Shooting a menu that keeps changing

A cafe adds three specials for autumn, puts two prices up, drops the drink nobody ordered, and changes its oat milk supplier. That is a normal quarter, and it happens again in twelve weeks. Any photography plan built around a single session of good pictures is out of date within one menu cycle, which is why so many boards end up as one professional photograph from 2023 surrounded by five phone snaps taken in a rush on different afternoons.

The fix is not better pictures. It is making a new picture cost four minutes, so that keeping up with the menu stops being a project.

Build a rig, not a shoot

Decide once, then stop deciding. A named spot for the cup, a named spot for the camera at a fixed height, one direction of light, one ground, one crop. Tape on the counter for the cup, tape on the floor for your feet, a note in your phone with the settings. That is the rig.

Once it exists, photographing a new drink is: make the drink, put it on the mark, take six frames, done. The variability that normally ruins a menu, one item shot from further back than the next, one warm and one cold in colour, one on the wooden counter and one on the marble by the window, is gone by construction rather than by care.

The rig also solves a problem that is not about cost at all.

Drinks have a working window and it is shorter than your set-up

Latte art holds its definition for well under a minute before the foam begins to lose its edge. A cappuccino's head goes from textured to slack while you are finding the right angle. An iced latte separates, and the condensation on the glass goes from clean beads to a wet puddle on the surface in a few minutes, which is the difference between a drink that looks cold and a drink that looks abandoned. Ice cubes visibly round off.

So the single largest quality improvement available to most cafes is not equipment. It is ordering the work correctly: build the set completely with an empty stand-in cup, get the light and the framing right against nothing, and make the real drink last. If you are still moving a lamp when the drink lands, you are photographing a dying drink and you will do it again tomorrow.

Hot and cold want opposite light

This is the part most cafes get wrong, because they light everything the way they light the flat white.

Cold drinks are translucent. The subject is what is happening inside the glass: the espresso falling through the milk, the layers, the structure of the ice, the colour of a syrup in suspension. Light has to pass through the drink for any of that to read, which means the source belongs behind or well behind to one side. Lit from the front, an iced latte is a beige cylinder.

Hot drinks in ceramic are opaque. There is nothing to shine through. The subject is the surface: crema, the texture of the microfoam, the art if there is any, the meniscus at the rim. That wants light coming across the surface from the side, so the texture casts tiny shadows and becomes visible. Backlight the same cup and you get a dark drink with a bright rim, which is a silhouette of a mug.

If your rig has one light position, put it side and slightly behind. That is the compromise position that treats both reasonably, and then rotate the glass rather than the light when you move to the cold half of the menu.

Steam is real or it does not exist

Steam is only visible against something dark and only when light comes from behind or hard from the side. On the warm cream ground that suits coffee it will not appear at all, no matter how hot the drink is, because there is nothing for it to be brighter than.

That gives you two honest options. Keep a second short set-up for the two or three shots that genuinely need steam, dark ground and a light behind, or leave steam out of the whole menu. What you should not do is add it afterwards. Drawn steam is the most immediately template-looking element available in this category, and customers who have never thought about photography in their lives will still register that the picture looks like clip art.

Not every item needs a photograph

This is the counterintuitive one, and it saves more effort than anything else here.

Six photographs of six brown drinks in the same cup is a worse menu than three photographs and three well-set names. Three reasons, and they compound.

A photograph on a menu is a signal that says look at this. Spend it on everything and it says nothing, and the drink you actually want to sell loses the emphasis it would otherwise have had. Near-identical images invite comparison, and because the differences between them are tiny, what the eye picks up instead is the inconsistency: this one is slightly darker, that one sits slightly lower in the frame. Variety was the intention and unevenness is the result. And every extra photograph is another thing to reshoot the next time the cup design changes or the recipe moves.

Photograph the drinks that genuinely look different from each other, plus the one with the highest margin. Set the rest as type, and set it well.

Three things date a menu photo, and one of them is avoidable

A change to the cup or glass. This one is unavoidable and it is brutal, because a cup redesign invalidates every image at once. It is worth knowing before you approve new cups, and it is a reason to shoot tight enough that the branded panel is either fully current or fully out of frame rather than half visible and wrong.

A change to the drink that is actually visible. A new syrup with a colour, a switch from a glass to a keep cup, a topping. That needs the camera, and with a rig it needs four minutes.

A price printed into the image. This one is self-inflicted. Prices, names, "new" tags and allergen notes should sit as type over the photograph rather than baked into the pixels, so that a price rise is a text edit and the photograph survives the quarter untouched. That is the same argument as keeping headlines as editable layers on a poster, and on a menu it is worth more, because menus change far more often than posters do.

Pick a lane, because the two cafe looks do not mix

There are two coherent visual grammars in this category and they share almost nothing except the product.

The independent cafe lane is warm and busy. Brown, cream and one accent, the drink on one side and a block of type on the other, a row of small labelled icons, an address, opening hours, an offer, an ordering instruction. It is designed to be read.

The single-origin lane is quiet. One deep field, usually near-black or dark green, the cup at modest size, one small line of type, and a great deal of empty space. It is designed to be believed.

Mixing them satisfies nobody. A minimal dark field with a promo code sitting on it is the most common failure, because a promo code cannot be quiet. The rule of thumb is easy: if there is a price, a discount or a code anywhere in the brief, you are in the busy lane, so commit to it and give the offer its space.

One note on the icon row, since it is the busy lane's signature. It works when the three things are specific: roasted four miles away, oat as standard at no extra charge, kitchen closes at four. It is pure decoration when the three things are quality, taste and coffee, and in that case cutting it improves the design.

Two small colour notes

Avoid pure white as a background field. Coffee wants cream. White reads clinical and works against the warmth the entire category trades on, and it has a practical downside too: phone white balance drifts between sessions, and a white ground shows every bit of that drift as a faint blue or pink cast while a cream ground hides it.

Then keep the cup the same size in the frame across the whole board. Unlike a social feed, a menu is a comparison grid, and everything on it is seen simultaneously. Drinks photographed at different distances read as different sizes, which is a real question a customer will have about a real product they are paying for. Scale is worth taking seriously anywhere a buyer is deciding from an image alone, and a drinks menu is the place it is most often quietly wrong.

The quarterly routine

When the menu changes, block out half a morning rather than a day.

Set the rig up once with a stand-in cup. Group the new items by light setup, all the hot ones together, then rotate the light and do the cold ones. Make each drink at the moment you are ready for it, never before. Shoot six frames, check the highlight on the foam, move on. Then update the type, which is the only part that touches the items you did not reshoot.

Three new drinks is under an hour of counter time. That is the whole point of the rig: it turns the menu changing from something you dread into something you handle on a Tuesday before opening.