A website for a food business has one job.
Not to win anything. To open on a phone, in a parking lot, in under three seconds, and let somebody read the menu and order without thinking about it. That is checkable, so here is the check.
Nobody has ever chosen where to eat because a website was beautiful.
They chose it because they could read a price at a red light, tap the number without copying it out, and see that you were open. A site does not make anybody want your food. It stops you losing the people who already do — and that is a smaller job than most people selling websites will admit, which is exactly why it can be measured instead of admired.
So this is the list my own report runs when somebody types their name into the box on the front page. Ten checks, the threshold each one is measured against, and what it costs when it fails. You can run it on yourself in about ten seconds and never speak to me.
Ten things I measure, and the line I measure them against.
- 01
Whether your listing points anywhere at all.
I start at your Google listing rather than at your site, because that is where most people start. If there is no website on it, everybody who looks you up lands on somebody else's page instead of yours — an app, an aggregator, a directory that scraped your menu wrong in 2019 and never came back.
The line: A website on the listing, or nothing else on this list applies.
- 02
Whether it loads at all.
I ask for the page the way a phone would and I wait. If nothing has come back in eight seconds I stop waiting and treat the site as down, because that is what a customer does. A site that is up on your office wifi and dead on a phone with two bars is down.
The line: Eight seconds, then I call it down.
- 03
How long it takes when it does load.
Anything past three seconds counts against you. That number is mine, not a study — I picked it because it is roughly where somebody in a parking lot decides you are broken and goes back to the search results. It is almost always one enormous background photograph that nobody ever resized.
The line: Three seconds for the page to arrive.
- 04
Whether it loads securely.
If the site still runs on plain http, the browser puts a warning in front of it before anybody sees your name. People do not read that warning. They leave, and they do not tell you why.
The line: https, or it fails.
- 05
Whether it knows it is on a phone.
One line in the page tells a phone to lay itself out at phone width. Without it, the phone shrinks the desktop version to a postage stamp and everybody pinches and drags. Most of your visitors are on a phone, so that is exactly where it breaks.
The line: The mobile viewport tag is either there or it is not.
- 06
Whether anybody can tap to call you.
A number typed on the page as plain text is a number a customer has to copy by hand and paste into the dialer. It should be a link that opens the phone. The easiest thing a hungry person can do is call you, and a lot of sites make them work for it.
The line: At least one tap-to-call link somewhere on the site.
- 07
Whether your menu is a page or a picture.
A PDF, or a photograph of the printed menu taped onto the page. Neither one is a page. Google reads them badly if at all, so your own dishes do not come up when somebody searches for them by name, and on a phone the customer is dragging around a layout that was designed for paper.
The line: The menu is real text, not a PDF and not an image.
- 08
Where your order button actually goes.
I follow the outbound links and see whose logo you land on. If every path to ordering ends on a marketplace, a cut of every one of those tickets leaves the building before you have paid for the chicken — and the customer's number stays with the app rather than with you.
The line: Is there any way to order that belongs to you?
- 09
Whether the page tells Google what you are.
A short block of machine-readable markup states your name, address, phone, hours and category outright, so Google is not inferring them from your layout. Most food business sites do not have it. It is an afternoon of work and it never has to be done again.
The line: Business markup present, or Google is guessing.
- 10
What year the bottom of the page says.
If the copyright line in the footer is two years old, a first-time visitor reads it as a business that might have closed. It is the cheapest thing on this list to fix and it is the one nobody ever fixes, because nobody scrolls to the bottom of their own website.
The line: Older than last year and it counts against you.
Four things I do not check, and one I cannot.
I do not grade your design.
Nobody has ever bought a plate of food because the corners were the right radius. If the menu opens and the number dials, an ugly site beats a beautiful one that does neither.
I do not hand you a performance score out of a hundred.
Those numbers are produced on a simulated connection and they move five points between two runs. Three seconds on a real phone is a fact. Ninety-four out of a hundred is a mood.
I do not count keywords.
There is no density to optimise. Write the dish names, the prices and the neighbourhood the way you say them out loud, and you are done — that is genuinely all the on-page part is for a business your size.
And here is what my check cannot see.
It reads what your server hands a browser on the first request. It does not run your site's code afterwards. So if your site assembles itself in the browser — a lot of newer builders do — some of these ten will read worse than your site really is. When that happens I say so instead of letting a red mark stand, because a scary report I cannot defend is worth less than no report.
The version of this that is worth paying for.
Everything on the list above, fixed by building the thing correctly the first time rather than by patching it.
- The menu as a real pageEvery dish is a line of text a phone can read and a search engine can index, with the prices sitting where you can change them.
- The number, tappable, on every screenOne tap opens the dialer. It is the cheapest fix on the whole list.
- Photographs of your actual foodTaken on the day I come out. Not the stock plate of pasta that shipped with the template, which every customer has already seen on four other websites this month.
- Hours that match your listingThe same hours in both places, including the holidays, because the two disagreeing is how somebody ends up at a locked door writing a review about it.
- Ordering, if it is worth it for youOn your own site, at your counter prices, with the order landing where your orders already land. Whether that pays for itself depends on your volume and I will tell you straight.
- The markup underneathName, address, phone, hours and category, written where Google reads them rather than left to be inferred from a layout.
- Your own domain, in your nameBought on your card so it belongs to you. If you ever decide you are done with me, the site and the domain leave with you and nothing has to be negotiated.
What a website cannot do for you.
It cannot make anybody want your food.
That part is yours and it always was. Everything above is about not losing the people who already decided.
It cannot fix your Google listing.
They are two different things and most people find you through the listing first. That is its own job and its own page. the listing, and what it actually controls.
It cannot be finished once and left.
Prices move, the special changes, you close for a week in July. A site that was right in March and wrong ever since is the same problem as no site, just quieter about it.
I am in Marietta and I work across metro Atlanta. A build starts with me coming out, because I cannot photograph your room or read the handwriting on your specials board over email, and the visit is where half the decisions above get made.
The rest of it
- Every systemEight systems and the website under them: what each one does, what it needs from you, and its price.
- PricingThe packages for trucks and restaurants, the discount for building more than one at once, and the terms.
- Online orderingWhy marking a plate up on the app raises the fee with it, and what your own order page has to do to be worth having.
- AI phoneWhat an AI phone line does on a real call, what it gets wrong, and how to catch it getting things wrong in week one.
- Google listingHours, categories, the attributes people filter on, photos and reviews — and the one ranking factor nobody can sell you.
- Food trucksWhat it costs to run a text list, why a location page has to admit when it is out of date, and when a truck is better off not buying one.
- Start hereThe homepage, and the free report — type your name, it pulls your Google listing and your site and tells you what it finds. About ten seconds, no email.
Text me your business name and I will run all ten of those on your site tonight.