> ## Documentation Index > Fetch the complete documentation index at: https://www.laurence.com/docs/llms.txt > Use this file to discover all available pages before exploring further. # Demand and Markets > Where we see shopper demand, what each source misses, and how we define markets, niches and product lines No single source shows all the demand on Amazon. Every source we use is good at telling which search is bigger and poor at telling how big it is, so we read levels as ranges and never treat a search that is missing from a source as a search nobody makes. ## What we mean by demand Demand for a search is two numbers: how many times shoppers make that search in a month (its **search volume**), and how many purchases start from it. A **search** (or search term) is the exact text a shopper types, such as "gps dog collar". We see demand through three sources: Amazon's Search Query Performance report, Amazon's Brand Analytics search terms, and the search-term report of our own ads. ## Search Query Performance (SQP) SQP is a report Amazon gives a brand about its own products, as part of Brand Analytics. For each of the brand's products (ASINs) and each search, over a week, a month or a quarter, it reports the search's total volume, impressions, clicks, add-to-carts and purchases across all sellers, and the product's own share of each. It counts organic and Sponsored Products traffic from the search results page, and leaves out widgets such as "Top Rated" or "New Arrivals". SQP is the only source with Amazon's own counts, but it is cut short in three ways: * **It stops at 100 searches per product and period.** No product in our data has more than 100 searches in a week, month or quarter, and 45% of monthly lists are full. Amazon documents a cap of 1,000 for the brand-wide view; the per-product view holds 100. * **The 100 are the product's own best searches, not the biggest searches.** The list follows how much the product itself was seen and clicked on each search far more than how often shoppers make the search. * **Each period has its own 100.** A product's week, month and quarter lists differ, so their union holds a median 2.5 times the month's list for a product at the cap. So SQP misses most of the searches our own ads get clicks on, but mostly small ones: 81% of distinct searches, 29% of clicks and 15% of impressions are missing from the product's monthly list. Of searches with a single ad click, 90% are missing; of searches with more than 20 ad clicks, 9% are. > [!WARNING] > > A search missing from SQP is not a search with no demand. Read it as "below this product's top 100". SQP and the ads report also count different things. SQP counts organic and paid traffic together, so its clicks for a product usually exceed the ads report's; yet on 27% of searches SQP shows fewer clicks than the ads report does, and SQP counts an impression every time the product is displayed in search results, which is not how the ads report counts it. Never reconcile the two line by line. Across all sellers, SQP shows how often a search ends in a purchase: purchases divided by clicks, on searches with 100 or more clicks, is 6.5% at the 25th percentile, 14.5% at the median and 23% at the 75th, about the same for small and large searches. ## Brand Analytics search terms Brand Analytics also ranks every search on Amazon by how often it is searched, the **search frequency rank** (SFR): 1 is the most searched. SFR ranks searches, not products. A product climbing the results for a search does not change that search's SFR; an SFR moving toward 1 means the search itself is getting more popular. For each search, Brand Analytics also lists the three products shoppers clicked most and their share of the search's clicks and purchases. Those three products convert at about 0.75 of their click share: their share of purchases is three quarters of their share of clicks. ## Third-party estimates Third-party tools estimate search volume for any brand. They put searches in nearly the right order but often get their size wrong, so read their numbers as ranges. ## Our search-term report Amazon reports, for our own ads, every search that triggered them: impressions, clicks, spend and attributed orders per search, per day, for Sponsored Products, Sponsored Brands and Sponsored Display, up to 15 months back. It is exact for our own ad traffic and silent about every search we did not advertise on. Its orders keep arriving for days after a click (see [Measuring Results](https://www.laurence.com/docs/amazon/measuring-results.md)). ## Which source for which question | Question | Use | Why | | - | - | - | | Which of two searches is bigger | Any source | All of them rank searches well | | How many shoppers make a search, for a connected brand | SQP, ideally the union of week, month and quarter | Amazon's own count, but only the product's top 100 | | What our ads got on a search | Our search-term report | Exact, for our ads only | ## Markets, niches and product lines These three words have fixed meanings at Laurence. * **Market.** An Amazon subcategory: one node of Amazon's browse tree, the category a listing is filed under. * **Niche.** A cluster of searches that share one shopper intent, such as "gps dog collar" and "dog tracker collar". A market contains many niches, and a search belongs to a niche, not to a market. * **Product line.** The listings a shopper chooses between by form, which differ only by colour, scent or size. A product line sells in one market. ## Branded and generic searches A **branded search** names the brand, or a misspelling of it; a **generic search** does not. Branded searches return far more per ad dollar: on two accounts we checked, branded keywords returned about three times the sales per ad dollar of generic ones. How much of a branded sale would have happened without the ad is not measured; see [Organic and Paid](https://www.laurence.com/docs/amazon/organic-and-paid.md). ## What we have measured | Fact | Evidence | Source | | - | - | - | | SQP counts organic and Sponsored Products traffic from search results, excluding widgets | `Amazon-documented` | Amazon's SQP help pages | | SQP's brand view is capped at 1,000 searches per report | `Amazon-documented` | Amazon's SQP FAQ | | SQP holds at most 100 searches per product and period; 29% of week, 45% of month and 56% of quarter lists are full | `Measured` | All SQP data, Jan 2025 to Sep 2026, run Oct 13 2026 | | SQP's list follows the product's own engagement: rank correlation of Amazon's score with the product's clicks −0.56, its impressions −0.48, the search's total volume −0.29 | `Measured` | 217 full monthly lists, run Oct 13 2026 | | The week, month and quarter union holds a median 2.52 times the month's list (p25 2.06, p75 2.84) | `Measured` | 67 product-months at the cap with all periods present, run Oct 13 2026 | | SQP omits 80.8% of the distinct searches our ads got clicks on, 28.8% of clicks, 14.6% of impressions; 89.7% at one click, 8.8% at 21 or more | `Measured` | 234,926 product-search-months, Sponsored Products Jun to Aug 2026 | | SQP clicks are below the ads report's search clicks on 27.2% of searches with 10+ ad clicks | `Measured` | 7,149 product-search-months, Jun to Aug 2026 | | Conversion per search across all sellers is 6.5% / 14.5% / 23% at p25 / p50 / p75, flat across volume tiers | `Measured` | 4,442 US searches with 100+ clicks, SQP monthly | | SFR ranks searches, not products; 1 is the most searched | `Amazon-documented` | Brand Analytics | | The top three products' purchase share is 0.75 of their click share (0.72 to 0.80) | `Measured` | Brand Analytics | | Market = subcategory; niche = cluster of searches with one intent; product line = listings differing only by colour, scent or size | `Our convention` | Our definitions | This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.