Menus Written in Shopper Language, Not Org Charts

A good menu is a set of predictions. Each label tells a visitor, in their own words, what they will find behind it. Menus fail when they mirror the business instead: departments, supplier names, internal ranges or clever brand words that mean nothing to a first-time shopper. Name top-level categories by product type first, then by use, person or occasion where shoppers genuinely think that way. Keep the top level short enough to scan in one glance, put the most sought items where eyes expect them, and test the labels with people who do not work in the business.

Shoppers divide a catalogue by the job in front of them. A parent thinks school shoes, not footwear by department. A gift buyer thinks by recipient and occasion. A tradesperson thinks by task and compatibility. Your range may support several divisions, but the main menu must choose the one most visitors use most often, and support the others through search, filters and category pages. Evidence comes from search logs, support questions and watching people browse. If three sources point to the same grouping, trust it over a tidy internal structure that looks elegant in a planning document.

A label passes when a stranger can guess the contents and, just as importantly, guess what is not there. Plain product words usually beat invented range names. If a branded sub-range matters, pair the brand word with a plain description until shoppers learn it. Avoid labels that overlap so heavily that every product could live in two places; overlap creates hesitation at the exact moment the menu should create confidence. Where two groupings are both natural, choose the primary home and cross-link from the secondary route, rather than duplicating the whole category in two places with diverging contents.

Every extra level asks the visitor to keep faith that the product is still ahead. Two levels handle most small catalogues well; a third should be rare and clearly signposted. Long fly-out menus that vanish when the pointer slips are a known frustration, and on mobile a deep tree hidden in a drawer can become a maze with no view of the whole. Show the structure where you can: a category landing page with its subcategories visible often beats a clever multi-level menu. Breadcrumbs then help the visitor understand where they landed, as covered in breadcrumbs and wayfinding.

Put the categories most visitors seek earliest, using your own movement data and sales knowledge rather than alphabetical tidiness or internal importance. Keep the order stable once published. Returning visitors navigate partly from memory, and a menu that reshuffles each season taxes the loyalty you want to reward. Sale or seasonal entries can appear, clearly labelled and time-limited, but they should not displace the permanent routes shoppers rely on. When you do change order or labels, note the date and watch category entry and onward movement for the following weeks before judging the change.

A wide menu for a large catalogue can help if it groups subcategories under clear headings, shows only levels the visitor needs, and loads without shifting or trapping focus. It hurts when it becomes a sitemap dumped into a dropdown: dozens of links, equal weight, no guidance about where most shoppers go. Restraint means choosing. Feature the routes with real demand, link to the full category page for completeness, and leave rarely used groupings to search and filters. A menu is an editorial product, not an inventory. Its job is to start journeys well, not to prove that every destination exists.

The cheapest menu test needs no software. Write ten product-finding tasks from real searches and support questions. Ask people outside the business where they would look, using only your labels, and note hesitations and wrong first choices. Wrong choices that repeat are label problems, not user problems. Fix the words first; they are cheaper than a redesign and often solve the issue. Only when the words are right and people still struggle is the structure itself at fault. Record the tasks and results so a future menu change starts from evidence, not from a new opinion.