In most teams, the question of how deep a discount has to go gets answered from memory. Someone remembers a competitor at 30 percent last November, someone else remembers that 15 percent was enough in spring. Both are true, and neither helps, because both are single observations.
This post answers the question by counting instead. The basis is 29,128 newsletters sent by 343 German online shops between 1 October 2025 and 31 August 2026, eleven full months. Of those emails, 12,991 name a percentage, which is 44.6 percent of them. Those 12,991 numbers are the subject here.
The result up front, and it is the opposite of a movement: the median fixed discount is 20 percent. Not on average across the year, but in every single one of the eleven months. What changes is not the depth. It is how many emails name a number at all, and how that number is phrased.
One note before the data: we see what was sent, not how it performed. This post says nothing about opens, clicks or revenue, because we do not have those numbers. And we do not name the shops. The point is the pattern, not the sender.
Eleven Months, Eleven Times 20
Two kinds of number need separating. A fixed discount is a commitment: 20 percent off the range. An up to is a ceiling: up to 50 percent, reachable on selected items. The two behave completely differently, which is why they sit side by side in this table.
| Month | Emails | Shops | With a number | Of those, "up to" | Median fixed | Median "up to" |
|---|---|---|---|---|---|---|
| October 2025 | 1,710 | 210 | 38.8 % | 27.6 % | 20 % | 50 % |
| November 2025 | 2,320 | 202 | 54.0 % | 39.9 % | 20 % | 50 % |
| December 2025 | 2,219 | 203 | 44.9 % | 39.5 % | 20 % | 50 % |
| January 2026 | 2,172 | 215 | 46.9 % | 39.1 % | 20 % | 50 % |
| February 2026 | 2,122 | 222 | 45.4 % | 27.2 % | 20 % | 50 % |
| March 2026 | 2,629 | 253 | 42.0 % | 29.8 % | 20 % | 50 % |
| April 2026 | 2,990 | 260 | 39.8 % | 33.9 % | 20 % | 50 % |
| May 2026 | 3,224 | 263 | 41.9 % | 30.1 % | 20 % | 50 % |
| June 2026 | 3,182 | 261 | 46.2 % | 39.7 % | 20 % | 50 % |
| July 2026 | 3,231 | 285 | 49.1 % | 44.3 % | 20 % | 50 % |
| August 2026 | 3,329 | 282 | 41.9 % | 40.5 % | 20 % | 52 % |
Eleven rows, and the two right-hand columns stand still. The median fixed discount is 20 percent in every month, the median ceiling is 50 percent in ten of eleven. In between sit a Christmas season, a January sale, a summer clearance and a Black Week.
The upper quartile of fixed discounts sits at 25 percent in eight of those months and reaches 30 only in December, January and August. So the season everyone expects to be a race to the bottom moves the upper quartile by five points, not twenty.
Where the Numbers Cluster
If the median is that calm, the distribution underneath is worth a look. It turns out to be even tighter.
| Fixed discount | Share of all fixed discounts |
|---|---|
| 20 % | 33.6 % |
| 10 % | 17.0 % |
| 15 % | 13.7 % |
| 25 % | 11.3 % |
| 30 % | 10.0 % |
| all other values | 14.4 % |
Five round numbers carry 74.2 percent of every fixed discount in the year. Slightly more than half, 54.3 percent, sits in the corridor between 15 and 25 percent. Only 14.2 percent goes beyond 30. A fixed discount of 35 or 45 percent is an exception in German retail, not an instrument.
Ceilings live somewhere else entirely, and just as firmly:
| "Up to" | Share of all ceilings |
|---|---|
| 50 % | 37.0 % |
| 70 % | 16.0 % |
| 60 % | 15.3 % |
| 30 % | 11.3 % |
| 40 % | 9.4 % |
| all other values | 11.0 % |
Two in three ceilings, 67.2 percent, start at 50 percent or higher, and half of all ceilings fall on 50, 60 and 70 alone. That is the real finding of this section: the market runs two separate number worlds. The number that commits to something lives between 15 and 25. The number that commits to nothing lives between 50 and 70. Comparing the two in one table means comparing a pricing decision with a headline.
Every Shop Has a House Number
So far, the calm median could be an averaging effect: many shops with very different numbers cancelling each other out. So we looked at shops individually. In eleven months, 188 of them sent at least ten emails carrying a fixed discount, enough for a median of their own.
- For 76 of those 188 shops, their own median is exactly 20 percent.
- 124 of 188, two thirds, sit between 15 and 25 percent.
- 29 sit below, 35 above.
The cancelling out never happens. Shops are not scattered around a shared average, they sit individually on the same number. It gets clearer when you ask how loyal a shop stays to its own figure. At the median, 46.5 percent of a shop's discount emails carry its single most used value, and 86 of the 188 shops put at least half their emails on one number. That house number is 20 for 78 shops, 10 for 34 and 15 for 21.
One example from the panel, standing in for the pattern: a beauty retailer sent 69 emails with a fixed discount across the eleven months. 66 of them said 20 percent. What changes between October and August is not the number but the occasion and the urgency in front of it: "Dein Rabatt: –20 %" (your discount: 20 percent off), "NUR HEUTE: –20 % auf vieles" (today only: 20 percent off many items), "Nur noch heute: 20 % Rabatt!" (last day: 20 percent off). The same commitment, eleven months long, in new packaging every time.
For competitor monitoring this means the discount depth of any single competitor is not a daily decision you need to chase weekly. It is a constant you learn once. It only gets interesting in the rare moment it changes.
What Moves Instead
There is movement, it just sits in other columns of the same table.
The share of emails naming a percentage at all ranges from 38.8 percent in October 2025 to 54.0 percent in November 2025. That is the Black Friday effect, and it is a volume effect: during Black Week the same shops do not discount deeper, more shops discount at all.
The share of ceilings among all numbers swings harder, from 27.2 percent in February to 44.3 percent in July. In the summer clearance and in Black Week, almost twice as many emails phrase their number as an "up to" compared with a quiet February. The number in the subject line grows in those weeks without the offer behind it getting better.
Together that gives a usable planning picture: peak season in German retail means more emails, more numbers and bigger numbers with an "up to" in front of them. Peak season does not mean lower prices.
Where the Corridor Ends
The 14.2 percent above 30 are not random noise, they sit in recognisable corners of the market. Ten of the 188 shops have a median of 30 percent, individual ones reach 40, 50 and 60. Two grocery retailers sit at 30.5 and 33 percent, a wine merchant at 37.5. Those are ranges with short shelf life, or with a margin structure that survives a clearance.
At the other end sit 17 shops with a median of exactly 10 percent, among them clothing, electronics, furniture and travel. Ten percent is not a weak offer there, it is what the margin allows.
The practical consequence: a benchmark across all industries, as tight as this one turns out to be, does not replace a comparison with your own competitive set. It tells you where the market stands. It does not tell you where your five most important competitors stand.
What to Take Away
- Depth is the dullest variable. Eleven months, a median of 20 percent, without a single exception. Building a plan around "how deep" optimises the number that moves least.
- Two number worlds, not one. Fixed discounts live between 15 and 25 percent, ceilings between 50 and 70. An "up to 50 %" is not twice as good as 25 percent, it is a different kind of statement.
- Round numbers win. 74.2 percent of all fixed discounts are 10, 15, 20, 25 or 30 percent. The 17 or 22 percent that sounds like fine tuning in a meeting barely exists in the market.
- Competitors have house numbers. 86 of 188 shops put at least half their discount emails on a single value. Know that value per competitor and every deviation becomes obvious.
- Season shifts volume, not price. In November 54.0 percent of emails name a number, against 38.8 percent in October. The median stays where it was.
How We Counted
The basis is the Reyo panel: newsletters from German online shops that we receive as regular subscribers and evaluate. The window runs from 1 October 2025 to 31 August 2026, eleven full months, and contains 29,128 emails from 343 shops.
An email counts once, no matter how many offers it carries. Where an email names several percentages, the value counted is the one the extraction identifies as the email's leading offer. Emails without a percentage, such as free shipping, a gift above a basket threshold or pure content, are part of the 29,128 but not of the 12,991.
We report medians, not averages. Single emails at 80 or 90 percent on clearance stock would pull any average upward and produce a picture that matches no shop at all. Monthly values are reported for full months only, which is why September 2026 is missing.
We do not quote personalised subject lines, nor any subject line that identifies the sender. No voucher code appears verbatim.
What You Need for This
A benchmark is a starting point. It becomes actionable once you know the house number of each individual competitor and notice when it changes.
- Competitor Monitoring reads your competitors' newsletters, on-site promotions and ads, and records discount depth, mechanic and duration per campaign.
- Market Predictions projects from that history when a competitor runs their next campaign and with which mechanic.
- Campaign Planner puts your own plan next to the market's expected campaigns, so you do not launch the same number on the same day.
For the same data at the level of individual emails, see Newsletter Examples: 12 Emails from German Retail. For what happens to these numbers during Black Week, see Black Friday Email Examples.
If you want to know where your competitors sit in this corridor, we will show you on your own competitive set: see the demo.