Promotion reporting: did it actually work?

The Tuesday set menu ran for six weeks, and the manager is pleased: Tuesdays averaged seventy covers, up from forty, and fifty guests a night ordered the set menu. The owner asks whether it worked. The honest answer is that nobody knows yet. Some of those fifty guests would have come anyway and paid full price. Some came on Tuesday instead of Thursday. Only some were truly new. Promotion reporting exists to find out which was which.

Most promotion reports count how many guests used the offer. That number is easy to get and almost meaningless on its own. A promotion used by fifty guests a night might have created thirty extra covers or none at all, depending on how many of the fifty would have come anyway. The only question that matters is how many covers the offer created, and how many it merely discounted.

This guide covers how to separate the covers a promotion created from the ones it discounted or moved, how to set a baseline, how to do the sum after the event and how to put the result on one page so the next decision is easier. Every venue and figure in it is an example.

A promotion costs money: the margin given away on every cover that uses it, plus any extra costs. It earns money only from the covers it adds, the guests who would not have come without it. The question is whether the second is bigger than the first.

Everything in a promotion report should serve that question. How many guests used the offer, how busy the night was and how much the guests enjoyed it all matter, but only as evidence for the main question: did the offer create more value than it gave away?

Incremental covers are the covers a promotion created: guests who came because of the offer and would not otherwise have come at all, that night or any other. They are the promotion's real result, and the only covers that justify its cost.

They are hard to count directly, because nobody can ask every guest what they would have done without the offer. So they are estimated, by comparing what happened during the promotion with what would have happened without it. The comparison is the heart of promotion reporting.

Discounted covers are guests who used the offer but would have come anyway, at full price. A regular who comes every Tuesday and now orders the set menu is a discounted cover: the venue served the same guest on the same night and earned less.

Discounted covers are the promotion's hidden cost. They make the offer look popular while quietly reducing margin. A promotion with many discounted covers and few incremental ones can look like a great success on a busy night and still lose money.

Moved covers are guests who came during the promotion instead of another time. A guest who would have come on Thursday at full price comes on Tuesday for the set menu. The Tuesday looks busier, the Thursday is quieter and the week's total is unchanged, at a lower margin.

In marketing, this shifting of sales from one product or time to another is called cannibalisation. To spot it in a restaurant, look beyond the promotion night at the whole week, and at the weeks before and after.

The baseline is what would have happened without the promotion. It is the number every result is compared with. A good baseline for a weekly promotion might be the average covers on the same night in the four to six weeks before it started, or the same weeks last year if trade is seasonal.

Adjust for anything unusual. A heatwave, a local event, a nearby closure or the school holidays can change covers regardless of any offer. If the promotion weeks were unusual, say so in the report rather than crediting or blaming the offer for something it did not cause.

Count how many covers used the offer. For a set menu, count the set menus sold. For an offer booked in advance, count the bookings that mention it. For anything redeemed at the bill, record each use at the till. Without this count, the rest of the sum cannot be done.

Make counting easy for the team. A set menu that appears on the till as its own item is counted automatically. An offer applied informally by servers, with nothing recorded, cannot be counted at all.

New guests who used the offer are more likely to be incremental: they had not been coming before. Regulars who used it are more likely to be discounted: they were coming anyway. Separating the two gives a better estimate than treating all offer guests the same.

Look at how many offer guests had visited before, and how often. If most were regulars, the promotion mainly discounted existing trade. If most were new or lapsed guests, it probably brought in people who would not otherwise have come.

Before the promotion, you worked out the offer margin: how much margin each offer cover would give away, and how many incremental covers were needed to make it pay. After the event, do the sum again with real numbers.

Compare them. If the promotion needed twenty incremental covers a night to break even and delivered thirty, it paid. If it delivered ten, it did not. Our guide to offer margin: the sum to do before you publish covers the sum to do before; this one is the same sum, after.

Take the Tuesday set menu. The baseline is forty covers. During the promotion, Tuesdays averaged seventy, with fifty guests ordering the set menu. So the night gained about thirty covers on the baseline. Looking at the fifty set-menu guests, around twenty were regulars who had been coming on Tuesdays before.

Now check the rest of the week. Thursdays fell by about five covers on their baseline, suggesting some guests moved. So the estimate of incremental covers is around twenty-five a night: thirty gained on Tuesday, less five moved from Thursday. The cost is the margin given away on fifty set menus; the gain is the margin earned on twenty-five extra covers. Whether that pays depends on the menu's margin, which is the final step.

Promotions change what guests spend, and the report should show it. A set menu might lower spend per cover, because guests who would have ordered three courses and wine now order the set menu and a glass. Or it might raise it, if guests who would have had one course now have three.

Compare spend per cover for offer guests with the baseline spend for that night. A promotion that brings extra covers but lowers spend for everyone can earn less than it seems. One that brings extra covers and holds spend steady is doing well.

Always look at the whole week. A promotion that fills Tuesday while emptying Wednesday and Thursday has moved guests, not added them. A promotion that fills Tuesday while the rest of the week stays steady has added covers.

Compare each night's covers during the promotion with its own baseline. Small falls on other nights are common and may be chance. A steady pattern of falls on the nights next to the promotion is a sign of moved covers.

Some promotions pay off later. New guests who first came for the set menu may return at full price in the following weeks. If they do, the promotion's real value is larger than the promotion weeks show.

Look at guests who first came during the promotion, and see how many returned in the next two or three months, and whether they paid full price. A promotion whose new guests became regulars was worth more than its first sums suggest. One whose guests never came back was worth less.

Promotion reporting works best on one page: the promotion and its aim, the dates, the baseline, covers during the promotion, offer covers, the estimate of incremental, discounted and moved covers, spend per cover, the margin given away and earned, new guests who returned, and one line on what you learned.

Fill it in for every promotion, using the same layout. After a year, the pages show which kinds of promotion work for your venue and which merely discount it, which is the most useful marketing knowledge a restaurant can have.

Campaign reporting and promotion reporting answer different questions. Campaign reporting asks whether an email or a post brought guests. Promotion reporting asks whether an offer created covers, whichever channel told guests about it. A good email can promote a bad offer, and a good offer can be let down by a weak email.

Report them separately, then together. Our guide to campaign reporting: from send to seated guest covers the email side; this guide covers the offer itself.

Small venues have small numbers, and small numbers are noisy. A promotion that seems to add eight covers a night may be adding nothing, with the difference down to chance. A few weeks of data are rarely enough to be sure.

Present results as ranges rather than single figures: "the promotion added somewhere between ten and twenty covers a night". Run promotions long enough to see a pattern, and be cautious about drawing firm conclusions from a single short run.

A useful way to test a promotion is a holdout: run it on some weeks and not others, or on one night but not a similar one, and compare. If Tuesdays with the set menu consistently beat Tuesdays without it, by more than the normal variation, the promotion is creating covers.

Holdouts need a little planning and some patience, but they give far clearer answers than comparing with a baseline alone. They are especially useful before committing to a promotion for a whole season.

Once several promotions have been reported in the same way, compare them by cost per incremental cover: the margin given away, plus any extra costs, divided by the incremental covers. A promotion that cost three pounds per incremental cover was a better use of money than one that cost fifteen.

The comparison also guides the next decision. If set menus consistently cost less per incremental cover than percentage discounts, run more set menus. If events bring the most new guests who return, run more events. Our list of fifteen restaurant promotion ideas that are not 2 for 1 offers alternatives to test.

Some promotions are worth running even when their short-term sums are modest, because they build something that lasts. A monthly supper club may break even at first, then gather a group of regulars who book every month. A neighbourhood night may slowly turn local residents into regulars.

Report these over months, not weeks. Count the regulars they create and the full-price visits those regulars make. A promotion that builds a loyal group of guests can be one of the best investments a venue makes, even if its first month looked ordinary.

The report should also tell you when to stop. Signs include a promotion whose covers are mostly regulars, a promotion that moves guests from other nights, a promotion whose guests never return at full price and a promotion that has become a habit guests now expect.

Stopping a promotion that does not work is not a failure; it is the point of reporting. Our guide to restaurant discounts: the habit they create covers how to withdraw an offer without losing the guests it brought.

Share the results with the team, in plain terms. "The Tuesday set menu added about twenty-five covers a night, mostly new guests, and about a third of them have come back since" tells the team their effort mattered. It also helps them talk about the promotion to guests.

Ask the team what they saw. Servers often know which guests came for the offer, which were regulars and which seemed likely to return. Their observations can fill gaps the numbers leave.

Keep a log of every promotion: the aim, the offer, the dates, the baseline, the results and what you learned. Over a few years, the log becomes a record of what works for your venue, and it stops the same unsuccessful promotion being tried again because nobody remembers it.

Keep it with the marketing plan, and look at it before planning the next promotion. Our guide to a restaurant marketing plan on one page covers how the log and the plan fit together.

A good promotion creates many incremental covers and few discounted or moved ones, holds spend per cover steady and brings new guests who return at full price. Its margin given away is clearly smaller than its margin earned, and it does not create a habit guests come to expect.

Few promotions are perfect, but measuring them this way shows which come close. Over time, a venue that reports honestly runs fewer promotions, better ones, and gets more from each.

A new venue, or a new night of service, has no past to compare with, so there is no baseline. In that case, collect one before the promotion starts: open the night normally for a few weeks, count the covers and then launch the promotion.

If that is not possible, run the promotion for a defined period and then pause it, comparing the weeks with and without. It is less precise than a long baseline, but it gives far more information than reporting the promotion weeks on their own.

The margin given away is only part of a promotion's cost. There may be extra staff for the busier nights, printed menus or signs, paid posts to announce it, special ingredients or a guest chef's fee. Each belongs in the sum.

Add them to the report alongside the margin given away. A promotion that looks profitable on margin alone can turn out to lose money once the extra costs are counted, and one that looked marginal may be worth it if its costs were low.

In a group, a promotion can be tested at one venue before it runs at all of them. Run it at one venue, report it with the same rules and compare with a similar venue in the group that did not run it. That gives a clearer answer than any baseline.

Report the group's promotions in the same format, so they can be compared. A set menu that adds covers at one venue but not another may suit one neighbourhood and not the other, which is useful to know before rolling it out everywhere.

An invented bistro runs a Tuesday set menu for six weeks and reports it on one page. The baseline is forty covers; Tuesdays averaged seventy. Of the fifty set menus sold each night, about twenty went to regulars who already came on Tuesdays, and Thursdays fell by about five covers. The estimate is around twenty-five incremental covers a night.

At the set menu's margin, the twenty-five extra covers earned more than the margin given away on the fifty, so the promotion paid, though by less than the busy Tuesdays suggested. Three months later, about a third of the new guests had returned at full price. The bistro keeps the set menu, limits it to Tuesdays and records the result in its promotions log.

servearound shows the covers side of the sum but not a discount side, because its till has no discount function: a promotion is usually a set menu built on the menu, or something added. The menu report shows units sold, revenue and share of sales for every dish over any dates, so the dishes in an offer can be counted, and reservation analytics show covers by day and hour, bookings by source and each headline figure against the period before, which gives the baseline.

Guest analytics show new and returning guests by day and the return rate by the week guests were first seen, which is the evidence for whether promotion guests came back. Campaign reporting ends at the click, so the join between an email, the offer and the seated guest is one you make with your own rules. See Analytics and Campaigns.

To find out whether an offer created more value than it gave away.

Covers from guests who came because of the offer and would not otherwise have come.

Guests who used the offer but would have come anyway, at full price.

What would have happened without the promotion, such as the average for the same night in previous weeks.

Look at the other nights of the week. Falls next to the promotion night suggest guests moved.

Long enough to see a pattern, usually several weeks, and longer for small venues.

Running the promotion on some weeks and not others, then comparing.

By cost per incremental cover: margin given away plus extra costs, divided by incremental covers.

Promotion reporting has one real question: how many covers did the offer create, and how many did it merely discount or move? Counting guests who used the offer is not enough. Set a baseline from previous weeks or last year, count the offer covers, separate regulars from new guests, look at the whole week for moved covers and check spend per cover.

Do the offer margin sum again with real numbers, look at whether new guests returned at full price and present results as ranges when numbers are small. Put every promotion on one page, compare them by cost per incremental cover, use holdouts when you can, stop the ones that only discount and keep a log so each promotion makes the next one better.