with no subscription, and an image costs about 0.8 cents, so three hook variants at a few generations each cost well under a dollar in credits. The real budget is the ad spend and your time, so cap the test before you start: one product, three hooks, a fixed daily spend you can afford to lose, and a date when you will decide. Finish the set in the small-shop photo guide workflow, and use the upscale guide only if a winning frame needs to be larger for a feed version.