What Is RTB House?
According to the company’s technical publications and official press releases, RTB House is an independent, globally operating marketing technology (AdTech) and programmatic retargeting provider built 100% on Deep Learning algorithms.
Whereas traditional retargeting companies (Criteo, Google Display Network) rely on rule-based segmentation models or standard linear machine learning models, RTB House states in its technical reports (RTB House Whitepaper, 2021) that it uses multi-layer artificial neural networks to predict a user’s browsing rhythm, on-page interactions, and purchase probability.
Deep Learning vs. Traditional Machine Learning
Note: The comparison below is based on the architectural descriptions in the technical whitepapers published by RTB House; the company’s internal algorithm weights are a proprietary trade secret (proprietary technology).
| Dimension | Standard Retargeting (Rule-Based / Standard ML) | RTB House (Deep Learning Model as Described by the Company) |
|---|---|---|
| User Modeling | Static segments (Cart abandoners, Visitors in the last 7 days) | Sequential Journey Vectors (Sequential Embeddings); the user’s consecutive browsing signals |
| Creative Selection | Manual templates and rule-driven product matching | Dynamic Creative Optimization (DCO); automated selection of color, layout, and copy |
| Bidding | Fixed/static bids based on average segment value | Dynamic eCPM bids driven by Individual Conversion Probability (CVR) estimation |
| Feature Extraction | Rules defined by human analysts | Latent features extracted from raw data by deep neural networks |
RTB House Workflow (Conceptual Illustration)
[User Site Interaction] ──► [Deep Learning Model (Prediction)] ──► [Auction & Bidding]
│ │
▼ ▼
[creativecdn.com Data Collection] [Dynamic Creative Generation (DCO)]
│
▼
[Personalized Impression](The diagram above is a conceptual flow illustration summarizing the process described in the company’s public technical documentation.)
1. Signal Collection and Behavioral Tracking
The IDs of the products or projects viewed by a visitor are collected via a JavaScript tag added to the website. Requests are routed to the RTB House infrastructure through creativecdn.com (AMS, FRA, IAD edge servers).
2. Conversion Probability Estimation (CVR Engine)
According to company statements, recurrent neural networks (RNNs) and attention mechanisms (Attention/Transformer models) analyze the user’s page-to-page navigation rhythm to calculate a conversion probability score per impression.
3. Dynamic Creative Optimization (DCO)
When a bid is won, the banner served is not a static image; at the creativecdn.com edges, assets (image, headline, call to action) matching the category and price range the user browsed are assembled on the fly.
Cookieless Developments and Privacy Sandbox Experiments
Google’s revision of its timeline for fully deprecating third-party cookies, and its announcement of a new model that gives users a choice, have prompted the industry to update its transition strategies.
- Protected Audience API (Formerly FLEDGE) Trials: In 2022–2023, RTB House was among the first buyers to join Google Chrome’s experimental Privacy Sandbox Origin Trial processes and ran on-device, in-browser auction tests (
navigator.joinAdInterestGroup). - Current Status and Industry Reality: Due to technical barriers such as latency, browser processing overhead, and limited industry adoption, the Protected Audience API has not yet become a mainstream production standard capable of fully replacing third-party cookies in the global programmatic market. RTB House and other industry stakeholders use both first-party data identity solutions (UID2, ID5) and traditional hybrid cookie-based/cookieless models in parallel.
Applying RTB House in the Real Estate Sector
For high-ticket real estate projects with long sales cycles, RTB House technology is configured with a strategy that differs from classic e-commerce:
- Inventory and Portfolio Feed (XML / JSON Feed): The project’s unit types (1-bedroom, 2-bedroom, Penthouse, Detached Villa), current prices, floor plans, and delivery dates are connected to the RTB House engine via a dynamic data feed.
- Investor Intent Classification:
- Users who spend time on the project’s payment plan and currency-based return calculation pages are tagged as “Financial Return-Focused Investors.”
- Users who browse the social amenities, children’s playgrounds, and landscaping galleries are assigned to the “Family / Lifestyle-Focused Buyer” segment.
- Dynamic Real Estate Banner: As the user browses independent news or business websites, instead of a generic project logo, they see a dynamic banner showing the current installment amount and floor plan of the exact unit they viewed in that specific block.
- CRM Offline Conversion Integration: When the customer visits the sales office or signs a contract, this stage is reported from the CRM to RTB House via an S2S API. As a result, ad impressions to customers who have already purchased are stopped immediately, preventing wasted budget.
Citation
Yalçın, K. (2026). What Is RTB House? How Does It Work? KorayYalcin.org Publications & Research.
