How to organize delivery menu categories that convert
On a phone, your category order is your menu design. A guest sees roughly four items per screen, has no server to ask, and is comparing you against six other restaurants in adjacent tabs. Here is the sequence that converts, the item counts that work, and the dishes that should never leave your kitchen in a bag.
TL;DR — Key Takeaways
Delivery menu categories are the grouped sections — Most Popular, Mains, Sides, Drinks, Desserts — a guest scrolls through in a delivery marketplace or your own ordering page. On a phone they replace the entire layout of a printed menu: the category row is the only navigation a hungry person gets.
The stakes are the whole channel. Nearly 75% of restaurant traffic now happens off-premises, and it is a larger share of sales than in 2019 for 58% of limited-service and 41% of full-service operators, per the National Restaurant Association.
A delivery menu is not a dine-in menu on a screen. A printed menu shows forty items at once and comes with a server who can steer you; a delivery menu shows about four items per screen and is compared against six restaurants in adjacent tabs.
The best practice is specific: lead with a Most Popular or signature category of five to seven items, keep every category under eight, place sides immediately after mains, and never let bundles fall below the first two screens.
Where Intermenu fits: Intermenu lets you reorder categories, cap item counts and push photo and price changes live in seconds — turning category sequence into something you test rather than guess.
Why delivery menu order is different from dine-in
Delivery menu order matters more than dine-in order because the guest sees far less of the menu at once, gets no help from staff, and abandons at zero cost. A printed menu is a single canvas; a delivery menu is a queue, and everything below the first two screens competes for attention it will probably never get. All of this assumes the menu already exists as structured data; if it does not, start by digitising your menu first.
In the dining room a guest has a server standing there to say "the short rib is the one people come back for." Online they have one thumb, a six-inch screen, and a partner asking what's taking so long.
Two constraints drive every rule below: screen space, and travel time, since the food has to survive 20 to 35 minutes in a bag. A third is that the marketplace ranks you — category names get matched against what people type, so "From the Josper" is invisible to anyone searching "grilled chicken." For the structure logic underneath, start with how to organize a menu.
How guests actually scan a delivery menu
Guests scan a delivery menu in a fast top-down sweep: store header, category chips, then the first two or three items of the first category, then progressively less of each category below. Attention is front-loaded and decays sharply, so list position is close to a proxy for order volume.
Nielsen Norman Group's research on the F-shaped reading pattern found people read the top of a page horizontally, make a shorter second pass lower down, then scan down the left edge — skipping harder once a page feels repetitive. A delivery menu is the most repetitive page a person reads all day.
The category chip row is your front page. Whatever fits before it scrolls out of view is what most guests treat as your menu — four to six chips, the sixth already half-cut.
Items one and two get read; item seven does not. Put the dish you most want sold at the top, not the middle.
Photos are scroll-stops. An un-photographed item below a photographed one is functionally invisible.
The category sequence that converts
The sequence that converts is: Most Popular, signature category, bundles, mains by format, sides and add-ons, drinks, desserts, extras. Each position exists for a reason, and the reasons are about the order a guest makes decisions in — not how your kitchen is organized.
1. Most Popular / Bestsellers (5–7 items). Your highest-value real estate. It answers the only question a first-time guest has — "what do people order here?" Rotate it monthly from order data.
2. Your signature category. What you are known for, in the guest's words. It reinforces the intent that brought them to you.
3. Bundles and meal deals. A bundle is a shortcut past the whole decision — a guest who takes one never has to scroll. Buried at the bottom, bundles get ordered by almost nobody.
4. Mains, split by format. Burgers, Bowls, Pasta, Grill — not "Entrées." Formats are how guests think and how app search works.
5. Sides and add-ons. Right after mains, the moment the guest has committed to a protein and is deciding how big the order gets.
6. Drinks. Almost nobody adds a drink before choosing food; asking early wastes a scroll.
7. Desserts. A checkout impulse on delivery, so it performs better near the cart than at the top.
8. Extras — sauces, bread, pantry items. Low-consideration and cheap to add. Perfect for the bottom.
The first-screen logic ties it together. On a phone, the header plus the chip row plus roughly two items is everything before the first scroll. If Most Popular is not in that space, your best sales tool is below the fold.
MOST ORDERED— Double Smash · Buttermilk Chicken · Loaded Fries · Vegan Smash
BURGERS— 6 items
MEAL DEALS— Burger + Fries + Drink · Two Can Dine · Family Box
SIDES— Onion Rings · Slaw · Mac Bites · Pickles
DRINKS— 6 items
DESSERT— Brownie · Basque Cheesecake
Seven categories, about thirty items, decision reachable inside two screens. To choose which dishes earn the top slots, use menu engineering in 2026.
How many items should each category have?
Five to seven items per category, with a hard ceiling of eight, and 25 to 40 items total across six to eight categories. Beyond that, decision time rises, conversion falls, and ticket times get worse at exactly the moment a driver is waiting at the pass.
The evidence is old and still the most useful thing in menu design. In Iyengar and Lepper's 2000 study "When Choice is Demotivating", a tasting table offering 24 jams drew more people than one offering 6 — but 30% of those who stopped at the 6-jam table bought a jar, versus 3% at the 24-jam table.
Under 5 items: merge into a neighbor. A three-item category wastes a chip slot.
5–7 items: the sweet spot — two screens, scannable in one pass.
8 items: your ceiling, and only for the signature category.
9 or more: split or cut. "Pasta (14)" becomes "Baked Pasta (6)" and "Fresh Pasta (6)", with two dishes deleted.
Over 45 total: a prep problem, not a menu problem. Delivery amplifies every SKU — more waste, longer tickets, more items 86'd at 8pm.
Cut by contribution margin and order frequency, not gut feel: the method is in how to reduce menu items. Cutting rarely cuts revenue, because delivery demand concentrates — removing the bottom fifth pushes that volume into dishes you already make, so the cut shows up as higher margin at flat sales.
Naming categories so they surface in delivery-app search
Name categories with the words guests actually type: "Burgers," "Chicken Wings," "Vegan," "Pasta." Marketplace search matches queries against category and item names, so a clever house name removes you from results you would otherwise win. Keep names under about 18 characters so chips do not truncate.
Use the noun, not the concept."From the Grill" loses to "Grilled Chicken." Brand voice belongs in descriptions, not navigation.
Diet terms are search terms."Vegan," "Gluten-Free" and "Halal" get typed constantly. With five or more qualifying dishes, give them a category; under five, use tags.
Front-load the keyword. Chips truncate from the right: "Wings & Tenders" survives, "Tenders and Signature Wings" becomes "Tenders and Sig…".
Never name a category after a station. Fryer, Wok and Cold Line mean something to your team and nothing to a guest.
Descriptions do measurable work too. DoorDash's merchant guidance reports that menus with item photos saw an increase of up to 44% in monthly sales, and menus with item descriptions earn up to 18% more each month— the formula is in how to write menu descriptions. If you publish in several languages, Intermenu keeps category names in sync so a rename happens once.
What travels well — and what should never go on a delivery menu
Dishes that hold heat under a lid travel well: braises, curries, stews, roasted meats, rice bowls, bone-in fried chicken, pizza in a vented box. Dishes defeated by steam, carryover cooking or plating precision do not: fries, tempura, rare steak, dressed salads, ice cream, anything assembled at the table.
Fries and anything battered. The enemy is trapped steam. Vented container, never sealed in the same box as a burger, a thicker 3/8-inch or coated fry, bagged last. If none of that works, swap to roasted potatoes or wedges.
Rare steak. Carryover cooking plus 25 minutes in a warm bag turns rare into medium-well. Cook one doneness under and say so on the confirmation, or restrict delivery to cuts that eat well at medium — flat iron, hanger, skirt.
Ice cream and gelato. Don't, unless you have a sub-10-minute radius and insulated packaging. Substitute something stable at room temperature: brownies, basque cheesecake, tres leches in a sealed cup.
Dressed salads, nachos, anything layered. Ship components — dressing in a 2oz cup, greens dry, chips bagged apart from toppings. Guests forgive 30 seconds of assembly; they do not forgive sogginess.
Delicate plating. Foams, tuiles and towers are gone by the time the lid comes off. Re-plate for the container or drop the dish.
Eggs and soups. Yolk keeps cooking, so brunch delivery moves to a soft scramble. Soups go to 80% with a tamper-evident lid, never bagged with anything crispy.
The commercial case comes from guests: in the same National Restaurant Association research, 90% said they would order a greater variety of items if food kept its on-premises quality in transit, and more than half would pay more for premium packaging. Before any item goes live, box it, drive it around the block for 20 minutes, then eat it in front of the team.
Where photos matter most
Photograph the first two categories, every bundle, and any item whose name does not describe itself. Photos act as scroll-stops, and their effect is concentrated at the top of the menu. Below the third category, photo lift falls away because far fewer guests scroll that far.
Most Popular: non-negotiable, every item shot.
Bundles: a photo does the explaining three lines of text cannot — the guest sees burger, fries and drink and gets the value instantly.
Ambiguously named items: house terms and regional dishes an out-of-area guest may not know.
High-margin sides: where a photo makes a $6 add-on look like a $9 one.
Skip: canned drinks, plain rice, sauce cups. A stock image is worse than no image.
One rule outranks those: be consistent within a category. Three photos out of eight suppresses the other five. Shoot the whole category or none of it — see how menu photos increase sales and how to add menu photos. In Intermenu you can generate them without a shoot when budget is the blocker.
Modifiers, combos and upsells without slowing the order
Keep modifiers to a maximum of two required steps per item, make everything else optional, and put attach prompts at the cart rather than inside the item. Every extra required tap between "I want this" and "add to order" is a place an order gets abandoned.
Required means required. Protein, size, spice level where it genuinely matters. If a default is right 80% of the time, set it and make the modifier optional.
Two-step ceiling. Base → protein → add. A third mandatory screen costs completions.
Combos get their own category, high up. Do not make guests build a meal out of modifier groups; most won't.
Attach at the cart."Add fries for $3.50?" converts better at the cart than buried in the item modal, because the guest has already decided.
Anchor the add-on. A $3.50 side beside a $14 main reads as small; offered first, it reads as an expense.
Cap modifier lists at six. Fifteen sauces is choice overload in miniature.
Build them with how to add menu modifiers and price them with menu pricing psychology. The same attach logic works in the dining room — see QR menu upsell tactics— but delivery is less forgiving, because no server recovers a missed add-on. Modifiers also carry a mis-pick risk in a rush, which drops when orders write straight into the till: restaurant POS integration.
How to test and iterate your category order
Change exactly one thing, run it for two weeks against a two-week baseline, and compare view-to-order ratio per category on matched weekdays. One variable per window is the whole method — change sequence and names together and you have learned nothing you can act on.
The metric to build around is view-to-order ratio per category: the share of guests who saw a category and ended up with an item from it in the cart. If your platform reports only item-level data, use item views to item adds.
Weeks 1–2, baseline. Change nothing. Record view-to-order ratio per category, items per order, average order value, and side and drink attach rate. Log confounders — weather, promos, holidays, a competitor opening.
Weeks 3–4, one change. Move one category. Not names, not photos, not prices.
Compare like for like. Two Tuesdays against two Tuesdays. Delivery demand is violently day-of-week dependent and an extra Friday flatters any change.
Respect your volume. Under about 200 orders per window you cannot separate a small shift from noise — run four-week windows instead.
Keep a log: date, change, result, decision.
Test in this order, largest effect first: sequence, then names, then item count, then photos, then modifiers. Most operators do it backwards and buy a photo shoot before checking whether Most Popular is on the first screen. Do not start a test the week you swap the menu for a season — run seasonal menu updates in a separate window — and measure with QR menu analytics.
On Intermenu, reordering a category takes seconds and goes live immediately, so a two-week cycle is realistic. If every change needs a support ticket, the sequence you shipped on day one becomes permanent by default.
Frequently asked questions
What is the best way to organize delivery menu categories?
Lead with Most Popular, then your signature category, then bundles, mains, sides, drinks, desserts and extras. Keep each category to five to seven items and put everything a guest needs to decide inside the first two screens.
How many categories should a delivery menu have?
Six to eight. Fewer forces long, unscannable lists; more pushes the chip row out of view, hiding everything after the sixth chip.
Should "Most Popular" always be the first category?
Yes, for almost every concept. It answers the first-time guest's real question. Rotate its contents monthly from actual order data.
Do delivery menu categories affect search inside delivery apps?
Yes. Search matches queries against category and item names, so plain nouns like "Burgers" or "Vegan" surface in results that "From the Grill" never reaches.
Should drinks come before or after sides?
After. Guests add drinks once the food decision is made, so drinks placed first push your highest-margin attach down the page.
What food should never go on a delivery menu?
Anything defeated by steam, carryover cooking or plating precision: unvented fries, rare steak, ice cream, dressed salads, assembled nachos and fine-dining plate work.
Do I need a photo for every item?
No, but you need photos for every item in a category you choose to photograph. Mixed categories suppress the un-photographed items.
Should my delivery menu match my dine-in menu?
Prices and branding should match, the item list and category order should not. Delivery needs fewer items, travel-tested dishes and a sequence built for a phone.
Intermenu gives you one menu you can reorder, cap, photograph, translate and update in seconds. Build your delivery menu free with Intermenu →