Mechanical keyboards: a cut pays one and a half times, a small rise shows no clear cost
I measured 1,705 price changes across 216 mechanical keyboards and compared each one against the products that held their price. A 10 percent cut improved the sales rank by about 15 percent within two weeks, while a 10 percent increase cost about 8 percent. Rises below 15 percent showed no clear effect at all. This is report 5 of twelve in this quarter's consumer electronics series.
In short
- A cut moves the rank more than one for one. A 10 percent cut improved the rank by about 15 percent (range 13 to 17) within two weeks.
- Raising the price costs less than cutting gains. A 10 percent increase cost about 8 percent of rank (range 6 to 10). The two ranges do not overlap.
- A small rise showed no clear cost. Increases of 5 to 15 percent left the rank where it was, within the range I can measure. Rises of more than 25 percent cost 40 percent.
- Deeper cuts bought more. A cut of 5 to 15 percent improved the rank by 20 percent, 15 to 25 percent by 32 percent, more than 25 percent by 44 percent.
- No price group stands out. All three groups gained from a cut, and their ranges overlap.
Price cuts
A 10% cut improved rank by roughly 13 to 17 percent.
Price increases
A 10% increase cost roughly 6 to 10 percent of rank.
Small rises go unnoticed, every cut counts
A cut of 5 to 15 percent improved the rank by 20 percent, and 15 to 25 percent by 32 percent. These two steps are clearly apart. Beyond 25 percent the rank improved by 44 percent, which overlaps with the step before.
On the other side, a rise of 5 to 15 percent showed no clear change. Its range runs from about 6 percent better to 6 percent worse. A rise of 15 to 25 percent cost 17 percent of rank, and more than 25 percent cost 40 percent. Each of these steps is clearly apart from the one before.
By price group, budget keyboards gained about 16 percent per 10 percent cut, mid ones about 11 and premium ones about 17. The ranges overlap, so the dip in the middle is an observation only.
The full picture
01 · Does discount depth matter?
Yes. Small rises showed no clear cost, larger ones did, and deeper cuts bought more.
The chart at the top of this report shows all six steps. The pattern holds on both sides, with one exception. A rise below 15 percent is the only move I cannot tell apart from no change.
02 · Will competitors follow?
A little more than they move on any other day.
A cut adds about one in ten extra moves
I looked at 1,864 cuts on 245 products. Within 14 days, some competitor changed a price after 97 percent of them. Against a normal two-week window, competitors moved about 11 percent more often after a cut.
After an increase it was also about 11 percent more. 62 percent of first moves fell on the same day as the trigger, which I cannot read as a reaction.
03 · Where can a new product get in?
Nowhere I can show on solid evidence.
A stable shelf with few newcomers to judge
No corridor: no price band in this category met the entry criteria, so none is marked recommended.
In the last 24 months, 60 new products entered this shelf. 8 of them now rank in the top third, and 2 have left again.
The strongest band in my scoring is €119 to €141, where two of three judged newcomers reached the top third. Three newcomers are too few to recommend a band. Only two newcomers left the shelf again.
04 · What does the shelf look like?
A shelf that changes slowly, spread fairly evenly from €30 to €140.
Of the 263 products I could place on the price map, 94 sit in the three cheapest bands, up to €50.48. 49 sit in the four most expensive bands, from €119.
The typical product in a band has between 56 and 204 reviews. A typical price held for 32 to 183 days, depending on the band.
Data appendix
I started from a list of 3,413 listings on amazon.de. I removed 1,545 listings that were not mechanical keyboards, such as keycaps, switches, wrist rests, keyboard and mouse sets, tablet keyboards and membrane keyboards. I also removed 19 multipacks and sets, 414 listings outside €30 to €250 and 506 listings with fewer than 20 reviews. To keep single brands from dominating, I kept at most three products per brand, which removed 614. Of the remaining 315 listings, 218 were variation groups that cannot be bought themselves. I replaced each with one sellable variant, the one with the longest price history, and checked the variants against the same rules. That gave 290 products. After the price history came in, I removed 12 whose typical price fell outside €30 to €250, which left 278. Of these, 12 never showed a price on amazon.de, so they add nothing to the figures.
I compare each product only against products in the same Amazon categories. Amazon files this product type in two parallel branches, "Gaming-Tastaturen" and "Tastaturen", and I accept both. That is stricter than title keywords, because a membrane keyboard often lists "mechanical feel" in its title. 15 products sat outside these categories, and their 61 price changes were left out. The product list is available on request for methodological review.
| Step | Events | Removed because |
|---|---|---|
| Price changes found | 3,786 | n/a |
| Gaps in the price history | −147 | the price history had holes around the change |
| Too little ranking data | −505 | the rank was missing on too many days |
| Overlapping price change | −776 | another price change fell inside the measuring window |
| Deal or coupon markers | 0 | a deal or coupon ran at the same time |
| Stockout signals | −229 | the product looked out of stock |
| Burst of new reviews | −325 | reviews jumped at the same time |
| Listing too young | −106 | the listing was less than 60 days old |
| Too few control products | −100 | fewer than 5 comparable products held their price |
| Passed every check | 2,130 | one event can fail several checks |
| Used for the figures | 1,705 | minus 376 with a trend already running before the change and 61 from products outside the category, 12 of them counted in both |
price change ≥ 5% and ≥ €0.30, held ≥ 7 days · comparison: 14 days before, days 3 to 14 after · control products: same category, price moved less than 2%, at least 5 of them · rank coverage ≥ 70% of days · listing older than 60 days · review burst = more than 10% and at least 15 new reviews · ranges are 90%, from a bootstrap clustered by product · study protocol v1.1
The price changes I found range from −63 percent to +222 percent. Moves that large are rare. They do not drive the figures, which come from the dense middle. 79 percent of the used changes were cuts or increases of up to 25 percent.
Limits
- LIMITS: I measure the ranking, not units sold, and only over two weeks after a change. Every figure holds only for the price moves sellers actually made in this category. The price groups are thirds of this set, so "premium" means €85 to €236.19 here. Competitor reactions are counted as any price change by another product in the category, which says nothing about intent. The entry view looks back 24 months and judges newcomers only once they have been listed long enough.
- 278 products
- 3,786 price changes found
- 2,130 passed every check
- 1,705 used for the figures
- 216 products in the fit set
- ref 0ce1130f
The Monthly
Each weekly report measures a single category. The Monthly is where they get compared. One subscription: The Monthly every month and the Quarterly every three months.