Amazon 70% Off Deals October 2026: What the Badge Actually Means

We read three unrelated Amazon aisles on 17 October 2026 — 70 LED bulb and batten listings, 87 school stationery listings and 33 hand cream and body lotion listings, logged out, from India, with nothing bought — and asked one question of all 190: what does the discount badge actually tell a shopper?

The short answer: it tells you how big the seller's claim is, not how much you are saving. The 69 tiles badged 70% or deeper had a median price of Rs. 190; the 42 badged 30% or shallower had a median of Rs. 299. Deep-badge tiles are the cheaper things on the shelf, and their stated list prices sit 4.23 times above what they charge against 1.17 times for the shallow ones.

Three shelves also behave differently enough that a habit built on one of them will mislead you on the next. Here is what each looked like.

The exact-fraction column is the one that surprised us. On the beauty shelf, 20 of 33 tiles priced at an exact fraction of their stated MRP — eleven at MRP times 0.3 (badged 70%) and nine at exactly half (badged 50%). On the bulb shelf and the stationery shelf, one tile each did that. Formula pricing is real, common, and completely shelf-dependent.

The stationery shelf carries the widest claim range, from 1.01x to 100x. That 100x is a single Rs. 170 geometry box stating Rs. 17,000, which reads as a misplaced decimal rather than a price history; it is excluded from every figure on this page and named here rather than dropped silently. An exact 100x multiple on a cheap item is the clearest data-error signal on any shelf.

Evidence coverage is near-total on all three. 187 of the 190 listings published a star average and a rating count, which is not true of every retailer — on two fashion shelves we read the same morning, one published none at all across 190 listings.

Read the third and fourth columns together. The deeply badged group is Rs. 109 cheaper at the median and its claimed list price is 3.6 times further from its selling price. Both facts have the same cause: these are low-priced goods from sellers who print large list prices. The badge is measuring the gap between two numbers the seller chose, and one of those numbers is unverifiable.

And here is the claim that did not survive. Within individual shelves it looked as though deep badges sat on thin evidence — the stationery shelf had an 81% badge with seven ratings behind it. Across all 190 listings that reverses: median rating count 627 for the deep group against 378 for the shallow one. We are reporting the version that failed rather than the neater one, because a rule that only works on the shelf it was invented on is not a rule. Use the rating count to judge a product, not to catch a badge.

One: hold the spec still before comparing anything. On the bulb shelf, seven plain single 9-watt bulbs sold from Rs. 42 to Rs. 142 — same wattage, same cap, 3.4 times apart. A price is only information once the wattage, the shade count or the millilitres are fixed. Most apparent bargains are category differences in disguise.

Two: divide the pack before you believe either number. A six-bulb pack in this read sold at Rs. 505, which is Rs. 84.17 a bulb against a stated Rs. 500 a bulb, and its 83% badge was computed on the pack. Multipacks are the one place where both printed figures need dividing before they mean anything.

Three: divide the stated MRP by the price. About 1.2x means the two printed numbers nearly agree. Above 4x the badge is being produced by the claim. An exact multiple — precisely 2x, precisely 10x, precisely 100x — is either a pricing formula or an error, and on this read it was both, on different shelves.

Four: read the rating count, and read what it is attached to. It is the only figure on a tile that does not come from the seller, and 187 of these 190 listings carried one. On other retailers' shelves that number is sometimes absent entirely, and on some it is shared across several near-identical listings — so check whether the same count appears on more than one tile before treating it as independent evidence. This page read prices on one morning; nobody here bought, opened or tested anything.