31 creator codes checked: 48.4% applied a different discount than advertised
Pohex's software tested 31 codes at checkout between 2026-09-23 and 2026-10-29; 16 matched the advertised discount and the rest differed in value or in kind.
By Pohex Editors. Published October 29, 2026.
What the checkout tests found
The comparison covers 31 listed creator codes that carry both an advertised discount and a checkout measurement. Of these, 16 matched the advertised discount in kind and value, which is 51.6% of the cohort. The other 48.4% did not.
Among the codes that differed, 8 (25.8%) were measured lower than advertised and 1 (3.2%) was measured higher. Another 6 (19.4%) came out as a different kind of discount, for example advertised as free shipping but measured as a percentage. Each code is classified once, from its most recent measurement.
How large the gaps were
Among the percent-off codes, 9 showed a mismatch between the advertised and the measured value. The median gap across those 9 was 10 percentage points.
The direction was uneven: 8 codes came out lower and only 1 came out higher. With counts this small, that is a description of this cohort and not a rate that can be assumed elsewhere. In addition, 1 code was left out because its advertised value could not be read as a percentage, an amount or free shipping.
Differences by advertised kind
Percent-off codes made up 27 of the 31 and 55.6% of them matched. The 1 amount-off code matched, which is too small a sample to say anything about that kind.
The 3 free-shipping codes did not match at all, at 0%. This is built into the method: the checkout tests exclude shipping, so a free-shipping code can never match. That result reflects the test design and says nothing about how those stores behave.
Limits of the comparison
Only a small share of listed codes state an advertised value, so the cohort is 31 codes and the shares move a lot with each code. The measurement history runs from 2026-06-30 to 2026-10-29, and nothing earlier is on record.
Each measurement uses 1 cart, 1 variant, quantity 1, no shipping, as a new customer. A percentage code read on a single item can read differently on a larger cart if a store applies tiers or caps. The measured value is also the latest one, so a store that changed a code after it was advertised appears here as a difference. The data shows that the values differ, not why.
Method
Question: How often does the discount a checkout actually applies differ from the discount the code was advertised with, and by how much?
Cohort: Every listed creator code that carries both an advertised discount (the value stated in the code's listing, from data/coupons.json or the offers feed) and a checkout measurement (data/resolved-discounts.json).
Denominator: Codes with both values. Codes whose advertised value cannot be read as a percentage, an amount or free shipping are left out and counted.
Observation window: Measurements from the first to the newest checkout measurement in the cohort. In this edition: 2026-09-23 to 2026-10-29.
What is counted: Each code is classified as a match (same kind and the same value), measured lower or measured higher (same kind, different value) or a different kind (for example, advertised as free shipping but measured as a percentage).
Censoring: None: each code is classified once, from its most recent measurement.
Sampling: One cart, one variant, quantity 1, no shipping, as a new customer. A percentage code measured on a single item can read differently on a larger cart if the store applies tiers or caps.
Counts: 31 codes compared; 1 more had an advertised value that is not a percentage, an amount or free shipping and were left out.
Limits of the data
- The measurement history runs from 2026-06-30 to 2026-10-29; nothing before 2026-06-30 is on record.
- 46 codes were measured once and never tested again.
- 376 events across 210 codes carry the date their data file was committed rather than the date of the test (day precision at best).
- The comparison rests on 31 codes, because few listings state an advertised value.
- Only a small share of listed codes carry an advertised value at all, so the cohort is small and the shares move a lot with each code.
- An amount-off code is compared in dollars and a percentage code in percentage points; a free-shipping code can never match, because the checkout tests exclude shipping.
- The measured value is the latest measurement; a store that changed the code after it was advertised shows up as a difference.
Data: download the CSV (31 rows). The same history feeds the verification feed.