You put the mains up by £1.50 on the first of the month. By the end of week two there are four reviews using the word expensive, one of which mentions the old price specifically, and a manager has sent you a screenshot with no caption.
The question everyone asks at this point is do we reverse it. The honest answer is that you cannot possibly know yet, and acting now is how price rises get rolled back for reasons that would have disappeared on their own.
Value complaints after a price change follow a shape. Knowing the shape is most of the skill.
The curve
Across the price changes we have watched, the pattern is consistent enough to plan around:
| Week | What happens | Why |
|---|---|---|
| 1 | Almost nothing | Reviews lag the visit by days; regulars have not been in yet |
| 2–3 | Sharp rise in price mentions, peak here | Regulars arrive, notice, and are the most likely to comment |
| 4–5 | Still elevated, wording shifts | Comparisons to the old price fade; general value language remains |
| 6–8 | Returns towards baseline | New guests have no old price to compare against |
| 9+ | New baseline visible | Whatever remains is the real effect |
The peak at weeks two to three is not a verdict. It is a transition artefact, and it is caused by something specific: the people most able to notice a price rise are your regulars, and they are also the people most likely to write about you. For a few weeks your review stream is disproportionately written by the one group holding a direct before-and-after comparison in their heads.
After about six weeks, most of the people eating with you never saw the old price. They are evaluating £18.50 on its merits rather than as a £1.50 insult, and the complaint has nothing to attach to.
The two numbers to write down before you change anything
This is the whole discipline, and it takes ten minutes on the day before the new menu goes up.
Baseline price-mention rate. Of your last ninety reviews, how many mention price, cost, value, expensive, or cheap? Express it as a percentage. Most restaurants sit somewhere between eight and twenty per cent, and where you sit is the only comparison that matters.
Baseline comparison rate. Of those, how many compare you to somewhere else by name or by price? This one is the more diagnostic of the two and nobody tracks it.
Write both down with the date. Without them, week three is unreadable — you will have a number of complaints and no idea whether it is three times normal or completely ordinary.
Reading week eight
At the eight-week mark, pull the same two numbers and compare.
Mention rate back within a few points of baseline. The rise landed. This is the common outcome for an increase of up to about eight per cent on an established menu, and it is the outcome you should expect rather than hope for.
Mention rate elevated, but comparison rate flat. Guests notice the price and are not shopping around on it. Usually fine. Keep watching the frequency of return rather than the reviews — the risk here is quiet attrition rather than complaint.
Comparison rate up sharply. This is the signal that matters. Guests naming a competitor and a number are telling you that you have crossed a local threshold on a specific item. Note which item. It is almost never the whole menu — it is the burger, or the glass of house red, or the kids' portion.
Mention rate still climbing at week eight. Something other than the price change is going on. Check whether portion sizes or a spec changed at the same time, which happens more often than anyone admits, because cost pressure tends to produce both at once.
The item-level trap
The most common real mistake in a price rise is not the size of it. It is applying it evenly.
Diners do not hold a price expectation for your menu. They hold one for maybe five items — the burger, the roast, the pint, the flat white, the children's main — because those exist everywhere and are therefore comparable. The rest of your menu is effectively unpriceable by a guest, which means it carries an increase almost invisibly.
Put 8% on everything and the anchor items absorb all of the complaint while contributing a fraction of the gross profit gain. Put 3% on the anchors and 11% on everything else, and you get a better result with a quieter review stream.
If your week-eight comparison rate is up, go and look at which items the comparisons name. It is a short list, and moving those four prices back while leaving the rest almost always resolves it.
What not to do in the panic window
Do not reverse it in week three. Beyond losing the margin, you teach your team that pricing decisions are provisional, and the next necessary rise becomes an argument.
Do not announce the rise in your review replies. Explaining input costs to a guest who said you were expensive reads as an argument, and it is read by everyone deciding whether to book. A plain, short reply that does not mention the increase is better.
Do not add a discount to cover it. A twenty per cent Tuesday to offset a price rise tells your guests the real price is the discounted one and leaves you worse off on both sides. Discounting into a value complaint reliably deepens it.
Do not change anything else for eight weeks. Not the portion, not the garnish, not the supplier, not the plate. You are running one experiment and this is the only period in which you will be able to read it cleanly.
Where the curve does not apply
Low review volume makes the curve unreadable. At eight reviews a month, week three is one or two reviews and the whole exercise is noise. Below about twenty reviews a month, use the sixty days before and the sixty days after as single blocks rather than tracking weekly.
Seasonality overlaps. A January rise lands in your quietest, most price-sensitive month, with a different guest mix. Compare against the same period last year as well as against the preceding ninety days, or you will attribute winter to your menu.
The complaint may not be about price at all. Plenty of "expensive" reviews are really about a room that signals a lower price band, or about a bill that contained something the guest did not expect. Those will not decay, because nothing about them was caused by the increase. Guests price you before they read the menu, and the charges discovered too late are their own category.
A rise done alongside a portion cut is not a price rise. It is a value cut, guests detect it accurately, and it does not decay. Portion complaints are usually about the plate rather than the weight, but a genuine reduction is a different matter and the reviews will say so for as long as it is true.
The practical version
Before the new menu: two numbers, dated. For eight weeks: nothing. At eight weeks: the same two numbers, and the list of items guests named.
That is a more disciplined approach to pricing than most restaurants manage, and the hard part is entirely the middle bit — holding still for eight weeks while a manager sends you screenshots.
OMMU tracks price-mention and comparison rates as running percentages with the dates of your menu changes marked on them, which makes the before-and-after a glance rather than an argument. Ninety reviews and a tally does the same, as long as you write the baseline down before you print the menu, which is the step that gets skipped.