Conversion optimization for restaurants, mostly off the internet

The conversion happens offline and nothing online can see it

Between a website visit and a meal there is a walk, a drive and a decision that leaves no digital trace.

Somebody looks at a menu, decides, and turns up two hours later. Nothing connects those events, which means standard conversion measurement is close to useless in this business.

The consequence is that restaurants either measure nothing or measure the wrong things, optimizing for whatever the site can count rather than for covers.

The measurable proxies are worth using even though they are imperfect. Direction requests, phone calls from the profile, reservation completions, online orders and menu page views all correlate with visits.

The offline half can be asked about. A question at the table or on a receipt about how somebody heard of you is crude and it is the only bridge available.

Accepting the limitation is more useful than pretending it away. A restaurant that knows its measurement is directional will make better decisions than one treating a website metric as a proxy for business.

The menu decides the visit, so treat it as the conversion page

Almost every visitor is there to answer one question, and what they find determines whether they come.

People look at menus to establish whether there is something they want to eat, whether it suits the occasion, and whether the prices match what they expected.

A menu that hides prices loses people. Somebody who cannot establish the price range assumes it is more than they want to spend and moves on, and the assumption is usually worse than the reality.

Dietary information does disproportionate work, because the person in a group with a restriction frequently chooses the restaurant. Marking what is available plainly converts the whole table.

Currency matters. A menu showing dishes you no longer serve produces a disappointed customer at the table, which is worse than the visit you would have lost.

Photographs help and only if they are your food. Stock imagery of dishes you do not serve is recognizable and it undermines everything else on the page.

Reservations get abandoned in the last two steps

Somebody who reached the booking screen had decided, and the drop-off there is almost always mechanical.

The common failures are simple: no availability shown for the time somebody wants with no alternative offered, a form asking for more than is needed, or a system that requires an account.

Offering the nearest available times when the requested one is full recovers a meaningful share. Showing nothing sends somebody to the next restaurant rather than to a different hour.

Party size limits deserve attention. A group larger than the online system accepts is usually a valuable booking, and it should route to a person rather than to a dead end.

Confirmation and reminder messages reduce no-shows, which in a restaurant is a table that could not be sold. That is a conversion issue with a direct revenue number.

Check what happens on a phone at a busy hour, on a mobile connection. Reservation flows are frequently tested on a desk and used in a car.

The review ask is an operational step, not a marketing campaign

Reviews drive discovery in this category, and they only accumulate if somebody asks consistently.

A restaurant with steady recent reviews looks busy and current. One with an excellent old rating looks like it used to be good. That difference is produced by whether anybody asks.

Asking works best while the customer is still there and pleased, and it has to be built into the service routine rather than left to whoever remembers.

The mechanism should be frictionless. A short link on a receipt or a card outperforms an instruction to search for you later.

Do not incentivize in ways that violate platform rules or attract regulatory attention, and do not filter who gets asked. Both are visible and both cost more than they gain.

Responding is part of the same loop, and it is read by prospective customers rather than by the reviewer. A calm, specific reply to a complaint does real work on the person deciding.

Repeat visits are the whole business

Acquiring a diner is expensive relative to the check, and the economics only work if they come back.

A single visit may barely cover what it cost to attract, which means the profitable customer is the one who returns without being paid for again.

Most restaurants do very little about this deliberately. There is no record of who came, no reason for them to hear from you, and no way to notice that a regular stopped appearing.

The mechanisms that work are ordinary. Something worth joining, a reason to hear from you occasionally that is not a discount, and recognition when somebody returns.

Whatever you collect brings obligations attached, since email and text marketing carry consent and unsubscribe requirements. Worth setting that up properly rather than assembling a list informally.

The measure worth watching is the share of covers from returning customers, which most restaurants cannot state. Even a rough sense of it changes where effort goes.

Local conditions shape what any of this can promise

The sequence above holds anywhere. What you can operate and advertise does not.

Health permitting, alcohol licensing conditions and patio rules are administered locally and change more often than a website does, and they affect hours, capacity and what can be offered.

Anything on your site that depends on them is worth confirming with the local authority rather than carried forward from last year.

We work through local specifics market by market, and we are glad to go through yours with you.