Most of CDP work as SaaS and you don’t have access to raw data. Open CDP provides direct access to collected behavior and profile data. Your engineers get high quality datasets for machine learning in real-time.
Open CDP collects data from any source: website, mobile app, POS, NPS tablets, loyalty programs, payment gateways. It processes data and stores secured.
Matching versus deduplication
Instead of outdated process of deduplication (store duplicated data and join it with resources consuming operations), AI-matching operates duplicated data from all sources on-fly. Matching is performed by 2 identifiers: device ID and user ID. User ID can by any kind of identifiers: email, phone, loyalty ID, passport ID or driver license.
Track any users behavior events: product view, add to cart, purchase, fill form, sign up, add to favorites, payment, purchase return, etc. All events are attached to users profile and stored securely. Your data science team will appreciate.
Open CDP creates profile for every visitor, including anonymous one. New behavior and properties enrich the profile. Matching ensures all users devices are connected and events are attached to the profile. Raw data is available for AI/ML specialists.
Build dynamic segments based on profiles data and behavior events. Export to marketing platforms or use in machine learning.
Best open source technologies for your team. ElasticSearch stores structured profiles data. ClickHouse with SQL interface provides fast access to raw data of any size. Parallel processing in PHP: easy reading code and ability to modify it.
Platform is ready for horizontal scaling out of the box. Use single node for ElasticSearch and ClickHouse or attach large clusters for data lake: it will work. Run more workers by simple adding new servers. Just don’t forget to control free space.
Open CDP is open source. You have access to source code and database structure. Learn, modify, improve. Make it better. Or don’t and just use as is. You decide.