Welcome to BPP
Bytek Prediction Platform (BPP) is a customer prediction platform that helps brands and agencies turn first-party data into smarter audience targeting and more accurate conversion measurement.
BPP connects to your data warehouse, applies AI models, resolves customer identities across sources, and pushes enriched segments and conversion signals to your ad platforms.
What you can do with BPP
- Audiences — build precise customer segments based on who your users are, what they have done, and what AI predicts they will do next. Sync them to Google Ads, Meta Ads, and other platforms for remarketing, Customer Match, and lookalike seeding.
- Signals — define conversion events and business milestones from your data (e.g., closed deals, subscription renewals, in-store purchases). Send them to ad platforms for offline conversion attribution and value-based bidding.
- AI Models — activate predictive models (purchase propensity, predicted customer lifetime value, RFM segmentation, interest classification) that score your customers and surface insights you can act on.
- Connections — connect BPP to Google Ads, Meta Ads, Microsoft Ads, Anonymised, or any custom endpoint.
- Identity Resolution — BPP automatically links all your customer identifiers (email, cookie, CRM ID, etc.) to a single unified profile, so your audiences and signals are always based on real people, not fragmented records.
How BPP works
- Your first-party data lives in your data warehouse (Google BigQuery), in user tables and event tables.
- BPP reads your data and unifies customer identities across all sources — every person gets a single, stable profile.
- AI models run on a daily schedule and score your customers (propensity, LTV, RFM, interests).
- You build Audiences and Signals by filtering on any attribute — including AI model scores.
- BPP exports your audiences and signals to your connected ad platforms automatically.
Next steps
- Quickstart — set up BPP and run your first export in minutes.
- Key Concepts — understand the core building blocks of the platform.