Retail personalization: implementation comparison
Recommendation systems governed two ways: review gates, or a halt on each targeting rule, plus reasons and an answered challenge where targeting sets a price or an offer.
Where this fits
A boundary case: recommendations and personalization
No single recommendation changes a person's status, access, money, or risk, so there is no decision to challenge. The framework applies where personalization sets a price, a credit offer, or who is shown a job.
A retailer's targeting decides who sees which price, which offer, and which job ad. A shopper who is shown worse terms because of their postcode or age never sees the better ones, so has nothing to challenge.
Standard governance relies on policy constraints and review gates with limited authority at runtime. It excludes sensitive attributes such as gender or age, runs cross-functional ethics reviews before deployment, and publishes notices about data use and tracking. A person signs off on large changes to the recommender, and a periodic audit checks whether some groups are shown worse prices or offers. None of it lets the shopper see what they were not shown.
Ethotechnics implementation
What changes, and what to verify before launch
Where targeting sets a price, an offer, or who sees a job, it is treated as a decision about a person.
- A shopper can turn personalization off, and it stays off. Turning it off does not cost them a price or an offer they would otherwise get.
- Where targeting sets a price, a credit offer, or who is shown a job, the person can see the main reasons it applied to them.
- A challenge to that price or offer goes to a named owner who answers by a stated date. That is what contestability means here.
- Upheld challenges are counted by targeting rule. The count goes to the rule's owner, who reviews the rule and records what changed.
- The owner can halt one targeting rule while it is reviewed, and a rollback target says how long it takes to undo a targeting change.
Before launch, publish the answer date for a challenge to a price or an offer. Measure time to restore for mistaken targeting changes, and log each rollback in the repair log.
Cite this implementation example · v1.0.0 · updated Feb 1, 2025 · CC BY-SA 4.0
APA
Ethotechnics Institute. (2025). Retail personalization: implementation comparison. Ethotechnics Institute. https://ethotechnics.org/use-cases/examples/retail-personalization
MLA
Ethotechnics Institute. "Retail personalization: implementation comparison." Ethotechnics Institute, 2025, https://ethotechnics.org/use-cases/examples/retail-personalization.
Chicago
Ethotechnics Institute. "Retail personalization: implementation comparison." Ethotechnics Institute. Feb 1, 2025. https://ethotechnics.org/use-cases/examples/retail-personalization.
BibTeX
@misc{ethotechnics_use_cases_examples_retail_personalization,
title={Retail personalization: implementation comparison},
author={Ethotechnics Institute},
year={2025},
howpublished={Ethotechnics Institute},
url={https://ethotechnics.org/use-cases/examples/retail-personalization},
version={v1.0.0}
}
RIS
TY - WEB TI - Retail personalization: implementation comparison AU - Ethotechnics Institute PY - 2025 UR - https://ethotechnics.org/use-cases/examples/retail-personalization ER -