---
title: "How does Smart Recommender Work?"
slug: "how-does-smart-recommender-work"
description: "Explains Insider One's Smart Recommender 5-step process: catalog/event integration, algorithm results, strategy setup, campaign creation, and analytics."
updated: 2026-01-09T06:30:26Z
published: 2026-01-09T06:30:26Z
canonical: "academy.insiderone.com/how-does-smart-recommender-work"
---

> ## Documentation Index
> Fetch the complete documentation index at: https://academy.insiderone.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How does Smart Recommender Work?

Understanding how [Smart Recommender](/v1/docs/what-is-smart-recommender) works helps you confidently launch personalized product recommendation campaigns across web, app, email, and more.

This guide walks you through the complete journey, from connecting your product data to designing recommendation widgets, so you can create campaigns that boost conversions and engagement.

Smart Recommender automatically turns raw data into real results. By understanding the process, you can:

- Launch campaigns faster with fewer dependencies.
- Select the most suitable strategy for each placement and goal.
- Optimize performance with insights and analytics.
- Personalize across every touchpoint confidently.

Whether you’re just starting or scaling personalization, this guide helps you unlock the full value of Smart Recommender. Here’s the 5-step process that powers every recommendation:

## 1. Catalog and Event integration

Smart Recommender needs two data inputs before it can recommend anything:

- **Catalog Data**: Product details like names, images, categories, prices, and stock availability.

Synced via XML, Catalog API, Clickstream, or Shopify.

- **Event Data**: User interactions such as product views, add-to-carts, purchases, or even offline transactions.

Collected from web, app, and offline channels through Insider’s integration setup. Together, these datasets give Smart Recommender a live view of your products and user behavior.

## 2. Algorithm result generation

Once data flows in, algorithms process it into recommendations. Depending on your strategy, the engine can:

- Recommend top-selling or trending items.
- Recommend contextually (based on the page a user is on).
- Personalize using user affinities and behavior history.
- Combine multiple strategies into a hybrid model.

The result is a continuously updating recommendation list, always fresh and relevant.

## 3. Recommendation Strategy creation

In the InOne panel, you define what to recommend and how to control it:

- Select algorithm(s).
- Define the number of items to display.
- Add filters (price, availability, categories).
- Set exclusions (e.g., don’t show already purchased or viewed items).

This step aligns the engine’s intelligence with your business goals.

## 4. Campaign creation

Your strategy is delivered through campaigns, which define:

- **Where**: Web, Mobile App, Email, Web Push, or InStory.
- **Who**: Target audience, segments, or triggers.
- **When**: Schedule, frequency, and behavior-based timing.
- **How**: Widget design (carousel layout, titles, CTAs, styling).

Campaigns are the vehicle that brings your strategies to life across channels.

## 5. Performance tracking & analytics

Once live, Smart Recommender measures results in real time through [Smart Recommender Analytics](/v1/docs/smart-recommender-analytics):

- Impressions
- CTR (Click-through rate)
- Add-to-cart rate
- Conversion rate
- Revenue influenced by recommendations
- A/B test performance

With these insights, you can continuously optimize and prove the ROI of personalization.

Smart Recommender is more than an engine—it’s a closed-loop system that connects product data, user behavior, algorithms, campaign execution, and analytics to help you drive growth at scale.

## Related

- [What is Smart Recommender?](/what-is-smart-recommender.md)
- [Key Benefits of Smart Recommender](/key-benefits-of-smart-recommender.md)
- [Smart Recommender Quick Setup Guide](/smart-recommender-quick-setup-guide.md)