An automatically discovered domain becomes an internal candidate. It is not listed, ranked or described publicly until it passes deduplication and editorial review.
1. How discovery works
ECIndex may identify candidate stores from approved public directories, industry and association listings, search or merchant datasets whose terms permit the intended use, partner referrals, existing public links, and submissions from businesses or community members.
Each candidate retains its discovery source and time. We normalize the domain, compare it with published and pending records, and look for public evidence that the website supports commerce. The initial pilot focuses on Pakistan fashion, electronics, and health and beauty stores.
2. What the discovery crawler observes
The classifier requests a bounded copy of a public homepage and may record the response status, title, meta description and the presence of signals such as Product or Offer structured data, product links, prices, cart or checkout language, commerce technology markers, and public delivery, payment or returns information.
These signals produce an internal discovery-confidence level used to prioritize editorial review. Discovery confidence is not an ECIndex score and does not measure traffic, revenue, reputation, safety or customer satisfaction.
3. Technical and ethical boundaries
- We request public pages only and use bounded response sizes and timeouts.
- We do not sign in, create accounts, submit forms, add products to a cart or place orders.
- We do not attempt to evade access controls, bot protections, paywalls or rate limits.
- We do not seek customer lists, orders, private analytics or personal shopping information.
- We maintain per-domain limits and intend to respect applicable robots directives and source terms.
- Suspicious, harmful, illegal or clearly irrelevant candidates can be quarantined or rejected.
4. Deduplication and editorial review
Before review, candidate domains are compared with existing stores and other candidates. A reviewer can accept, reject, merge, defer or request more evidence. Acceptance moves a domain into the normal observation process; it does not guarantee a category, score or ranking position.
Commercial customers, affiliates, partners and service-provider relationships receive no discovery or ranking advantage. Review decisions and later material changes are intended to retain an audit trail.
5. Crawler identity
Our current automated requests identify themselves with user-agent names including EcommerceIndexDiscoveryBot/0.1 for candidate classification and EcommerceIndexBot/0.1 for observations of approved stores. Requests include a link to ECIndex information pages.
Crawler names and versions may change as the system develops. This page is the authoritative public description of their intended activity.
6. Controls for website operators
Website operators can use standard robots directives to communicate crawler preferences. Because discovery sources, cached evidence and third-party referrals can exist independently of a crawl, blocking automated access may prevent a new observation but does not necessarily remove an existing factual directory record.
Authorized owners, managers and employees can request a factual correction or refresh from the relevant store profile. Requests involving exclusion, legal rights or sensitive circumstances will be reviewed individually.
7. Questions, corrections and exclusion requests
Contact hello@ecindex.org and identify the domain, your relationship to it and the requested action. We may verify authority through a business-domain email or other reasonable evidence before changing a record.
Related disclosures are available in our Methodology, Privacy Policy and Terms of Use.