Menu engineering for cafés: pastries, add-ons and the coffee ladder

A café sells a 3 € product two hundred times a day, which means every design decision on the menu is multiplied by two hundred before lunch. A pastry that sells to one extra guest in ten is 20 croissants a day. A 0.40 € oat milk surcharge that half your latte drinkers accept is another 30 to 40 € daily. Restaurant menu engineering usually talks about steaks and margins in the tens of euros; café menu engineering is about small numbers that compound, and the menu board or card is the lever.

Build the coffee ladder first

Every café menu has an implicit ladder: espresso at the bottom, a large flavoured latte at the top, with eight to twelve rungs in between. The ladder should climb in even steps of 0.30 to 0.50 € so each next drink feels like a small upgrade rather than a jump. Espresso 2.00, macchiato 2.30, cortado 2.70, flat white 3.20, cappuccino 3.40, latte 3.60 is a ladder; espresso 1.80 then cappuccino 3.80 is a cliff.

Order the board by that ladder, top to bottom or left to right, so the eye reads a progression. Cost each rung honestly: a double espresso uses 18 g of coffee at roughly 0.45 € for specialty beans, a 12 oz latte adds 250 ml of milk at about 0.25 €, plus the cup if it leaves the shop. The margin difference between the espresso and the latte is smaller than the price difference suggests, which is why milk drinks are the profit engine of most cafés.

Sizes: two is enough, three is a decision

A single size removes a decision and speeds up the queue, which is right for a specialty bar where the barista controls the ratio. Two sizes with a 0.50 to 0.70 € gap is the standard café pattern and the middle option effectively disappears, so most guests pick the larger one. Three sizes only make sense with high takeaway volume, and then the middle size should be the one you want to sell, priced closer to the large than to the small.

On the board, list sizes once as column headers, not on every line. A menu that repeats small, medium, large twelve times is unreadable from two metres away, and a café menu is read standing, from two metres away, in the fifteen seconds before the guest reaches the counter.

Add-ons are pure margin if they are visible

Alternative milks, syrups, an extra shot and decaf are the four add-ons most cafés sell. Their food cost is small, between 0.10 and 0.30 €, and their price is typically 0.40 to 0.70 €. The only thing that limits sales is whether the guest sees them before ordering.

• Group add-ons in a short block directly beneath the coffee list, in the same type size as the drinks, not in tiny print at the bottom.
• Price them as a flat surcharge, plus 0.50, rather than listing a separate price for every drink with every milk. The grid version quadruples the board without adding information.
• Name the milk brand or type if it is a good one. Barista oat from a recognised roaster-grade producer justifies its surcharge; anonymous plant milk invites the question of why it costs extra.
• Decaf at no surcharge is the modern default. Charging for it reads as a penalty and the volume is too low to matter.

Pastries sell where they are seen twice

The pastry case does the heavy lifting, but the menu should back it up. A guest who reads croissant, 2.60, on the board while queuing has already half decided by the time they see the case. If the board only lists drinks, the pastry becomes an impulse buy that depends entirely on how full and fresh the case looks at that moment, which by 11 am it often does not.

Put a short food block on the menu, six to ten items, with prices. Lead with the two items that carry the best margin and the fewest waste problems: often a plain croissant or a cookie rather than a filled pastry that must be sold the same day. Describe one or two with a single specific detail, cardamom bun, brown butter, rather than adjectives.

Then bundle. A coffee and pastry combination at a 0.50 to 0.80 € discount off the separate prices is the most reliable way to raise the average ticket in a café. The bundle should be visible as its own line, priced as a whole number where possible, and the discount should be real, because regulars will do the arithmetic.

Board or card: the layout rules differ

A wall board is read at distance in seconds, so the type has to be large and the structure has to be immediate. As a rule of thumb, drink names need a cap height of at least 25 mm to be read at 2.5 m, which means roughly 90 pt at print size; prices can be one step smaller. Three columns maximum, aligned to a strict grid, no more than 25 lines in total. Anything else belongs on a table card.

A printed table card for a café with seating is a different object. It is read at 40 cm by someone who has already sat down and has time. It can hold the full food menu, the origin of the coffee, a short line on the roaster and the brewing methods. An A5 card, single or double sided, on 300 to 350 g board is the standard, and it should be replaced every four to six weeks because cafés are hard on paper: coffee rings, sugar, wet tables.

Test the ladder against the till

Café menu engineering is easy to check because the volume is high. Change one thing, wait two weeks, compare the sales mix. If the cappuccino share rises after you moved it above the latte on the board, keep it there. If the add-on line doubled sales of oat milk, it paid for the reprint in a day. Cafés that treat the menu as a living document, adjusted every couple of months, consistently outperform those that printed a board at opening and never touched it.

The reprint cycle is where good intentions usually stall, because redoing the card each time means finding the file and the designer. In Menu Atelier the Café Chalk style is built for exactly this: paste the drinks and food as a list, add prices, and the layout handles the ladder and the add-on block. A5 single or double sided is one click, and the export is a print-ready PDF with bleed and crop marks, so the six-week reprint is a ten-minute job rather than a project.