Web Relevance Engine

BloomReach’s core technology is our Web Relevance Engine, which harnesses the power of big data, machine learning and large-scale systems science to match relevant products and services to consumers at scale.

Discovering content across multiple channels (web, social, mobile) is an unprecedented challenge due to volume, volatility and the hierarchical architecture of websites (learn more in the Digital Discovery Divide white paper). The Web Relevance Engine is the core of all BloomReach products, continuously predicting and adapting in real time to changes in consumer behavior, queries, site content, competitive content and inventory.

Go inside the Web Relevance Engine to see how it adapts your sites to consumer demand in real-time.

The Web Relevance Engine has six core components: Data Gatherer, Semantic Interpreter, Relevance Scorer, Site Optimizer, Intent-rich Page Generator and Feedback System.

Data Gatherer

The Web Relevance Engine constantly crawls the web to capture general data from websites and consumer queries. The Web Relevance Engine, deeply crawls customer sites (and their competitors’ sites) to collect extensive information. Additionally, using our JavaScript pixel, we anonymously observe aggregate consumer behavior. Finally, customer product feeds are directly integrated into the Web Relevance Engine.

Semantic Interpreter

The Web Relevance Engine develops semantic understanding of the web site content, queries and consumer behavior using a sophisticated learning language model. BloomReach interprets 1 billion web pages and 1 billion consumer interactions daily with 1000 Hadoop jobs. The Semantic Interpreter then delivers the relationships and interpretations into the relevance scorer.

Relevance Scorer

Using the data and analyses from the first two components of the Web Relevance Engine, BloomReach’s algorithms create a score of likely intention versus available information to ascertain relevance and its applicability for each individual user. The highest scoring combinations—which are both unmet and non-excluded—are targeted for Site Optimization. BloomReach algorithms are constantly tuned to maximize quality and efficacy based on both traffic and conversion metrics.

Site Optimizer

The Web Relevance Engine determines what content is needed on an existing or new page to create the best user experience and maximize its relevance for consumers with a specific intention. This includes internal links, related products and relevant content from related pages.

Intent-Rich Page Generator

BloomReach creates semantically rich pages in real-time using proscribed optimizations prioritized by relevance score reflecting relevant content for specific behavioral intentions. Our customer’s websites call our API upon requesting a page for the most up-to-date site optimization.

Feedback System

We continuously monitor consumer behavior – both in general and onsite, specifically from the site optimizations introduced by BloomReach. Our goal is to provide the most relevant and compelling consumer experience while generating revenues, shortening consideration cycles and positively reflecting on our customers’ brands. Through vigorous quality engineering, BloomReach adjusts and adapts to consumer intentions, inventory changes and new relationships between information.