Like-for-Like Sales: How to Calculate Comparable Growth in a Retail Chain

Your chain sold more this year than last. The question every owner, lender and buyer asks next is how much of that came from the stores you already had, and how much from the ones you opened. Like-for-like sales growth, also called comparable or same-store sales, answers the first part: it compares sales from the same stores over the same periods, so a new store cannot flatter the result and a closed one cannot drag it down.

The formula is simple: like-for-like sales this period, divided by the same stores' sales in the comparison period, minus one. The hard part is deciding which stores are "the same", and the companies that report the number every quarter do not agree on that.

Total growth and like-for-like growth tell different stories

Take a chain that closed one store and opened two during the year. Its total sales grew by 7%, a figure that looks healthy. The stores it had all along grew by half that:

Both numbers are true. Total growth says how big the business got; like-for-like says whether the existing stores are doing better, which is what tells you if the format works. Listed retailers publish the reconciliation between the two for exactly this reason.

Which stores count as like-for-like

Every definition starts with a minimum trading period, so a store is compared only once it has a full comparison period behind it. The thresholds differ, even among Europe's largest grocers and general retailers:

• Ahold Delhaize counts a store once it has been open for a full 56 weeks¹.
• Marks & Spencer counts stores that have traded for at least 52 weeks with no change of more than 10% in floor space².
• Sainsbury's counts stores that have been open for more than one year³.
• Carrefour counts stores open for at least twelve months and excludes temporary closures⁴.

For a chain with a few dozen stores, a year of trading is the usual floor, and a few extra weeks help: a store that opened in the middle of a month has a partial first month, and comparing it with a full month a year later overstates its growth.

Refits, relocations and franchise moves

The harder cases are stores that changed. Here the definitions pull in opposite directions. Marks & Spencer takes a store out of the comparable base when its floor space changes by more than the threshold above². Ahold Delhaize keeps remodeled stores in, and keeps a store that was replaced by another in the same market area, but takes out any store that moved between its own network and franchisees¹. Sainsbury's treats an Argos store moved into one of its supermarkets as new space, while the supermarket itself stays like-for-like³.

Neither approach is wrong; what matters is choosing one and applying it the same way every time. One question settles each case:

VAT, fuel and currency

What counts as a sale changes the number too. Sainsbury's switched its like-for-like measure to exclude VAT in its latest annual report and restated the year before on the same basis; it also publishes the figure with and without fuel, 3.9% against 2.5%³. Carrefour reports like-for-like at constant exchange rates and without petrol⁴, and Kingfisher at constant currency, including revenue streams such as its marketplace once they have a year of history⁵.

For a chain trading in one country, the VAT question is the one that bites: when a rate changes, sales including VAT move without a single extra item sold. Compare sales excluding VAT. If you trade in more than one currency, translate both periods at the same rate, or the exchange rate shows up as growth. The guide to consolidating POS data across stores covers defining a sale once for the whole chain.

Easter, holidays and leap days

The calendar moves the number on its own. Carrefour's second-quarter sales growth in 2026 carried a calendar effect of -0.6%, which it strips out of like-for-like⁴. Kingfisher reports an underlying like-for-like of +1.4% that excludes calendar and leap-year impacts⁵. Costco said that Labor Day falling a week later cut its August total and comparable sales by a little less than 0.75 percentage points⁶. The UK's statistics office had to explain the same thing for Easter in 2026: Good Friday fell in its March period and Easter Monday in April, while a year earlier both fell in April⁷.

Three habits keep the calendar out of the result. Compare weeks with the same weeks a year earlier, shifted by 364 days, so a Saturday is compared with a Saturday. Mark the periods a moving holiday falls into, and say how much it moved the number when you report it. And compare a period that contains Easter with one that contains Easter, even if that means looking at March and April together.

Publish the definition, and the bridge

Like-for-like is not an accounting standard. In the EU it is an alternative performance measure, and when ESMA reviewed how listed companies reported such measures in 2025, most of the problems regulators found were in definitions, reconciliations and explanations⁸. A private chain is not policed the same way, but its lender, its board and a future buyer will ask the same questions.

So write the definition down once: minimum trading period, the refit threshold, VAT, currency and how you treat the calendar. Show the bridge from like-for-like to total growth whenever you report the number. When you change the definition, restate last year on the new basis, as Sainsbury's did, so the two years still compare. How to compare individual stores once the base is set is covered in the retail operations dashboard guide.

How marql shows like-for-like

marql's business review compares each store over the period with the period before it, and counts as like-for-like only the stores with sales in both. Stores that opened or closed in between are listed by name, the network total is marked as not like-for-like, and transfers between the chain's own locations are left out of sales.

When you compare two periods in analytics, including the same dates a year earlier, a like-for-like growth figure for the stores present in both periods sits next to the total.