Campaign reporting: from send to seated guest

The owner asks a simple question: did last month's email campaign work? The answer comes back in pieces. The email tool says it was delivered to two thousand guests and clicked by ninety. The booking system says there were more bookings than usual that week. Nobody can say how many of those bookings came from the email, or how many guests actually sat down and spent money. Campaign reporting that stops at the click cannot answer the question the owner asked.

The question is really about a chain: an email is sent, delivered, opened, clicked, and then, somewhere, a guest books, arrives and pays. Most of that chain is visible. The hard part is the join between the email and the booking, because guests rarely book in a straight line. They read on a phone, book later on a laptop, ring the restaurant or simply turn up. A useful report accepts that and sets clear rules for counting.

This guide covers the chain from send to seated guest, why it breaks, the attribution rules to agree before you start, how tracking and booking links help, how to match bookings to recipients and how to put the whole thing on one page. Every venue and figure in it is an example.

Campaign reporting has to answer three questions. Did the campaign bring guests who would not otherwise have come? Which guests were they? And what did it cost to bring each of them? Everything else, from delivery rates to clicks, is useful only in so far as it helps answer those three.

Most reports answer something easier instead: how many emails were delivered, how many were opened, how many clicks there were. These are signs of interest, not results. A campaign with a high click rate that brought no extra bookings did not work, however good its numbers looked.

The full chain has six links. The email is sent. It is delivered to the guest's inbox. The guest clicks a link. The guest books a table. The guest arrives and is seated. The guest pays, and spends something. A complete report follows a campaign along every link, from the first to the last.

The first three links are measured by the email tool. The last three are measured by the booking system and the till. Campaign reporting is the work of joining the two halves, so that a booking and a bill can be traced back, at least roughly, to the email that prompted them.

The chain breaks because guests do not behave in straight lines. A guest reads the email on the bus, mentions it to a partner that evening and books the next day on a laptop, by typing the restaurant's name into a search. Another guest phones. Another walks in on Thursday without booking at all. None of them clicked.

So a report that counts only bookings made through the email's link will undercount the campaign's effect. A report that counts every booking made that week will overcount it. Neither is right. The answer is to agree rules in advance, apply them consistently and accept that the result is an estimate.

Before running the report, agree how bookings will be credited to a campaign. In marketing, this is called attribution, and there is no perfect method. What matters is choosing a method that suits a restaurant and sticking to it, so that this month's figures can be compared with last month's.

A simple set of rules might be: a booking counts towards a campaign if it was made by a guest who received the email, within seven days of the send, for the night or event the email was about. Bookings made through the email's own link count directly; other bookings by recipients count as likely. Write the rules down, and use them every time.

The attribution window, how long after a send a booking can still be credited to it, should fit how guests book. For an email about a weeknight dinner, a few days is usually enough, because guests decide quickly. For an event announced three weeks ahead, the window should run until the event. For a seasonal menu, perhaps two weeks.

Too short a window misses guests who booked after a few days' thought. Too long a window credits the email with bookings it had nothing to do with. Choose a window for each kind of campaign, and keep it the same each time you send that kind.

It helps to report two figures. Direct bookings are those made through the email's own link, which you can be fairly confident the email caused. Indirect bookings are those made by recipients by another route, such as a phone call or a search, within the window. Some of those the email caused; some it did not.

Reporting both keeps the report honest. The direct figure is the floor: the campaign brought at least this many. The direct and indirect figures together are the ceiling. The truth is somewhere between, and over several campaigns you will get a feel for where.

Tracking links are web addresses with extra tags that tell a website or booking system where a visitor came from, such as which email, which campaign and which link. When a guest clicks a tagged link and books, the booking can carry the tag, and the report can count it directly.

They are useful, but they break easily. A guest who clicks the link, leaves the page and comes back later by typing the address loses the tag. A guest who forwards the email to a friend passes the tag on. Treat tagged bookings as a reliable minimum, not a complete count.

A booking link that takes guests straight to the booking page, rather than the home page, does two jobs. It makes booking easier, which raises the share of clicks that become bookings. And, where the booking system can record where a booking came from, it lets you see how many bookings arrived through that link.

Use one clear booking link in each email, pointing to the right page. Our guide to booking links that tell you where guests came from covers how to set them up.

Where links cannot carry the source, matching by email address can fill the gap. A guest's booking usually carries their email address, and so does the campaign's list of recipients. Comparing the two shows which recipients booked within the window, whatever route they used.

It can be done by hand for a small list, or with a spreadsheet: copy the recipients, list the bookings in the window and look for addresses that appear in both. It will miss guests who booked under a different address, but it catches most of the indirect bookings that tracking links miss.

A booking is not a result until the guest sits down. Some bookings are cancelled; some guests do not turn up. A campaign that brought twenty bookings, of which five were cancelled and three did not show, brought twelve seated tables, not twenty.

Report seated covers as well as bookings. It is the number that matters for the business, and it sometimes changes the picture: an offer that attracts many bookings but a high rate of no-shows may be bringing in guests who were never very committed.

The last link in the chain is what campaign guests spent. Multiply the seated covers by the average spend per cover for those tables, or, if you can, add up their actual bills. That gives a rough figure for the revenue the campaign brought in.

Be careful with it. Some of those guests would have come anyway, and some of their spend would have happened on another night. Treat the figure as gross revenue from campaign guests, not as extra revenue the campaign created, unless you have a good reason to believe the guests would not otherwise have come.

The most reliable way to measure what a campaign added is a control group: a small, random share of the list that does not receive the email. If guests who received it booked noticeably more than guests who did not, the difference is the campaign's real effect.

On a restaurant's list of a few thousand guests, a control group is small and the results are noisy, so treat them with caution. Where a control group is not practical, compare the campaign period with similar weeks before it, such as the same night in the previous few weeks. It is less precise, but far better than crediting the campaign with everything.

Email open rates are often put at the top of campaign reports, but they no longer belong in the chain. Some email apps register opens automatically, and others hide them, so open rates can rise while results fall. They tell you little about whether guests read the email, and nothing about whether they booked.

Leave them out of the main report, or show them only with a warning. Start the chain at delivery and clicks instead. Our guide to email open rates and the numbers that matter more covers why.

A campaign report should fit on one page: the campaign, the audience and the date; emails delivered; guests who clicked; direct bookings; indirect bookings within the window; seated covers; estimated spend; unsubscribes; the cost of the campaign; and one line on what you learned.

That is enough to answer the owner's question, and short enough to fill in every time. A longer report tends to be filled in once and abandoned. A one-page report, kept for every campaign, becomes a record of what works for your venue.

Campaign reporting matters most when deciding where the marketing budget goes. A campaign that cost a hundred pounds in staff time and tools and brought twenty seated covers cost five pounds per cover. An advert that cost four hundred pounds and brought twenty first-time guests cost twenty.

Comparing cost per seated cover across channels, honestly, shows where the next pound should go. It also shows the value of the list itself: email to guests who know the venue is often one of the cheapest sources of covers a restaurant has. Our guide to retention or acquisition: where to spend first covers how to compare the two sides.

To report cost per cover, you need the cost of the campaign. Include the email tool, if you pay per send or per month; the time spent writing, designing and checking the email; and the cost of any offer, such as a free glass of wine for every guest who books.

Staff time is often the largest cost and the most often forgotten. An email that took a manager three hours to write and check has a real cost, even if the tool was free. Counting it honestly shows which kinds of campaign are worth the effort.

Automated emails, such as a monthly note to regulars or a message to newly lapsed guests, run without anyone pressing send, so they are easy to leave out of reports. Report them separately, using the same rules: delivered, clicked, bookings within the window, seated covers.

Because they run continuously, look at them over months rather than individual sends. An automated email to lapsed guests might bring back a handful each month, which adds up to a meaningful number over a year, and is easy to miss if nobody looks.

Report results by audience as well as by campaign. The same email may bring many bookings from regulars and few from new guests, or the other way round. Knowing which audiences respond to which kinds of email helps you decide whom to send what.

Over time, a pattern appears: regulars respond to news about favourite dishes, lapsed guests to a new menu, new guests to an invitation to try something specific. Reporting by audience turns those hunches into evidence.

Sometimes a campaign shows no clear effect: a few clicks, a few bookings, nothing that stands out from a normal week. That does not always mean it failed. Small lists produce small numbers, and a small effect can be hidden by the weather or a local event.

Look at several campaigns of the same kind before concluding. If none of them shows an effect, the kind of campaign may not suit your guests. If some do and some do not, look at what differed: the content, the audience, the timing.

Some campaigns pay off long after the report is written. A lapsed regular brought back by an email may return every month for the next year. A new guest who came for an event may become a regular. The one-page report, with its seven-day window, sees only the first visit.

Look back occasionally at guests who first returned after a campaign, and see how often they have visited since. It is a rough exercise, but it shows the long-term value that short-term reports miss, especially for win back and second-visit campaigns.

When several campaigns and channels run at once, the same booking can be credited more than once: to this week's email, to last week's email, to a social media post and to an advert. Added together, the channels appear to have brought more bookings than the venue actually took.

Decide how to handle overlaps before they happen. A simple rule is to credit a booking to the most recent campaign the guest received within its window. Whatever the rule, apply it consistently, and check from time to time that the credited bookings do not add up to more than the bookings you actually had.

When reporting to an owner or a team, be clear about what is measured and what is estimated. "Ninety guests clicked; twenty-two booked through the email's link; another fifteen recipients booked within the week; we estimate the campaign brought between twenty-two and thirty-seven bookings" is honest and useful.

Avoid presenting estimates as facts. A report that claims a campaign "generated" a precise sum of revenue invites decisions based on false precision. A range, with its rules stated, invites better ones.

Review the attribution rules once a quarter. Is the window right for each kind of campaign? Are indirect bookings being counted sensibly? Is double counting under control? Adjust where needed, and note the change, so that figures before and after are not compared as if they were the same.

Keep the rules simple enough to explain in a sentence. Rules that only one person understands will not survive that person's holiday. Our guide to a restaurant marketing plan on one page covers how reporting fits into the wider plan.

An invented bistro sends a campaign about its new Thursday supper to its list of two thousand. The email tool shows ninety guests clicked. Using its rules, a seven-day window and matching by email address, the bistro finds twenty-two bookings made through the email's link and fifteen more by recipients who booked by phone or by searching.

Of those thirty-seven bookings, thirty-one were seated, with an average spend of about forty pounds a cover. Comparing Thursdays in the previous month, the bistro estimates that the campaign added around twenty to thirty covers. The one-page report records the range, the cost of the manager's time and a note that phone bookings were higher than expected.

servearound does not join a campaign to the bookings it produced: campaign reporting ends at the click, and bookings are not credited to a send. Campaign analytics show emails sent, the delivery rate, the guests who clicked, the click rate and where the clicks went, by destination, and each campaign's report lists its recipients and the clicks on each link, so the first half of the chain, from send to click, is there for every campaign.

The second half comes from the guest records: each booking adds to the guest's visits and last visit, so a recipient who booked shows it on their profile, and reservation analytics show covers by day and hour and bookings by source, from the booking page, the widget on your own site, phone, staff entered and walk-in. There is no campaign source among them and no tracked links for a venue's own channels, so the window and the matching in this guide are rules you apply yourself. See Analytics and Campaigns.

Whether the campaign brought guests who would not otherwise have come, who they were and what each cost.

Because many guests book another way, by phone, by searching or by walking in.

How long after a send a booking can still be credited to the campaign. Choose one per kind of campaign.

Both, but seated covers are the result that matters.

A small share of the list that does not receive the email, so you can compare. On small lists, treat the result with caution.

Not in the main chain. They have become unreliable.

Agree a rule, such as crediting the most recent campaign within its window, and check the totals.

Once a quarter, noting any change.

Campaign reporting has to follow the chain from send to seated guest: delivered, clicked, booked, seated, spent. The join between the email and the booking is where it breaks, because guests book by phone, search or walk in. Agree attribution rules before you start, with a window that fits each kind of campaign, and report direct bookings through the email's link alongside indirect bookings by recipients.

Count seated covers, not just bookings, estimate spend honestly, include the cost of staff time and compare cost per seated cover across channels when deciding the marketing budget. Keep the report to one page, leave open rates out of the chain, report automated sends and audiences separately and review the rules each quarter, always presenting estimates as ranges rather than facts.