Email open rates and the numbers that matter more

A restaurant sends its monthly email and checks the report the next morning: a sixty per cent open rate, the best ever. The team is delighted. But the Thursday supper the email was about has no more bookings than usual, and nobody clicked the button. A year ago the same email had a thirty per cent open rate and filled the room. Email open rates, once the first number every venue checked, no longer mean what they used to.

Opens are counted in a way that has become unreliable. Some email apps now register an open whether or not anyone reads the email; others hide real opens. The result is a number that can rise while results fall, and that cannot fairly compare one email with another. Venues that judge their emails by opens are judging them on noise.

This guide covers how opens are counted and why they have stopped being trustworthy, what an open rate can still tell you, and the three measures that matter more for a restaurant: clicks on the booking link, the bookings that followed and the health of the list. Every venue and figure in it is an example.

An open is usually counted when an email loads a tiny, invisible image hosted by the sender. When the guest's email app downloads the image, the sender's system records an open. It is a form of email tracking, and for years it was the main way businesses judged their emails.

The method always had flaws. Guests who read emails with images turned off were never counted. But for a long time the number was good enough to compare one email with another, and to see roughly how many people read each send.

Several changes have broken the open rate. Since 2021, Apple's Mail app has offered a privacy feature that loads emails' images in the background, which can register an open whether or not the person ever reads the email. Many people use it, so open rates for lists with many Apple users jumped overnight.

Other apps and services block images by default, so real opens go uncounted. Some company email systems scan incoming messages and load their content for security. Put together, open rates can be inflated for some guests and understated for others, and the mix changes from list to list and month to month.

Open rates are not useless, just unreliable for fine judgements. A sudden collapse in opens across several sends can still point to a real problem, such as emails landing in spam. A slow drift over many months, within your own list, may still say something about engagement.

What open rates can no longer do is compare one subject line or send time with another, or tell you how many guests actually read an email. For those questions, you need measures based on what guests do, not on whether an image loaded.

For a restaurant, three measures tell you whether an email worked. First, clicks: how many guests clicked the booking link or the main button. Second, bookings: how many bookings and covers followed, especially for the night or event the email was about. Third, the health of the list: how many emails were delivered, how many bounced and how many guests unsubscribed.

Each measures something an open cannot. Clicks show interest strong enough to act on. Bookings show results. List health shows whether the email strengthened or weakened your relationship with the guests who received it.

A click is a real action. A guest who taps the button to book, or to see the menu, has shown interest that no automatic image download can fake as easily. The click rate, the share of delivered emails that led to a click, is the best quick measure of whether an email engaged guests.

Look at unique clickers as well as total clicks. One guest who clicks five times is one interested guest, not five. And look at which links were clicked. An email whose clicks went mostly to the booking link did its job; one whose clicks went to the social media icons in the footer did not.

Clicks are more reliable than opens, but not perfect. Some company email filters follow every link in a message to check it for safety before it reaches the guest, which can register a click that no person made. These clicks often happen within seconds of delivery, on every link at once.

Counting unique guests rather than total clicks reduces the effect, as does looking at the links that matter, such as the booking button. And bookings, the second measure, are immune to it: a security filter does not book a table.

The real test of a restaurant email is whether guests booked. Look at bookings made in the day or two after a send, and at covers on the night or event the email was about, compared with a normal week. If an email about the Thursday supper was followed by more Thursday bookings than usual, it worked.

This is booking conversion, from email to table, and it is the measure closest to what the venue cares about. It is never perfectly clean, because weather, events and other marketing also affect bookings. But compared over several sends, it shows which emails bring guests. Our guide to booking conversion: from page view to confirmed covers the steps between a click and a booking.

Some booking systems and website tools can tell you where a booking came from, if the link the guest clicked carries a tag saying so. A booking link in an email that is tagged as coming from that email lets you count bookings from the send directly, rather than estimating them.

Where you can use tagged links, do. Where you cannot, compare bookings after the send with a normal period, and accept that the figure is an estimate. Our guide to booking links that tell you where guests came from covers how tagged links work.

The third measure is the health of the list after each send. How many emails were delivered? How many bounced, because the address no longer works? How many guests unsubscribed, and did any mark the email as spam? An email that brings bookings but causes a wave of unsubscribes may be costing more than it earns.

Watch these over time. A delivery rate that falls, or bounces that rise, can mean the list needs cleaning or that emails are starting to land in spam. Our guide to email deliverability: why your mail goes to spam covers what to do.

Campaign reporting does not need to be complicated. For each send, note five things: emails delivered, guests who clicked, the bookings and covers that followed, unsubscribes, and one line about what you learned. That fits on a single page for a whole quarter.

Keeping it simple makes it a habit. A report that takes ten minutes a month gets done; one that takes an afternoon does not. Over time, the page shows which kinds of email work for your guests, and which do not.

Compare like with like. An email to a small group of regulars will usually have a higher click rate than one to everyone on the list, because regulars are more engaged. Comparing them tells you about the audience, not the email.

Compare emails sent to similar audiences, of similar size, about similar things. And use rates rather than totals: a click rate per delivered email, not a raw number of clicks, so that a larger send does not look better simply because it was larger.

Email subject lines were traditionally judged by open rates. Now they are better judged by clicks, and by the bookings that followed. A subject line that brought more guests to the booking button did its job; one that only produced automatic opens did not.

Compare subject line styles across several sends, using click rates, and keep a log of what you tried. Our guide to email subject lines for a restaurant, tested covers which shapes tend to work and how to test your own.

Email send time was also traditionally judged by opens: send when people open the most. For a restaurant, the better question is when an email leads to the most bookings, which depends on when guests decide where to eat.

Test send times using bookings and clicks rather than opens. Our guide to email send time: the hour that fills Thursday covers a simple three-week test.

Not every guest who acts on an email clicks it. Some read about the Thursday supper and phone to book. Some mention it to a partner and book later through a search. Some simply walk in on Thursday. None of them appears as a click.

That is why bookings and covers matter alongside clicks. An email with a modest click rate may still fill a night, because guests act on it in their own way. Looking only at clicks would undervalue it.

Most restaurant lists are small, so the numbers from a single send are small too, and they jump around. A click rate of four per cent one week and six the next may mean nothing more than chance.

Look at trends over several sends, not at one. And make decisions on differences that repeat. If one kind of email consistently brings more clicks and bookings than another over three months, that is a real pattern. One good week is a hint at most.

It is tempting to put a revenue figure on each email, but for most restaurants it can only be an estimate. Take the extra covers you think an email brought, compared with a normal week, and multiply by your average spend. Be conservative, and say clearly that it is an estimate.

Overclaiming does harm. An email credited with every booking made that week will look far more valuable than it is, and may push the venue to send more emails than guests want. A modest, honest estimate is more useful than an impressive, inflated one.

Some numbers look important and tell you little. Open rates, for the reasons above. Total list size, which says nothing about whether guests read your emails. Total clicks, which can be inflated by a few enthusiastic guests or by security filters.

Stop leading with them. Report clicks per delivered email, bookings that followed and list health instead. A smaller list with high clicks and steady bookings is worth far more than a large one with impressive opens and empty Thursdays.

Once a month, look back at the month's sends. Which emails brought the most clicks per delivered email? Which were followed by more bookings than usual? Did any cause a rise in unsubscribes? What will you do differently next month?

Ten minutes is enough. The review turns a pile of numbers into a few decisions, such as more emails about specific dishes, fewer round-ups and a different send day, and keeps the emails improving month by month.

Published benchmarks for restaurant emails are mostly about open rates, which makes them less useful now. Even benchmarks for clicks vary widely, because they mix lists of very different sizes, ages and quality.

Compare your emails with your own past, not with an industry average. A click rate that has risen from three to five per cent over a year, with bookings rising alongside it, is a clear success, whatever the average elsewhere.

Sometimes the measures disagree. Clicks are high but bookings are low: guests are interested, but something stops them booking. Perhaps the night is already full, the booking page is slow, or the link takes guests to the home page rather than the booking form. Fix the path from click to booking.

Bookings are high but clicks are low: guests are acting on the email without clicking, perhaps by phoning or booking later. That is fine, but it suggests the booking button could be clearer. When the numbers disagree, the gap usually points to the problem.

When reporting email results to an owner or the team, use plain language and the measures that matter: "The Thursday supper email went to six hundred guests; forty clicked to book; Thursday covers were up by about fifteen on a normal week; three guests unsubscribed." Everyone understands that.

Leave out open rates, or mention them only with a warning. A report that leads with a sixty per cent open rate invites everyone to celebrate a number that may mean nothing.

Replies are one of the best signs that an email worked. A guest who replies to ask about the menu, to book a table or simply to say thank you is engaged in a way no open or click can show.

Make replies welcome: send from an address someone reads, and answer every reply. Count them informally. An email that brings a handful of warm replies has strengthened the venue's relationship with its guests, even if its click rate was ordinary.

Where an email has more than one link, the clicks on each show what guests wanted. If the menu link gets more clicks than the booking button, guests may want to see the food before they decide; next time, put a photo of the dish in the email and keep the booking button as the main link.

If most clicks go to the footer, such as the social media icons, the email's main message did not land. Look at which link each click went to, not just at the total.

Automated emails, such as a monthly note or an email to newly lapsed guests, run for months and are easy to forget. Measure them like any other send: clicks per delivered email, the bookings that followed and unsubscribes, compared over time.

An automation that worked well when it started may slowly lose its effect as its content ages. Review the numbers every few months, and refresh the email when clicks fall.

Beyond individual sends, look at the list as a whole every few months: how many confirmed subscribers it has, how many joined and left, and how many of them clicked anything in the last few months. A list where most guests have not clicked in a year is a list that has gone quiet.

Consider a gentle re-confirmation email to guests who have not engaged in a long time, and remove those who do not respond. A smaller list of engaged guests produces better numbers and better results than a large list of silent ones.

An invented brasserie has been judging its emails by open rates, which rose sharply last year while its email-driven bookings seemed to fall. It changes its reporting to three measures: click rate on the booking button, bookings and covers on the target night compared with a normal week, and unsubscribes.

Within two months, the picture changes. Its long monthly newsletter has a high open rate but few clicks and no rise in bookings. Its short emails about specific dishes have lower open rates, more clicks and noticeably fuller nights. The brasserie drops the newsletter, sends more short emails and reports results on one page each month.

servearound does not count opens at all, so its campaign reporting starts from what guests do. Campaign analytics show emails sent, the delivery rate, the guests who clicked, link clicks and the click rate, the bounce rate and the unsubscribe rate, with every campaign side by side, how each audience answered, automated sends against one-off ones and where the clicks went, by destination, with the clicks and the guests behind them.

Each campaign's report lists its recipients and the clicks on each link, and each venue's list health shows confirmed subscribers, net growth over thirty days and the unsubscribe rate across recent sends. servearound does not attribute bookings to a campaign, so the second measure in this guide comes from reservation analytics, which show covers by day and hour and bookings by source. See Analytics and Campaigns.

For spotting big problems, such as emails landing in spam. Not for comparing emails or counting readers.

Some email apps load images automatically, counting opens nobody made, while others block images and hide real ones.

Clicks on the booking link, the bookings that followed and the health of the list.

More than opens, but security filters can inflate them. Count unique guests and focus on the booking link.

Compare bookings and covers after the send with a normal week, or use tagged booking links where you can.

Compare with your own past instead. Most benchmarks are based on opens.

Look at the path from click to booking: availability, the booking page and where the link goes.

Monthly, in about ten minutes, on one page.

Email open rates no longer mean what they did. Some email apps count opens that nobody made, others hide real ones, and the result is a number that can rise while results fall. Use opens only to spot big problems, never to compare emails or count readers.

Judge restaurant emails by three measures instead: clicks on the booking link, counted by unique guests; the bookings and covers that followed, compared with a normal week; and the health of the list, from deliveries to unsubscribes. Compare like with like, look at trends over several sends, estimate revenue honestly and review it all on one page each month.