Constructor is committed to protecting the security and privacy of its customers and their end-users. Our entire data collection framework was built with security and privacy in mind.
Constructor does not collect or store any personally identifiable information (PII) about a site or app’s end users. All our learning algorithms are based on aggregated and anonymous user behavior.
Data Collection Overview
Constructor collects behavioral information from users on your site or mobile app, such as the searches they make, the items they click on, and their conversion behavior.
For each site or app user, Constructor creates a unique, random identifier which we use to track the behavior.
Constructor customers have the option of choosing an integration in which they pass their own user identifier to take advantage of data about past purchase history or loyalty programs. Constructor strongly encourages customers to use a one-way hashing algorithm or to otherwise obfuscate this identifier before sending it to Constructor so it cannot be used to tie back to a customer’s user database.
Behavioral data is sent via HTTPS using 2048-bit RSA encryption to the geographically closest Constructor data center.
Data Collection Details
The following is a list of data collected by the Constructor platform:
- Interaction events related to product discovery (search, autosuggest, browse, and recommendations), including:
- Search box focus
- Letters typed into the search box
- Search terms submitted
- Search filters applied
- Autosuggest suggestions clicked
- Product listing pages viewed
- Items clicked on product listing pages
- Product detail pages viewed
- Add-to-cart (or other interest signals specific to the site) clicks
- Conversions/purchases (if instrumented)
- Browser type and version
- Number of bytes returned by server
- HTTP information: protocol type, version, and response code
- IP address with last octet removed
- City, country, and postal code based on the original IP address
- Operating system type and version
- Random, anonymized browser ID
- Referring URL
- Session number (incremented after 30 minutes of inactivity)
- Optionally, an obfuscated user identifier passed by the customer
The data collected from the platform is aggregated for use in a variety of ways to improve and enhance the Constructor services, including:
- Learning to automatically correct misspelled search terms
- Learning synonyms that users commonly use
- Learning which items that groups of users click on and purchase together to power recommendations
- Learning which items users click on and purchase after a particular search to improve results for that searchterm
- Learning the preferences of individual users to boost items that interest them and recommend items to them
Behavioral data is retained by default within our data processing system for one year. Constructor saves raw behavioral logs in its archive system indefinitely. This archival retention policy can be configured on a per-customer basis.
Privacy Regulation Compliance
Constructor is fully committed to complying with all privacy regulations. We have processes in place to compile and send user data if requested, as well as to delete user data.