Your website audit flags things in the language Google and Lighthouse use — "LCP", "CLS", "schema", "tap targets". Here's what each of those actually means for your business, and why it's worth fixing.
No test at the end. You don't need to memorize any of this — just know where to find it when your audit (or your developer) throws one of these at you. Every definition lands first, then tells you plainly why you should care.
No terms match your search. Try a shorter query, or clear the search.
The handful of things every small-business website needs before anything else. If these aren't right, nothing else matters.
A menu built as real text on a web page, instead of as a PDF file or a photograph of a printed card. Real text means phones can render it at readable sizes, Google can index it, and screen readers can read it aloud.
Why it mattersPDF menus are the single most common reason a restaurant site loses mobile visitors. They download slowly, open in a separate viewer, force pinch-zoom, and often fail on older phones outright. An HTML menu is searchable, fast, updatable in seconds, and gives Google a crawlable list of what you actually serve.
See the research The 70% cart abandonment rate
The large top section of your homepage, above everything else. It's the only area most first-time visitors will see before deciding whether to keep reading or leave.
Why it mattersYour hero has about four seconds to answer: what kind of business is this, where is it, and what can I do here right now. If those three answers aren't clear without scrolling, you're losing visitors who were ready to become customers.
See the research The 70% cart abandonment rate
A phone number on your site that, when tapped on a phone, opens the dialer with your number pre-filled — instead of being plain text the visitor has to copy. Built with an href="tel:+1555…" link.
Why it mattersIf a customer has to copy your number, switch apps, paste, and dial, many won't. One tap, one ring — that's the shortest path from "I'm hungry" or "I need a haircut" to paying you.
See the research Usability of local business websites
Whether today's hours are visible on your homepage within five seconds of landing, without scrolling or clicking. Distinct from just having hours somewhere on the site.
Why it matters"Are they open right now?" is the single most common question a visitor asks. Hide the answer three clicks deep and they'll ask your competitor's site instead.
See the research Usability of local business websites
Your street address as real, selectable text on the page — not baked into an image or a graphic. Ideally wrapped in a Google Maps link so tapping it opens directions.
Why it mattersIf a customer can't long-press and copy your address on their phone, they can't paste it into Maps, Uber, or a text to a friend. Text addresses also feed search engines' local-business index; image addresses don't.
See the research The 70% cart abandonment rate
A single sentence — usually in the hero — that tells a first-time visitor your cuisine, neighborhood, and vibe. Example: "Wood-fired Neapolitan pizza in downtown Silver Spring, open late."
Why it mattersFour seconds after landing, a visitor has either classified your restaurant or left. A pitch sentence does the classifying for them so they don't have to guess from a logo.
The human-readable address people type to reach your site — "rowhousegarden.com". You buy it yearly from a registrar (Namecheap, Google Domains, Cloudflare) and point it at whichever host runs your site.
Why it mattersYour domain is the one piece of your brand that lives on signs, menus, Instagram bios, and printed receipts. Own it outright (not through a builder that "includes" it), keep the yearly renewal on auto-pay, and pick something short enough to say out loud without spelling.
The server, somewhere on the internet, that sends your website to a visitor's phone when they type your domain. Small-business hosting runs $5–$30/month (Cloudflare Pages, Netlify, SiteGround, DreamHost); restaurant-specific builders like Squarespace bundle hosting into their subscription.
Why it mattersBad hosting makes every other optimization pointless — a fast site on a slow host still loads slowly. Good hosting gives you free HTTPS, free global caching, and 99.9%+ uptime without you thinking about it.
See the research The 70% cart abandonment rate
Email that runs on your own domain (hello@rowhousegarden.com) rather than a free Gmail or Yahoo address. Usually $6–$12/month per mailbox through Google Workspace or Microsoft 365.
Why it mattersFirst impressions aside, free email addresses are five times more likely to land in spam when you reply to reservation inquiries or invoices. A domain-matching email also signals to customers that you actually own your online presence.
Small labels next to menu items indicating gluten-free, vegan, vegetarian, dairy-free, or nut-containing preparations. Best delivered as text or SVG symbols with a legend — not as tiny icons that vanish on a phone screen or screen reader.
Why it mattersDietary restrictions drive where groups eat. If a party of four includes one coeliac or vegan and your menu is silent, the group picks the restaurant that was clear. Labels also reduce staff interruptions: diners self-select before ordering instead of asking the server about every dish.
See the research The 70% cart abandonment rate
How your site performs on a phone, on cellular, in a real customer's hand. Most of your visitors are here — and Google ranks pages on this first.
Google's three performance scores for a real page load on a phone — one for loading speed (LCP), one for layout stability (CLS), and one for responsiveness (INP). You'll see them as colored bars in any audit report.
Why it mattersGoogle explicitly uses Core Web Vitals as a search ranking signal. Two restaurants with identical menus and identical Google Business Profiles will rank differently if one's vitals are green and the other's are red.
See the research How Lighthouse scores performance
How long it takes for the biggest visible chunk of your homepage — usually the hero image or the headline — to finish drawing on a phone. Under 2.5 seconds is "good", over 4 seconds is "poor".
Why it mattersLCP is what a visitor feels as "the page loaded". Every extra second pushes more of them to the back button. Google's own research puts the bounce-risk jump at ~32% when load time goes from 1s to 3s.
See the research How Lighthouse scores performance
A score that measures how much your page jumps around as it loads — when a banner slides in, an image arrives late, or a late-loading font nudges text down. Lower is better; under 0.1 is good.
Why it mattersIf a customer taps "Reserve" and the button jumps at the last millisecond because a photo arrived and pushed it down, they tap the wrong thing and leave. A stable layout keeps conversions intact.
See the research How Lighthouse scores performance
How long your site takes to respond after a visitor taps, clicks, or types something. Measured in milliseconds; under 200ms is good, over 500ms feels broken.
Why it mattersWhen a page "feels slow", this is usually the culprit — not load time. A heavy third-party chat widget or an untuned tracking pixel is often the fix.
See the research How Lighthouse scores performance
A single line of code in the <head> of every page (<meta name="viewport" content="width=device-width, initial-scale=1">) that tells phones "render this page at phone width, not desktop width".
Why it mattersWithout it, phones render every page as if it were a desktop monitor and shrink it — everything looks tiny, unreadable, unusable. The fix is literally one line; the before-and-after on mobile is night-and-day.
See the research The 3-second mobile load rule
Any clickable element — button, link, menu item — that a finger taps on a phone. Google and Apple both recommend a minimum 44×44-pixel tap area with space between neighbors.
Why it mattersCramped tap targets produce the "fat-finger" tap that opens the wrong thing. After one or two of those, most visitors leave. Bigger targets also make your site usable for older customers and anyone with motor or dexterity limits.
See the research Fitts's Law
The default size of regular paragraph text on your site. On phones, 16 pixels is the absolute floor; 17–18px is more comfortable for reading menus, descriptions, and post content.
Why it mattersBelow 16px, iOS Safari forcibly zooms your input fields when a visitor taps them, which breaks layouts. And anything smaller than that is simply harder to read on a phone in a hurry or in dim light.
See the research How Lighthouse scores performance
The open-source auditing tool Google ships inside Chrome and inside PageSpeed Insights. It simulates a mobile page load and scores your site across performance, SEO, accessibility, and best practices.
Why it mattersLighthouse is the same engine Google uses to rank your site — so its score isn't advisory, it's the report card. Our restaurant audit runs a real Lighthouse scan on your URL and translates the numbers into specific fixes for your business.
See the research How Lighthouse scores performance
Google's free web tool at pagespeed.web.dev that runs a Lighthouse audit on any URL and shows you your Core Web Vitals plus a prioritized list of fixes. Built on top of the Lighthouse engine.
Why it mattersPSI is the tool Google itself uses to look at your site, so its score is as close to Google's opinion of you as you can get without working there. Run your site through it once a month; run it again after any big change.
See the research How Lighthouse scores performance
The moment the browser paints the first piece of your page — usually a headline, a logo, or a patch of background color. Measured in seconds. Under 1.8s is good, over 3s is poor.
Why it mattersFCP is the "something is happening" signal. A blank white screen for three seconds reads as broken even if the page is about to load; first paint before 1.8s tells the visitor the site is working and buys you time to finish loading.
See the research How Lighthouse scores performance
How long it takes your web server to send the very first byte of a page after a visitor requests it. Measured in milliseconds. Under 200ms is good, over 600ms points at a slow host or a misconfigured site.
Why it mattersBad TTFB is almost always a hosting problem, not a design problem. Every other speed metric inherits it — you can't render what hasn't been sent yet. If your TTFB is poor, switch hosts or add a CDN before optimizing anything else.
See the research How Lighthouse scores performance
A global network of servers — Cloudflare, Fastly, AWS CloudFront — that keep cached copies of your site close to every visitor in the world, so the page travels 30 miles instead of 3,000. Most good hosts wire one up for free.
Why it mattersA CDN is the single biggest "I don't have to think about this" speed upgrade available. Even if half your customers are in one city, the other half get the same fast site, and your origin server does dramatically less work.
See the research How Lighthouse scores performance
A design approach where the same page automatically rearranges itself to fit whatever screen it lands on — phone, tablet, desktop — without a separate mobile site. The standard since about 2015.
Why it mattersRoughly two-thirds of your visitors arrive on a phone. A non-responsive site treats them as second-class. Google also explicitly penalizes sites that aren't mobile-friendly under its mobile-first indexing rule.
See the research How Lighthouse scores performance
A technique where images and videos further down the page don't start downloading until the visitor actually scrolls near them. Built into modern HTML via loading="lazy".
Why it mattersA homepage with twelve food photos doesn't need to download all twelve before the top of the page is usable. Lazy loading lets your hero paint in under a second while the bottom of the page gets ready in the background.
See the research How Lighthouse scores performance
Turning a visitor into a reservation, an order, or a phone call. The difference between traffic and revenue.
The specific button or link that asks a visitor to do the next thing — "Reserve a Table", "Order Online", "Book a Class", "Get a Quote". Usually bright, clearly labeled, and repeated at natural moments on the page.
Why it mattersTraffic without a clear next step converts at a fraction of the rate. Every page should have one unmistakable primary CTA — the one action you'd most like a visitor to take — and it should be visible on the first screen, not hidden at the bottom.
See the research Fitts's Law
A persistent link — in the nav, in the hero, and in the mobile footer — that starts a reservation flow. Ideally wired to whichever platform your hostess stand already trusts (OpenTable, Resy, Tock, Square) so bookings land in the same system your team already checks.
Why it mattersA reservation that has to be emailed, or lives three clicks deep, is a reservation half your visitors abandon. One tap, one booking, one confirmation — that's the standard diners have been trained on.
See the research The 70% cart abandonment rate
A button on your menu page that links directly to your pickup or delivery checkout — not a landing page, not an interstitial. Either a handoff to your POS provider (Toast/Square/ChowNow) or a direct checkout running through your own payment processor.
Why it mattersEvery extra tap between a hungry visitor and a checkout drops completion rates measurably. Direct checkouts also let you keep more of the margin than marketplace apps like DoorDash or Grubhub.
See the research The 70% cart abandonment rate
Tapping your address on a phone should launch Apple Maps or Google Maps with your business pre-filled as the destination. Built with a Google Maps URL (or the native maps: protocol) instead of plain text.
Why it mattersCustomers already standing on a sidewalk with your site open shouldn't have to retype your address into their map app. A working directions link is the shortest path from "I'm coming over" to "I'm here".
See the research The 70% cart abandonment rate
A thin bar pinned to the bottom of every mobile page holding one or two of your most important actions — usually Call, Reserve, or Order. Stays visible even as the visitor scrolls through the menu or the blog.
Why it mattersA visitor who's sold somewhere in the middle of your menu shouldn't have to scroll back to the top to act. A sticky footer keeps the next step one thumb-tap away at any moment.
See the research Fitts's Law
A quote, star rating, or press mention surfaced on your homepage — ideally in or near the hero, not buried on a Testimonials page — that tells a first-time visitor other people have tried and liked this place.
Why it mattersStrangers trust other strangers more than they trust your marketing copy. One real quote from a real review ("the best carbonara in Bethesda — Washington Post") above the fold does more conversion work than three paragraphs of your own prose.
See the research The 70% cart abandonment rate
The part of your homepage a visitor sees without scrolling. The term is borrowed from newspapers — the half of the front page that was visible before you unfolded it. On mobile that's usually the first ~600 pixels of the page.
Why it mattersA sizable fraction of your visitors never scroll. Whatever you need them to know or do — what you are, where you are, how to book — has to live above the fold, or it might as well not exist.
See the research The 70% cart abandonment rate
The percentage of visitors who land on your site and leave without clicking anything — no menu page, no reserve button, no second page. A high number means the page didn't answer what they came to ask.
Why it mattersBounce rate is an honest mirror. 70%+ on your homepage usually means the hero isn't answering the "what is this place" question fast enough, or the thing they came for (menu, hours, reservation link) isn't obvious.
See the research The 70% cart abandonment rate
The percentage of visitors who do the specific thing you wanted them to do — book a reservation, place an order, fill out a contact form. A 3% reservation conversion rate on a restaurant site is solid; a 10% online-ordering rate on the menu page is excellent.
Why it mattersTraffic is vanity; conversions are revenue. Doubling your conversion rate is usually cheaper than doubling your traffic — and it compounds every time a new visitor shows up.
See the research The 70% cart abandonment rate
The sequence of steps a visitor goes through before becoming a customer — landing page → menu page → reserve button → reservation form → confirmation. "Funnel" because each step loses some of the people at the one before it.
Why it mattersMost conversion problems live at one specific step, not everywhere. Mapping your funnel tells you whether visitors are bailing at the menu, at the reservation form, or at the confirmation email — so you know which one to fix first.
See the research The 70% cart abandonment rate
When a visitor starts filling out a form — reservation, contact, inquiry — and leaves before submitting. Usually caused by too many required fields, unclear errors, or a form that doesn't work on their phone.
Why it mattersA visitor who starts a form has already decided they want to give you their business. Losing them halfway through is a uniquely expensive failure. The fixes are almost always smaller forms, clearer labels, and mobile-tested flows.
See the research The 70% cart abandonment rate
A page on your site where someone can buy a gift card in minutes, usually delivered as an email with a code the recipient redeems in-house or online. Square, Toast, Shopify, and specialty providers like Yiftee or GiftFly all offer a drop-in widget.
Why it mattersGift cards are the single highest-margin transaction a restaurant runs — roughly a quarter never get redeemed, and the ones that do bring a full table, not a solo diner. A "Gift cards" link in the header adds a revenue stream that costs almost nothing and runs itself year-round, with a December spike that can cover a slow January.
See the research The 70% cart abandonment rate
A simple form on your site that collects an email address in exchange for an actual reason to give it — "first to know about the spring menu," "event invites," a small perk on first visit. Usually wired to Mailchimp, Beehiiv, ConvertKit, or Klaviyo.
Why it mattersA house email list is the only marketing channel you fully own — no platform can throttle it, charge you to reach your own audience, or disappear overnight. A thousand-address list becomes a reliable way to fill Tuesday seats, soft-launch a new menu, or sell out a ticketed dinner in a single afternoon.
See the research The 70% cart abandonment rate
A dedicated page explaining your catering or private-event offering — minimums, packages, sample menus, lead time, and an inquiry form that lands in the right inbox. Separate from the regular menu so a corporate planner can assess fit in under a minute.
Why it mattersCatering and buyouts are the most profitable covers you'll serve: guaranteed headcount, batched prep, and a single-point-of-contact client. A page that answers the obvious questions (how many people, what it costs, how much notice) filters tire-kickers out and delivers qualified leads — the ones ready to write a purchase order.
See the research The 70% cart abandonment rate
How many photos are attached to your Google Business Profile — visible in the photo carousel at the top of your Google search panel and Maps listing.
Why it mattersDiners scrolling the carousel decide in 4–6 photos. Under ten and the carousel runs out before they've seen a dish, the room, and the patio — they swipe over to a competitor whose 80 photos let them eat with their eyes first. Adding 15–20 strong photos (food, room, exterior, smiling staff) takes one focused hour with your phone and pays back for years.
Whether your site links out to at least one major third-party delivery aggregator (DoorDash, Uber Eats, Grubhub, Slice, ChowNow). Detected by following the order-online link from your homepage.
Why it mattersA hungry-at-9pm customer who can't find a delivery option opens DoorDash and orders from whoever IS on the platform. Aggregator fees hurt — but losing the customer entirely hurts more. Fast-casual restaurants in particular live or die on aggregator presence; fine-dining can skip it by design.
The preview card iMessage / WhatsApp / Messenger / Facebook show when someone shares your URL — a title, a short description, and a photo, pulled from <meta property="og:image"> and friends in your page <head>.
Why it mattersWhen someone texts your URL to four friends to plan dinner, that preview is half the click decision. Missing — blank box, just the domain — and half those friends never tap through. One image + one title in your platform's SEO settings; the upside is dozens of lost meal plans per year that suddenly become bookings.
The signals that tell a first-time visitor you're a real, careful, operating business — not a half-finished side project.
The little padlock icon in the browser's address bar. It means your site is served over a secure, encrypted connection — everything between the visitor and your server is scrambled in transit.
Why it mattersModern browsers flag non-HTTPS sites with a big "Not Secure" warning. Google has used HTTPS as a ranking signal since 2014. Most good hosts give it to you for free via Let's Encrypt or Cloudflare; if yours charges extra, move hosts.
Showing the price of every dish on your menu page — or, for tasting-menu formats, prominently stating the total. No "contact us for pricing", no hidden surcharges buried in small print.
Why it mattersFirst-time diners scanning your menu on a phone interpret hidden prices as either expensive or evasive. Neither wins the booking. Transparency converts better than mystery, every time.
Anything on your site that subtly tells a visitor "someone is still running this place" — a seasonal menu badge, a "this week's specials" block, a recent blog post, a dated press mention, or even a visible last-updated stamp in the footer.
Why it mattersA stale website feels like a closed business. Visitors who aren't sure if you're still operating don't drive over to find out — they pick the place whose site looks maintained.
A form where a visitor can type a question or an inquiry and actually get a response. "Working" is the important word — the form sends, the email arrives in an inbox someone checks, and a human replies.
Why it mattersBroken contact forms are everywhere, and every one is silently costing someone business. Test your own by sending yourself a message once a quarter. If nothing lands in five minutes in an inbox you actually check, replace it.
Photographs of your space, your food, and your team — not royalty-free stock images licensed from an agency.
Why it mattersVisitors can smell stock photography at a glance, and it reads as dishonest — "they wouldn't show their own place unless there was something wrong with it". One two-hour session with a local photographer outperforms a whole library of stock for years.
A structured shot list a restaurant owner hands to a photographer at the start of a session: which dishes get shot, where each shot is destined to publish (Yelp, Google, Instagram, the menu), what aspect ratio that destination needs, what angle and lighting the photographer should compose for, and how the files come back named.
Why it mattersThe single biggest gap between an independent restaurant's photo session and a usable result is the brief. Owners who hire a photographer without one walk away with 80 generic plate photos and 0 hero shots that actually fit a 16:9 web hero or a 1.91:1 OG card. Owners who walk in with a brief get the right 36 source frames in half a day, every one cropped to a known destination.
The ratio of an image's width to its height, expressed as w:h. Eight surfaces own canonical ratios: web hero 16:9, OG card 1.91:1, Yelp featured 3:2, Google Business cover 16:9, Instagram grid 1:1, Instagram story 9:16, menu inline 4:3, Apple Maps 4:3.
Why it mattersAn image at the wrong aspect ratio is either letterboxed or auto-cropped by the platform — neither is what the photographer composed for. Composing the source frame at 3:2 with awareness of how each destination will crop turns one shoot into eight usable images per dish, instead of one usable image and seven near-misses.
For a given destination surface (web hero, Yelp featured, Instagram profile-grid lead, Google Business cover), the single photograph that carries the visual identity of the restaurant — the one a returning visitor recognises before reading any text. Distinguishes from supporting shots (menu items, room interior, process) that round out the surface but don't anchor it.
Why it mattersThe hero is the photo a first-time visitor decides to come in by. Wansink (2005) and Cornell SHA (2014) put the descriptive-photo lift at 10–18% of selection rate; the hero is where most of that lift compounds because it's the photo every visitor sees once. Picking the wrong dish to hero is the single most expensive mistake an independent makes about photography.
A small file your web server uses to prove it's really you and to encrypt the connection to a visitor's browser. HTTPS is the result; the SSL (or more properly, TLS) certificate is the proof underneath it.
Why it mattersModern hosts issue these for free and renew them automatically via Let's Encrypt. If someone is quoting you for one — or worse, your certificate has quietly expired and visitors are seeing a scary red warning — both are solved by moving to a better host.
A page on your site (usually at /privacy) that plainly states what information you collect from visitors, what you use it for, and how you keep it safe. Legally required in the EU (GDPR), California (CCPA), and a growing list of US states.
Why it mattersEven if you only collect emails through a contact form, a missing privacy policy is a legal exposure and a trust red flag. The good news: a one-page, honest policy satisfies most rules. It doesn't need to be written by a lawyer for a site that just has hours and a menu.
The small bar or modal at the bottom of a page asking visitors to accept tracking cookies. Required by law in the EU and parts of the US if you run analytics or advertising pixels; unnecessary if you use a privacy-friendly analytics tool (like Plausible) that doesn't set tracking cookies.
Why it mattersCookie banners get dismissed in milliseconds and hurt your first impression. The cleaner answer for most small businesses is to use analytics that don't require one in the first place — your site stays legal and stays un-nagged.
A page covering the rules of using your site — reservation cancellation policy, gift-card terms, any disputes. Less strictly required than a privacy policy, but strongly recommended for any business that takes payments or deposits online.
Why it mattersA clear terms page short-circuits most chargeback disputes and cancellation arguments. Your processor (Stripe, Square) will also ask for a link to it during onboarding.
The percentage of your recent Google reviews to which the business owner has posted a reply — visible right under each review on your Google Business Profile.
Why it mattersDiners read the most recent 2–3 reviews before deciding. Zero replies signals an absentee operator and shifts borderline picks to a competitor who answers every one. Replying to even half — "thanks for coming in" on the 5-stars, "I'm sorry, please email me" on the 1-stars — measurably moves first-time visits.
The bundle of HTTP headers a modern site sends to tell browsers it should always load over HTTPS and how to handle script / frame loading safely. Graded A through F by Mozilla Observatory.
Why it mattersIf a diner hits your site and sees Chrome's "Not secure" warning at the top, the party-of-12 holiday booking dies before the booker has scrolled to your menu. The fix is usually one-line additions in your CDN / hosting settings — on Cloudflare, Vercel, or Netlify it's a checkbox.
The copyright year shown in your site footer (e.g. "© 2024 Joe's Pizza"). Diners use it as a subconscious freshness signal — particularly when planning a special meal or visiting from out of town.
Why it mattersA "© 2019" footer reads to an anxious diner as "are they still open after COVID?" It costs nothing to update and removes a silent doubt signal. Two-minute fix: change the year, or better, set it to auto-render from the current year so it never goes stale again.
How Google, Apple Maps, and voice assistants understand what your business is and where it is. Without this, customers don't find you in the first place.
Your business's free listing on Google Maps and Google Search — the panel on the right of a search that shows your name, hours, photos, reviews, and a link to your menu or booking. Claim and edit it at google.com/business.
Why it mattersFor most small local businesses, the Google Business Profile drives more discovery than the website itself. A well-tuned profile — right categories, good photos, complete hours, current menu link — is usually the highest-ROI hour anyone will spend on marketing.
See the research Usability of local business websites
The line of text that shows up as the clickable blue headline of your page in Google search results, and as the name of the tab in a browser. Set in the HTML with <title>…</title>. Under 60 characters, Google doesn't truncate it.
Why it mattersThe title tag is the single biggest on-page SEO signal you control. For a restaurant: restaurant name + cuisine + city is the formula ("Rowhouse & Garden — Italian Restaurant in Silver Spring, MD").
See the research Usability of local business websites
The short paragraph of text Google shows underneath the blue headline in a search result. Set in the HTML with <meta name="description" content="…">. Roughly 150–160 characters; any more gets cut off.
Why it mattersMeta descriptions don't directly affect ranking, but they heavily affect whether someone clicks. A sharp, specific description — with hours, neighborhood, and one differentiator — outperforms a generic one by double-digit percentages.
See the research Usability of local business websites
Invisible tags on your page that spell out, in a format Google can read directly, exactly what kind of business you are — "I'm a Restaurant, here's my cuisine, my address, my hours, my menu URL". Usually written as a JSON-LD script in the page head.
Why it mattersSchema is what lets Google show your hours, price range, and rating in the search result itself (a "rich result"), instead of a plain blue link. Without it you're guessing whether Google understood your business; with it, you're telling it.
See the research Usability of local business websites
A short written description of every image, set via the alt="…" attribute in the HTML. Screen readers read it aloud; Google uses it to understand what's in the photo.
Why it mattersGood alt text opens your site to blind and low-vision visitors (legally required in many jurisdictions), helps Google Images send traffic your way, and gives your site a shot at appearing in voice search results. "IMG_2041.jpg" helps nobody.
See the research Usability of local business websites
A file at /sitemap.xml on your domain that lists every public page on your site. You submit it once to Google Search Console and Bing Webmaster Tools so they know where to look.
Why it mattersWithout a sitemap, Google finds your pages eventually but slowly. With one, new blog posts, new menu pages, and new specials get indexed in days instead of weeks. It's a one-time setup that pays off forever.
See the research Usability of local business websites
The free Google dashboard that shows you which search terms people used to find your site, which pages they landed on, and whether Google has any indexing problems with your pages. Found at search.google.com/search-console.
Why it mattersSearch Console is the only way to know what Google actually thinks of your site. Every small business owner running a website should be claimed in it and checking it once a month — that's where you notice drops in traffic before they become problems.
See the research Usability of local business websites
A tag in a page's HTML that tells Google "this is the real address for this page, ignore any duplicates". Matters when the same content lives at multiple URLs (with/without www, with/without tracking parameters, printable versions).
Why it mattersWithout canonicals, Google can't tell which version of a page to rank and ends up splitting your SEO signal across three copies. A one-line tag concentrates all your search credit on the version you actually want people to land on.
See the research Usability of local business websites
A tiny plain-text file at /robots.txt on your domain that tells search-engine crawlers which parts of your site they can and can't read. Usually a five-line file; sometimes the quiet reason a page isn't showing up in Google at all.
Why it mattersA broken or accidentally-restrictive robots.txt is one of the most common "why is my site invisible on Google" causes. Get yours wrong once and half your pages disappear from search overnight.
See the research Usability of local business websites
A set of meta tags that tell Facebook, Instagram, LinkedIn, iMessage, and Slack how to preview your link when someone pastes it — which image to show, which headline, which description.
Why it mattersWithout Open Graph tags, a shared link shows as a grey box with your domain name, which people ignore. With a good food photo and a sharp headline, that same link becomes a thumbnail ad every time anyone posts about you.
See the research Usability of local business websites
Enhanced Google search results that show more than a plain blue link — star ratings, hours, prices, FAQ dropdowns, image carousels. Unlocked by having the right schema markup on your page.
Why it mattersA result with a 4.8-star rating and a "$" price next to it gets clicked far more often than a text-only result directly above it. Rich results are one of the few SEO levers where a small technical fix produces a visible, front-page win.
See the research Usability of local business websites
Your business name, address, and phone number written identically across your website, Google Business Profile, Yelp, Apple Maps, Facebook, and every other directory. "Suite 200" vs. "Ste. 200" counts as a mismatch to Google. Increasingly, hours-of-operation matters here too — Google treats day-of-week conflicts the same way it treats address conflicts.
Why it mattersGoogle cross-references your NAP across every mention of your business online and downgrades listings with inconsistent info — it reads as "this business might not be real". One tidy-up, done once, recovers rank you didn't know you'd lost. Open Hours covers the hours half; the other fields are a separate fix-up.
See the research Usability of local business websites
The structured representation of when a restaurant is open across the week — encoded variously as Google Business Profile fields, Yelp's Business Information, Apple Business Connect entries, Schema.org OpeningHoursSpecification on your website, and a handwritten card on your front door. Each surface has a different format; the underlying decision is the same.
Why it mattersWrong hours in any one place is the #2 driver of bad reviews after stale menus, and the #1 cause of Google Business Profile listing-quality decay. Diners check Google before they leave home; a "closed" listing on a Saturday lunch sends them to a competitor. Open Hours is the 15-minute fix-up that keeps the six places aligned.
See the research Usability of local business websites
A separate entity in Google's data model from regular hours: per-date overrides for holidays, private events, weather closures, and modified-hours days (early-close on Christmas Eve, brunch-only on Mother's Day). In Schema.org this maps to additional OpeningHoursSpecification entries with validFrom/validThrough dates.
Why it mattersMost owners enter regular hours once and never touch the data model again. Google flags listings where regular hours conflict with widely-known holiday closures (your "Mon: 11–9" still showing on Christmas Day) as "may have inaccurate info" — a soft demotion in local search. Maintaining special hours is the half of hours-management that requires a calendar, not a one-time fill-in.
See the research Usability of local business websites
The subset of special hours that recur on a calendar — closed on Thanksgiving and Christmas Day, early-close on Christmas Eve, brunch-only on Mother's Day, extended hours for St. Patrick's Day. Distinct from one-off closures only in that they come back next year. Captured in Schema.org as specialOpeningHoursSpecification entries with validFrom/validThrough set to the holiday date.
Why it mattersThe customer searching "is [restaurant] open today?" on Thanksgiving morning needs an authoritative answer. If your Google Business Profile or your site's JSON-LD still shows the standard schedule, the customer either takes the trip and finds you closed (negative review territory) or skips to the chain that bothered to post their holiday hours. Open Hours ships a holiday picker pre-populated with the dates that matter and produces an .ics calendar file with day-before alarms.
See the research Usability of local business websites
A specific Schema.org structured-data type for representing when a place is open. Lives inside a Restaurant JSON-LD block as one or more entries with dayOfWeek, opens, and closes fields. Holiday overrides use the same type via the parallel specialOpeningHoursSpecification property. The format Google's Rich Results parser actually consumes — the difference between "hours vary" and a full weekly grid in your search snippet.
Why it mattersThe customer searching "is [restaurant] open right now?" doesn't open your website — they glance at the rich snippet and decide. Without an OpeningHoursSpecification block on every page, Google falls back to "hours vary" or to whatever your Google Business Profile says, and when those two sources disagree your listing inherits the lower confidence. Open Hours generates a complete block (with the cross-midnight defensive pattern handled correctly) for any week + closure schedule you enter.
See the research Usability of local business websites
A small JSON block embedded in your website's <head> that mirrors the human-facing content (hours, address, menu items, reviews, prices) in a format search engines can index without parsing your design. The convention is Schema.org's vocabulary, expressed as JSON-LD; what you might recognize as the “rich result” preview Google shows is downstream of this data.
Why it mattersThe machine-readable copy decides what Google shows in your listing, your knowledge panel, and the first-screen result for "[restaurant name] hours". Pretty design doesn't matter if the structured data is missing or wrong — Google can't see CSS. Open Hours generates the hours portion of your structured data automatically; sending the block to a website builder usually takes them five minutes to install.
See the research Usability of local business websites
The non-Google map services — Apple Maps Connect (for iPhone users) and Bing Places (for Edge, Cortana, and ChatGPT-style Copilot) — where you can also claim and edit your business listing for free.
Why it mattersRoughly half of US phones are iPhones. An unmaintained Apple Maps listing sends every Siri-asking, iPhone-using customer to the wrong place or to your competition. Claiming these twice-a-year listings is an hour of work for years of payoff.
See the research Usability of local business websites
Google's rule, in effect since 2019, that it only looks at the mobile version of your site when deciding how to rank it. Whatever your desktop site does well no longer matters if the mobile version is bad.
Why it mattersEvery audit, every fix, every copy tweak has to be tested on a phone first. Desktop is a nice-to-have in 2026; mobile is the thing Google actually scores.
See the research Usability of local business websites
When the weekly hours your website declares (via OpeningHoursSpecification schema markup) don't match the hours Google shows on your Business Profile.
Why it mattersWhen your site says one set of hours and Google says another, Google's the one a Friday-night party of six trusts — they search "open now", see your name with the wrong hours, and either bounce or show up to a closed door. Among the most invisible reservation killers in independent dining. A 10-minute fix in Google Business Profile or your site's hours block prevents it permanently.
The restaurant audit tool tailors scoring and copy to ten segments. Each subtype maps to schema.org types, platform hints, and a specific set of priority-check weights — a pizzeria cares about direct-order commission in ways a fine-dining tasting-menu room doesn't, and vice versa. These entries explain what each segment covers and when it applies.
Tasting-menu or prix-fixe restaurants where the reservation is the conversion, the wine list is marketing, and OpenTable/Resy/Tock integration defines the booking funnel. Schema: Restaurant.
Why it mattersThe audit weighs reservations at 2.0× for fine-dining and the private-dining page at 1.5× — a single corporate holiday buyout pays for a slow week. Common platform fingerprints: Resy, Tock, SevenRooms, OpenTable. Keyword signals: tasting menu, prix fixe, chef's counter, sommelier, degustation, Michelin, omakase.
See the research Usability of local business websites
Sit-down restaurants that live in both worlds — reservations for dinner rushes, online ordering for lunch takeout. Neighborhood spots, family-friendly rooms, full-service bistros. Schema: Restaurant, FoodEstablishment.
Why it mattersMissing either reservations or direct ordering sends revenue to OpenTable, DoorDash, or a competitor with both. The audit weighs conversions at 1.5× and expects both flows. Common platforms: OpenTable, Yelp Reservations, Toast, Square. Keyword signals: dining room, full bar, lunch and dinner, signature dishes, family-friendly, neighborhood spot.
See the research Usability of local business websites
Counter-service restaurants where online ordering IS the business model. Every order through DoorDash/Grubhub costs 20-30% commission; every order through your own Toast or ChowNow keeps the margin. Schema: FastFoodRestaurant.
Why it mattersThe audit weighs conversions at 2.0× and menu-format at 1.5× — fast-casual menus ARE the conversion page; prices, item names, and the one-tap Order button need to live together. Platform fingerprints: Toast, ChowNow, Square, BentoBox, Popmenu, Slice, Menufy, Olo, Lunchbox. Keyword signals: order online, pickup, curbside, grab and go, fast-casual, counter service.
See the research Usability of local business websites
Coffee-led spots where morning traffic is 80% of revenue and the two most-read pieces of content are hours and location. Schema: CafeOrCoffeeShop.
Why it mattersThe audit prioritizes hours visibility, a tappable phone, and a legible menu; reservations are irrelevant, heavy online-ordering flows are nice-to-have but secondary. Platform fingerprints: Square (dominant for café POS), Toast, ChowNow. Keyword signals: espresso, cappuccino, pour-over, coffee shop, single-origin, cold brew. Roasters and bakeries with strong coffee programs usually classify here too.
See the research Usability of local business websites
Bakeries, patisseries, and dessert shops where custom-cake intake and wholesale accounts are the margin engine beneath the walk-in counter. Schema: Bakery, IceCreamShop.
Why it mattersThe audit weighs a wholesale / custom-orders flow at 2.0× for bakeries and treats allergen markers as a trust-and-safety signal, not just a marketing one — a customer ordering a wedding cake for a nut-allergic child needs to trust the cross-contamination policy before they book. Platform fingerprints: Square, Toast. Keyword signals: pastries, croissants, sourdough, patisserie, custom cake, wedding cake.
See the research Usability of local business websites
Bars, pubs, taprooms, breweries, wine bars, cocktail rooms — drink-led spots where private-party events and happy-hour clarity drive conversion. Schema: BarOrPub, Brewery, Winery, Distillery.
Why it mattersThe audit adds an age-gate check (weight 2.0× for bar-pub, 0 for every other subtype) and treats the private-events page as revenue-critical. Cocktail and draft lists rotate constantly, so the HTML menu check runs heavier here too. Platform fingerprints: Tripleseat for events, OpenTable/Resy/SevenRooms for reservations. Keyword signals: cocktails, craft beer, on tap, happy hour, gastropub, taproom, speakeasy, brewery.
See the research Usability of local business websites
Pizza-first restaurants where ordering IS the business and every Slice/DoorDash/Grubhub pie costs 20-30% in commission. Schema: Restaurant, FastFoodRestaurant.
Why it mattersThe audit weighs conversions at 2.0× and adds a delivery-radius check at 1.5×. A direct-order flow (Toast, ChowNow, Square, or Slice Direct) vs. routing everything through aggregators is the difference between ~78% and ~55% gross margin on a pie. Platform fingerprints: Slice is the strongest tell, then Toast, ChowNow, DoorDash, Grubhub. Keyword signals: pizza, neapolitan, wood-fired, coal-fired, sicilian, detroit-style, calzone.
See the research Usability of local business websites
Mobile kitchens and pop-ups where the primary content question is "where are you today?" — and static storefront maps are worse than no map at all. Schema: Restaurant, FastFoodRestaurant.
Why it mattersThe audit weighs the food-truck schedule page at 2.0× for this subtype and downweights maps because a pinned home address is the wrong answer — a dynamic "today's location" block, an Instagram feed, or a schedule calendar is what customers actually need. Reservations are irrelevant; catering inquiries are the margin-rich lead flow. Keyword signals: food truck, today's location, this week's schedule, pop-up, catch us, mobile kitchen.
See the research Usability of local business websites
Virtual brands and cloud kitchens that operate exclusively through aggregators — no dine-in, no pickup window, customers never visit. Schema: Restaurant, FastFoodRestaurant.
Why it mattersThe audit weighs the "explicit delivery-only marker" check at 2.0× for this subtype; without a clear "virtual kitchen / no dine-in" banner, customers show up in person to an empty address and leave one-star reviews. Aggregator links (DoorDash, Uber Eats, Grubhub, Caviar, Deliveroo) are the primary conversion; static storefront maps aren't useful. Keyword signals: ghost kitchen, virtual kitchen, delivery only, cloud kitchen, no dine-in.
See the research Usability of local business websites
Businesses that exist primarily to serve off-premise events — corporate catering, wedding catering, private-chef work, drop-off catering, pop-up dinners. Schema: FoodEstablishment, Restaurant.
Why it mattersThe entire site IS the catering page; the audit weighs that check at 2.5× for this subtype. Event planners comparing vendors search "catering [neighborhood]" and land directly on catering pages — packages, per-head pricing, minimum lead time, and a structured quote-request form are the critical surfaces. Platform fingerprints: ezCater, CaterTrax, Tripleseat. Keyword signals: catering menu, private events, corporate catering, wedding catering, drop-off catering, request a quote, off-premise.
See the research Usability of local business websites
The vocabulary of a restaurant's finances. The same terms show up in every lease negotiation, supplier conversation, and "free audit" pitch. Naming them cleanly is how you stop getting talked past.
Your food cost plus your labor cost, expressed as a percentage of sales. Healthy independents land between 55–65%; above 70% is operationally unsustainable without a hard change.
Why it mattersPrime cost is the single best one-number pulse of a restaurant's operating health. Everything else (rent, insurance, SaaS, loan payments) is more or less fixed — prime is the biggest lever you can actually move week to week. The Prime Cost Check in Margin Math plots your two inputs against the healthy band.
The cost of ingredients, alcohol, and consumables (paper goods, to-go packaging) expressed as a percentage of sales. Target 28–32% for most formats; pizzerias and cafés often run lower, steakhouses and sushi higher.
Why it mattersEvery point over target is margin walking out the kitchen door — and it's typically the fastest lever to fix, through portion specs, waste logs, supplier audits, and menu engineering. Most operators know their number to the tenth of a percent; many don't know what their peers' number is. The Plate Cost Calculator builds the per-dish food cost from a recipe in five minutes — the missing input behind a defensible aggregate number.
Wages, payroll taxes, and the tipped-employee contribution expressed as a percentage of sales. Target 28–32% for most formats, with quick-service at the low end and full-service at the high end.
Why it mattersRising labor markets make this the margin battlefield of the 2020s. Labor has faster turning radius than food cost — you can adjust a schedule this week; you can't renegotiate a tomato contract. When prime cost runs hot, this is usually where owners look first.
One customer seated and served. The restaurant industry's base unit of sales volume. "We did 180 covers on Saturday" means 180 customers came in and ate.
Why it mattersCovers per day × average check = revenue. Most restaurant math starts here. When comparing shifts, locations, or seasons, covers are the cleaner metric than revenue because they're not distorted by menu-pricing changes. The Break-Even Covers calculator answers "how many covers/day does my fixed overhead demand?"
Also called "average ticket." Total sales divided by total covers. If 180 covers spent \$4,500 total, average check is \$25.
Why it mattersThe lever that moves when you raise prices, upsell desserts, shift toward dinner from lunch, or redesign the menu to lead with higher-margin items. A 6% raise on the average check compounds into meaningful revenue without changing cover count — see the Price-Raise Simulator.
A third-party marketplace that takes orders on your behalf (DoorDash, Uber Eats, Grubhub) and keeps a commission per transaction. Not the same as direct online ordering through your own site.
Why it mattersThe single biggest margin pressure most independent restaurants face since 2020. Whether an aggregator is profitable for you depends on cover incrementality (does it bring new customers?) and commission tier. The DoorDash math post walks through the numbers; the Delivery Break-Even calculator runs them on yours.
The percentage an aggregator or payment platform keeps per transaction. Aggregators: 15–30% depending on tier (marketing, delivery, priority placement). Payment processors on direct orders: typically 2–3%.
Why it mattersOn a \$25 order at 25% commission and 2% processing, you keep \$18.25 before food cost. On the same order direct at 3% processing, you keep \$24.25. The \$6 difference per order is where the margin-recovery conversation starts.
A customer ordering through your own website or POS (Toast, Square, Olo, ChowNow), bypassing aggregators. You pay only the card-processing fee — no platform commission.
Why it mattersThe primary margin-recovery lever available to most restaurants. Every shifted order keeps the commission that would have gone to a platform. Two patterns work: aggressive website CTAs with obviously-better unit economics (e.g., free loyalty on direct), and in-restaurant nudges (QR codes on menus, signage near the door) pointing to the direct path.
A 50-year-old restaurant analysis (Kasavana & Smith, 1990) that plots every menu item on a 2×2 of contribution margin (\$ per sale) against menu-mix share (popularity), splitting at the median to label each dish a Star, Plowhorse, Puzzle, or Dog with a different recommended action.
Why it mattersOperators usually look at average prime cost or average price; menu engineering is what turns those averages into per-item decisions. Drop the Dogs, re-cost the Plowhorses, re-photograph the Puzzles, protect the Stars. The Menu Engineering Matrix runs the math on your own data.
For a single menu item: sale price minus the variable cost of producing one unit (plated food cost). Reported in dollars (CM \$) and as a percent of price (CM %). Distinct from operating margin and from food-cost percentage.
Why it mattersMost operators optimise on food-cost percent, but the dollar figure is what actually pays rent. A 30%-food-cost \$10 appetiser earns \$7 per sale; a 40%-food-cost \$30 entrée earns \$18 — the “worse-cost” entrée nearly triples the contribution. Menu engineering plots CM \$, not CM %. The Plate Cost Calculator suggests three menu prices at 28% / 30% / 33% and reports the dollar contribution margin at each.
For each item: units sold divided by total units sold across the menu in the same period. The denominator that turns a raw sales count into a velocity figure comparable across menus of different sizes. The y-axis of the menu-engineering matrix.
Why it mattersA dish that sold 80 covers tells you nothing on its own — 80 out of 200 is a Star, 80 out of 5,000 is a Dog. Menu mix is the math that lets a chef compare velocity across services, seasons, and menu sizes without recalibrating the eye each time.
What it costs you in ingredients to send one plate of a dish to the pass, after trim, peel, bone, and shrink. Computed per ingredient as edible-portion cost × portion used, then summed and divided by the recipe's portion count for batch recipes. Excludes labour and overhead.
Why it mattersPlate cost is the missing input every other restaurant-numbers tool assumes the operator already has. Menu engineering needs it as food cost; prime cost needs it aggregated across the menu; menu pricing needs it as the floor. Most independents work from the chef's gut estimate or the accountant's aggregate percentage; computing it dish-by-dish typically reveals 3–5 points of food cost the operator was eating without knowing.
For a given ingredient: the percentage of as-purchased weight that survives trim, peel, bone, fat, and cooking shrinkage. Romaine yields about 75%; a whole halibut closer to 50%; a whole chicken 60%. Sourced from culinary-school standard yield tables; varies ±5% by handler skill and product specification.
Why it mattersThe most-skipped concept in independent restaurant cost accounting. An owner who divides invoice cost by purchase weight under-prices most produce by 20–30%, and the loss compounds across every plate the recipe runs. Yield is the multiplier that turns AP cost into the real cost of a usable ounce — without it, every downstream number lies a little, in the same direction.
The cost of one unit of an ingredient after yield correction: EP cost = AP cost ÷ yield. Always ≥ the as-purchased cost; for low-yield items (whole halibut, whole chicken) it can be nearly double. The number that should drive menu pricing — using AP cost in its place under-prices most produce-heavy menus by 20–30%.
Why it mattersEP is the language of every culinary-school cost-control text and every printed yield table. Once an operator can read AP and EP fluently, they can audit any recipe's plate cost in their head from the invoice. It's the conceptual handle that lets a restaurant owner have an honest conversation with their chef and their accountant in the same vocabulary. The Plate Cost Calculator walks the AP→EP math row by row using ~80 canonical yields — override any one if your supplier's product yields differently.
The words next to each item — the descriptive sentence, the dish name, the price as it appears on the page — viewed as a separate craft from the cooking itself. A "house salad" can be the same plate of greens whether the menu calls it that or "Romaine, blistered cherry tomato, and 24-month parmesan, finished with house anchovy."
Why it mattersWansink's 2005 restaurant study found descriptive food names lift selection rate by 27% on average; Cornell's School of Hotel Administration has replicated the effect across casual and fine-dining tiers. Menu copy is one of the cheapest and least-touched conversion levers in the restaurant — most owners write their menu once and never audit it. The Menu Copy Inspector teaches how to read your own copy critically.
A modifier in a menu description that names a perceptual quality of the dish — flavor (smoky, briny, tangy), texture (crispy, silky, chewy), temperature (chilled, blistering, molten), or preparation method (charred, brined, dry-aged). Distinct from generic positives ("nice", "great", "amazing"), which carry no sensory information and are flagged by the Menu Copy Inspector as drag.
Why it mattersWansink's research isolated sensory adjectives as the specific lever behind the +27% selection lift. The cognitive mechanism: a sensory word lets the diner pre-experience the dish, which translates to higher willingness to order. Generic positives don't trigger the same imagined-taste response. Density target on a typical 12-word description: at least one sensory word, ideally two from different categories.
A pricing convention in which the cents portion ends in a number designed to make the price feel lower than the next round figure — \$19.95 instead of \$20, \$11.99 instead of \$12. Long established in retail; selectively used in restaurants depending on tier. The Menu Copy Inspector reports per-item pricing format and what each convention signals.
Why it mattersCornell's Kimes & Wirtz studies on restaurant pricing found that dropping the dollar sign entirely (writing "19" instead of "\$19") lifts average per-cover spend ~8%, while charm pricing communicates value but undercuts perceived quality at the upper price tiers. The right convention isn't universal — it depends on the segment, the average check, and the brand voice. Menu copy and price presentation are inseparable.
The sales volume at which revenue equals total costs. Below break-even, every cover loses money. Above break-even, every cover is profit. Expressed variously as "break-even sales," "break-even covers," or "break-even days."
Why it mattersKnowing the number below which you're losing money is the difference between "it was a slow week" and "it was a losing week." The Break-Even Covers calculator works backward from your fixed costs to show the covers/day floor.
Sales minus costs. Gross margin = sales − food cost. Contribution margin = sales − variable costs (food + hourly labor + card fees). Operating margin = sales − (food + labor + fixed costs). The word alone is ambiguous — always ask which margin.
Why it mattersConsultants, lenders, and vendors use "margin" loosely. Knowing which one is on the table in a specific conversation prevents you from agreeing to a number that sounds right but refers to a different line. When Margin Math says "contribution per cover," it means gross margin minus variable costs, which is the right input for break-even math — not operating margin.
How much cover count changes when prices change. Rough working rule in independent dining: a 6% menu-wide raise typically loses 2–3% of covers. Elasticity varies by item (signatures: low; anchors: high), daypart, and neighborhood.
Why it mattersThe whole math of a price raise lives in this one coefficient. Two restaurants can raise 6% and get opposite results — one gains margin; one loses covers. The Price-Raise Simulator lets you dial elasticity with a slider so you can model "if my guess is wrong by a point, what happens?"
Costs that don't change with sales volume: rent, insurance, loan payments, base utilities, SaaS subscriptions (POS, reservations, payroll, marketing tools), salaried management, accountant retainer. Independent of whether you did 40 covers or 400 last night.
Why it mattersThe floor your covers have to clear every month before profit starts. The first place to look when a business that "feels fine" keeps running negative months — fixed costs creep slowly (one new SaaS per quarter, a rent escalation clause) and nobody notices until the total moves the break-even.
Costs that scale with sales volume: food ingredients, hourly labor (cooks, servers, bussers, dishwashers), credit-card processing, aggregator commissions, variable utilities (gas while cooking), takeout packaging.
Why it mattersEvery sale has a marginal cost. Knowing what it is tells you which orders are actually worth taking — especially helpful when deciding whether a specific daypart, special, or channel pays for itself at the margin.
Four browser-side calculators — Delivery Break-Even, Prime Cost Check, Break-Even Covers, Price-Raise Simulator — plus a printable Monthly Margin Report, shareable URL-fragment scenario links, and an .ics monthly calendar reminder. Every input stays in your browser; no server endpoint receives your numbers.
Why it mattersEvery term above in this section is something Margin Math computes. The tool is designed to be run monthly — print the report, drop it beside your P&L, compare to last month, then hand it to your bookkeeper. The privacy claim is inspectable: no fetch() on input, no storage, no cookies. Right-click View source on the tool page — the math module is unminified and under ~500 lines.
A small vocabulary for evaluating whether any tool — ours or anyone else's — is honest about what it does with your numbers. Applies to every "free audit," calculator, or intake form you'll ever meet.
Computation that happens inside your browser tab, not on a server. A client-side tool takes your inputs, runs the math in JavaScript, and shows the result — without ever transmitting the inputs over the network. When the tab closes, the inputs are gone.
Why it mattersThe single technical test that separates tools that can see your numbers from tools that can only show them back to you. Margin Math is client-side across all four calculators — and invites you to prove it with DevTools before typing anything sensitive.
The browser's way of sending data to a server over HTTP. Visible in DevTools' Network tab. Modern tools may use fetch(), XMLHttpRequest, or a form submit to send input data onward.
Why it mattersIf typing a number into a free tool's form triggers a fetch request, your input just left your device. The fastest safety check on any tool: open DevTools → Network, start typing, watch the request list. If it grows with every keystroke, the tool is logging what you type. See how we verified this for Margin Math.
The portion of a URL that comes after the # symbol (sometimes called the "hash"). Per HTTP specification, browsers never transmit fragments to servers — they're read and used only on the client side.
Why it mattersA clean pattern for client-side tools that need a "save this scenario" URL: encode the inputs into the fragment, hand the reader a bookmarkable link, and keep the server entirely out of the loop. The permalink works; the server never sees what's in it. When Margin Math's share URLs ship, this is the design.
The small set of words every owner with a logo should know before handing it to a designer, a developer, a signage shop, or a print vendor. Most of these are invisible until they're missing.
The full visible system that carries a business across every touchpoint — logo, colors, typography, voice, and the rules for using them. A logo is one ingredient; a brand identity is the recipe that tells a designer, a developer, a print vendor, and a signage shop how to use it.
Why it mattersThe difference between "nice logo" and "recognizable business." Two restaurants can start with equally strong marks; six months later, the one with a written identity looks the same on its menu, its Instagram grid, its neon sign, and its takeout bag — and the one without it doesn't. Brand Suite generates a palette + tokens to start; a finished identity system is designer work.
A fixed composition of a mark, a wordmark, and sometimes a tagline, designed as one asset — spacing, alignment, and proportions locked in. The lockup is the asset designers hand to vendors 90% of the time. The bare mark alone covers the other 10%: the favicon, the social profile pic, an embossed coaster.
Why it mattersWhen a printer or developer asks you to "send your logo," the correct thing to send is a lockup — not a screenshot, not the mark stretched on its own. Identities that ship without a lockup get reassembled inconsistently by every vendor who touches them.
The minimum empty area that must surround a logo, usually measured in multiples of the mark's x-height or in a defined unit like "one cap-height of the wordmark." Nothing else — text, other logos, photos, borders — is allowed inside that zone.
Why it mattersThe first rule in every competent style guide. Clearspace protects the mark from visual noise and is the single rule most often violated by a well-meaning printer cramming "one more thing" next to the logo on a flyer.
The curated set of colors that belong to a brand — typically a primary, a secondary, a small number of accents, and a couple of neutrals. Each color has a defined hex value (and often CMYK / Pantone equivalents for print), a role, and a documented accessibility pair for text.
Why it mattersAfter the logo itself, the palette is the single largest carrier of brand feel. A restaurant with a defined palette renders consistently on its menu, its website, its Instagram grid, and its takeout bag; without one, every vendor invents their own. Brand Suite extracts a palette from an uploaded logo in the browser.
A readability standard, defined by the Web Content Accessibility Guidelines (WCAG), that requires body text to contrast at least 4.5:1 against its background and large text (18pt, or 14pt bold) at least 3:1. The AAA threshold is 7:1. Light grey on cream looks refined in a design mockup and fails badly in sunlight.
Why it mattersWCAG AA is the legal accessibility baseline in the US and EU, and it's also simply how readable your menu is on a patio at noon. Brand colors that look lovely in the logo but fail contrast on a menu shut out customers with low vision and invite ADA-compliance lawsuits. Brand Suite's contrast grid scores every pair in your palette.
The small icon displayed in the browser tab and in bookmark lists. Modern sites ship it at multiple sizes — 16, 32, 192, 512 pixels — because each context (tab, bookmark, home-screen icon) samples a different size. Usually a simplified version of the mark, not the full lockup.
Why it mattersThe tiniest touchpoint your brand has, and the most-viewed one once a customer bookmarks your site or adds it to their home screen. A missing or default-gray favicon is one of the first signals visitors use to judge whether a site is a real business.
The combination of a display typeface (used for headings and signage) and a body typeface (used for long-form reading) that together carry a brand's tone. A serif display + a geometric sans body is a classic pairing. Wrong pairings feel off to visitors without their being able to say why.
Why it mattersAfter color, the second-biggest tone carrier in a brand identity. The difference between "this restaurant feels handmade and old-world" and "this restaurant feels like a chain from 2008" is often 60% typography.
A single-color version of a logo, usually pure black and pure white (and occasionally a single brand color), designed for contexts where full color isn't reproducible: embroidery, foil embossing, small-scale print, laser engraving, a newspaper ad.
Why it mattersEvery logo needs a mono version. Its absence is the number-one signal to a designer that an identity wasn't finished — and the number-one reason a restaurant ends up with a muddy, pixelated reproduction on a staff polo.
A named reference for a brand decision — a color, a font size, a corner radius, a spacing step — exported as a portable file (CSS variables, JSON, or a Figma library) so the same value lives in one place and every surface that uses it stays in sync.
Why it mattersHex codes scattered across a stylesheet are how a brand drifts. Tokens are how it doesn't. When the palette is named once (--brand-primary) and referenced everywhere, swapping rust for olive is a one-line change instead of a search-and-replace across forty files. Brand Suite exports a starter set of tokens from any logo.
A version of a brand color whose lightness has been shifted just far enough — preserving hue and feel — that text rendered in it on a named background reaches the WCAG AA contrast threshold (4.5:1 for normal text). Two pairs are usually generated for each color: one on a light surface, one on a dark surface.
Why it mattersBrand teal looks great in the logo and unreadable as 14px body text on cream. Designing a usable system means knowing both the brand color *and* its accessible neighbor — and using the right one in each context. Brand Suite derives them automatically and exports both variants as CSS tokens.
A set of inherited vision conditions — protanopia, deuteranopia, tritanopia, and milder partial forms — in which one of the three cone-cell types is missing or weakened. The eye no longer distinguishes certain hue pairs (most commonly red from green), even when they look obviously different to other readers.
Why it mattersRoughly 8% of men and 0.5% of women experience some form. A palette that grades AA on luminance can still merge a "go" green into a "stop" red for a deuteranope, sinking accessibility despite passing automated checks. Brand Suite includes a Brettel/Viénot simulation toggle on its contrast grid so the math gets a vision-model sanity check too.
The relationship between the colors in a palette, measured by how far apart their hues sit on the color wheel. Named families — analogous, complementary, split-complementary, triadic, tetradic, monochromatic — each describe a specific spacing pattern that tends to feel intentional to a viewer, rather than accidental.
Why it mattersMost palettes that feel "off" are not using bad individual colors; they are using colors with no readable relationship between them. A harmony principle gives you a defensible reason to keep one color and drop another. Brand Suite's Palette Workshop generates candidate palettes from one anchor by applying these principles in OKLab hue space.
Two colors whose hues sit roughly 180° apart on the color wheel — terracotta and teal, mustard and indigo, brick and forest. The maximum-contrast pairing in color harmony. Split-complementary softens this by using the two hues 30° on either side of the opposite, producing a less-shouted variant.
Why it mattersBold and energetic by default — the right tool when a brand needs to feel alive (a tap room, a casual neighborhood spot, a bright daytime concept). Use sparingly: complementary pairs at full saturation can fight each other on the same plate of UI. Brand Suite's Workshop offers complementary candidates when "bold & energetic" mood is selected.
A palette built from hues 15–30° apart on the color wheel — terracotta with rust and ochre, forest with sage and juniper. Analogous palettes share a temperature and feel coherent because the eye reads them as variations on one decision rather than competing decisions.
Why it mattersThe right starting point when a brand wants to feel calm, considered, or warm — most independent fine-dining and natural-wine concepts land here. The trade-off is that the palette has less internal contrast, so a strong neutral and a documented accessible pair matter more than usual. Brand Suite's Workshop offers analogous candidates when "calm & considered" or "warm & welcoming" mood is selected.
A palette built from a single hue rendered at multiple lightness levels — a deep navy paired with a mid-blue, a pale ice, and a near-white tint of the same hue family. The hue stays constant; the lightness and chroma do the work of distinguishing the chips from each other.
Why it mattersThe most restrained of the harmony families, and the natural fit for a brand whose logo is already one strong color. Reads as sophisticated and disciplined; rarely reads as cheap. Brand Suite's Workshop offers monochromatic candidates when the uploaded logo is single-color, or when "refined & restrained" mood is selected.
A color space designed so that mathematical distance between two colors matches the perceived difference a human reader sees. Plain RGB does not have this property — two colors close in RGB can look very different, and vice versa. OKLab fixes that by separating lightness (L) from two opponent chroma axes (a, b).
Why it mattersEvery reliable color decision — k-means clustering for palette extraction, hue rotation for harmony generation, similarity scoring for "these two chips feel identical" warnings — needs a perceptual space to be honest. Brand Suite runs all of its color math in OKLab and only converts back to RGB for display and export.
Where the Cost Index gets its numbers — the public agencies behind every reading — and the methods that turn raw wholesale prices into a figure you can defend.
The U.S. federal statistics agency for prices, wages, and jobs. It publishes the Consumer Price Index (CPI) — what households pay at retail — and the Producer Price Index (PPI) — what businesses charge each other, including wholesale food — free and on a fixed schedule.
Why it mattersIt’s a public number nobody at your distributor controls. When a vendor says costs are up, the PPI shows whether prices businesses charge each other actually rose. The Cost Index uses BLS PPI/CPI to corroborate which way an ingredient’s price is moving — a check on the trend, not the primary wholesale level (that’s USDA).
The U.S. Department of Agriculture’s reporting service for wholesale food prices, run by its Agricultural Marketing Service (AMS). USDA reporters collect and publish actual transaction prices from terminal and shipping-point markets — produce, poultry, eggs, dairy, grain, and livestock — free to anyone, on a fixed daily or weekly schedule. The Cost Index uses these reports as its primary wholesale level for produce and poultry.
Why it mattersIt’s the closest thing to a public price check on the wholesale food market — built from real deals, not estimates, and not controlled by anyone selling to you. When a distributor says lettuce or chicken is up, the day’s terminal-market report shows what the actual market did. The Cost Index reads USDA Market News first for produce and poultry; BLS and FRED come in behind it to confirm the direction.
A USDA program under which meat packers are legally required to report their cattle, hog, and sheep purchase prices — and their boxed-beef and boxed-pork cutout values — to the government. Because nearly every large packer must report, the data is close to a census of the market rather than a survey of a sample, which is exactly what makes it hard to spin. It is public, free, and published on a fixed schedule.
Why it mattersThe price comes from a legal filing, not a phone call — the packers buying the cattle and selling the beef are the ones required to report, so no one along the chain gets to shade the number to their advantage. That is why the Cost Index uses USDA LMR as its primary wholesale source for proteins (beef and pork): when a vendor says meat is up, this is the public read of what packers actually paid and charged.
A weekly U.S. federal survey that collects actual sales prices and volumes for a handful of benchmark commodity dairy products — butter, cheddar cheese, dry whey, and nonfat dry milk. Those prices underpin the federal milk pricing system, and they’re public, free, and published on a fixed weekly schedule. The Cost Index uses NDPSR as its wholesale source for dairy.
Why it mattersIt’s the closest thing to a public, transaction-based price for commodity dairy — not an opinion or a phoned-in quote, but what was actually bought and sold that week. When your cheese or butter cost moves, NDPSR shows whether the commodity behind it moved too, or whether the change is happening somewhere between the plant and your invoice.
A free public database run by the Federal Reserve Bank of St. Louis. It gathers hundreds of thousands of economic series from many original producers — the BLS, the Census Bureau, the USDA, and others — into one searchable interface and a single open API. FRED distributes the data; it does not measure prices itself.
Why it mattersIt’s the convenient front door to numbers other agencies produce — one place to find a series and one open API to pull it, instead of chasing each agency’s own download page. The Cost Index uses FRED to pull certain public series programmatically, which keeps the build reproducible. Just keep the original measurer in mind: a price series on FRED is the BLS’s or the USDA’s work, re-served — FRED is the shelf, not the producer.
The U.S. federal agency for energy statistics. Its most restaurant-relevant series is the weekly retail price of diesel and gasoline — public, free, and on a fixed schedule.
Why it mattersThe Cost Index reads EIA diesel as a freight-cost pressure signal: when diesel climbs, delivery surcharges and freight-in costs tend to follow. It's a leading indicator on delivered cost — not a wholesale food price.
The U.S. agency that manages the nation’s ocean fisheries and publishes commercial landings data — the volume and ex-vessel (dock) value of seafood landed at U.S. ports. It’s part of NOAA, also called the National Marine Fisheries Service. Public and free.
Why it mattersIt’s the public read on what U.S.-landed seafood fetches at the dock, so the Cost Index uses it for seafood reference where it can. Be honest about the limit: most restaurant seafood is imported, and domestic landings reporting lags and is thin — so seafood is one of the index’s lower-confidence, harder-to-cover categories. We name the gap rather than paper over it.
The federal agency that regulates food safety: recalls, labeling rules, facility inspections, additive approvals, and the prevention rules under FSMA. It decides whether food is safe to sell and labeled correctly — not what it costs. The Cost Index draws zero prices from it; it’s a safety regulator, not a price source.
Why it mattersAn operator should absolutely watch the FDA — a recall can empty your walk-in overnight. Just not for pricing: the FDA publishes whether food is safe and labeled right, never what an ingredient sells for. The wholesale level comes from USDA, corroborated by BLS. Safety and cost are two different questions.
The U.S. commodity futures market, run by CME Group, where traders buy and sell standardized contracts on the future price of cattle, hogs, milk and dairy, and grains. The Cost Index deliberately does not use it: it tracks physical wholesale prices — the cash market, what a distributor charges this week — not financial futures.
Why it mattersA futures price is a bet on where a commodity is headed months out — not what shows up on your invoice, and it can swing on speculation detached from the cash market. Futures are for hedgers and traders; to tell a restaurant what it will pay, the Cost Index reads the cash side (USDA wholesale reports), not the futures screen.
A data-honesty taxonomy that labels every reading as exactly one of three states. Measured is a published series shown as-is — the highest confidence. Derived is a labeled estimate built from public inputs when no direct series exists — honest that it's an estimate, never passed off as a measurement. Absent is no usable public data, so the product says so plainly instead of inventing a number.
Why it mattersAny data product can adopt it, and the whole point is that you can always tell which of the three any reading is — a measured fact, a labeled estimate, or an honest gap — before you bet a decision on it. The Cost Index tags every ingredient as exactly one of the three: measured shows a solid line and a source date, derived shows a dashed line and an estimate badge, absent shows a greyed card that names the structural reason rather than a fabricated price.
Two separate certainties about a price: how sure you are of the dollar figure (level confidence) and how sure you are of which way it is moving (trend confidence). The two don't have to match — you can know prices are rising without knowing the exact dollar. The Cost Index scores both and shows the lower one as the headline.
Why it mattersLumping the two together hides what you actually know. A well-measured price shouldn't be marked uncertain just because last week's direction was noisy, and a clean direction shouldn't imply a dollar you can't name. Splitting them lets you act on the half you trust — and the Cost Index headlines the lower of level and trend so the number never claims more certainty than its weaker half can hold.
The publish-or-withhold gate behind the Cost Index. A reading ships only when it clears the bar — either a measured cross-market range where several public markets agree, or a single authoritative source whose direction is corroborated and whose noise is low. A lone, uncorroborated source is held back and shown as absent, not shipped as if it were solid.
Why it mattersIt's the rule that keeps a price honest. Any data product can ship every number it touches; a disciplined one keeps a publish-or-withhold gate, so a thin, uncorroborated figure never reaches you wearing the confidence of a measured one. In the Cost Index it means an ingredient earns a public reading only when the data supports an honest one — otherwise it's named and explained as absent, never faked. Ship complete, or not at all.
A calibrated uncertainty range, not a single guessed number. A prediction (or conformal) band is a range the true value should land inside a stated share of the time — 80% coverage means the real number should fall inside roughly 8 times out of 10.
Why it mattersA single forecast number hides how sure it is; a band shows it. Width is the honesty: a tight band claims a lot, a wide one admits doubt. The Cost Index publishes its derived forecasts as a band targeting 80% coverage (a split/EnbPI conformal method, widened until it holds). Walked forward across 84 ingredients and 5,929 scored steps, the band actually held about 84% of the time — covering a touch more than promised, which is the honest direction to err.
Testing whether stated confidence matches the real hit-rate. A forecast that says it is “70% sure” is well-calibrated only if calls like it come true about 70% of the time. The Cost Index publishes its own: its weak, medium, and strong direction calls came true about 48%, 51%, and 58% of the time across 1,877 scored calls, against a coin-flip baseline near 50%.
Why it mattersA confidence label is only worth reading if it’s earned. Calibration is how you check: grade a long run of past calls and see whether each tier comes true as often as it claims. The Cost Index does this in the open — “strong” calls land near 58%, “weak” ones near 48%, against a coin-flip near 50% — so “strong” genuinely beats chance and “weak” honestly barely does. The labels are earned, not decorative.
A way to estimate a number you can't measure directly by anchoring it to one you can, through a relationship that has held steady over time — and labeling the result an estimate, not a measurement. If the relationship isn't stable, you don't bridge.
Why it mattersIt's how you get an honest number when no direct one exists — without inventing it. The Cost Index builds a derived reading this way: when an item has no public wholesale series, it bridges from a related public series whose historical ratio has held, draws it as a dashed estimate with a confidence band, and labels it — never dressed up as a measurement. If the ratio isn't stable, it doesn't bridge; the item goes absent with the reason.
How current a number is. Public data publishes on a fixed schedule, so every figure ages between releases — and presenting a stale value as if it were today's reading is its own kind of lie. Series rot is a feed that quietly stops updating, leaving a last value that looks fine but isn't current.
Why it mattersA number's age matters as much as its value: an accurate figure read months late can be just as wrong as a bad one. The Cost Index stores every value's vintage, weighs each reading against its source's own cadence, and watches for series rot — a feed that stalls. A rotting or overdue series is flagged or dropped to absent with the reason, never shown as if it were today's number.
A leading indicator is something that reliably moves before the thing you care about, giving early warning. The Cost Index runs a leading-indicator pressure overlay: signals that lead delivered cost — diesel is the clearest, leading freight — published as an inferred direction of pressure (up, down, or flat), never as a dollar price, and only after the signal's own live track record proves out.
Why it mattersIt tells you which way cost is leaning before it lands on an invoice, without pretending to know the number. A leading signal like EIA diesel can warn that freight pressure is building — so a surcharge isn't a surprise — while the overlay stays honest about its limit: a direction, with a track record attached, and no dollar figure. It is the one part of the index that looks forward, and it is bounded hard.
Published numbers are versioned. A source agency releases a preliminary figure and later revises it to a final one, so any honest index built on them must revise with them — logging what changed and when, never silently overwriting history. Vintage is which version of a number you are looking at: the release it came from and the date it carried.
Why it mattersIf a number can change after you read it, you need to know which version you acted on. Treating every figure as a value plus a vintage is just good data practice — it keeps a past decision judgeable against what was actually known. The Cost Index follows a stated revision-and-correction policy: when an upstream series is revised, it re-vendors the new value with a fresh vintage and records the change rather than pretending the first number was always right.
A market price set by a private price-reporting agency's editors, who interview buyers and sellers and publish one judged figure that contracts then settle against. It is useful and widely used — but proprietary, usually paywalled, and resting on human judgment a reader cannot fully audit.
Why it mattersIt's the kind of number a vendor points to when they say "the market" moved — fast, granular, and contract-standard, but one you rent and can't fully audit. The Cost Index is deliberately not an assessed benchmark: it's built only from public, free, transaction-based government data (USDA and BLS), so anyone can re-pull the inputs. Knowing the difference tells you which kind of number you're looking at, and what you can check.