Overview

The Product Management service is the product catalogue and lifecycle engine for the Packt platform. It manages product metadata, pricing, format generation, and search for both legacy and Content Lake-backed products.

Historically, published products were rigid, high-effort things which took months or years to produce. Publishers are used to print books, audiobooks, etc. with multiple rounds of review.

Things have changed. In the future, the unit of a product may no longer be the book but a chapter, paragraph, quote, image or code block. Each LLM query could be a new product request which generates a unique response bundling different pieces of Content Lake content. If an LLM sends 1,000,000 queries a day, we would be creating 1,000,000 products a day — each with varying prices and royalties going back to authors.

Although flagship products will still be manually curated and distributed through storefronts like Amazon, Barnes & Noble or Waterstones, the majority of volume could come from automated agents transacting in milliseconds. We want a set of product management capabilities which allow us to meet growing volume demands.

Goals

• Dual Product Model — Support both legacy products with pre-existing distributable assets and Content Lake-backed products where content is assembled from Content Lake documents and formats are generated on demand.

• Real-Time Format Generation — Content Lake-backed products have no stored artifacts. PDF, ePub, and InDesign files are generated on the fly and streamed to the client.

• Version Flexibility — Each Content Lake document attached to a product can be independently pinned to a specific version or set to track the latest.

• Lifecycle Governance — Products move through a defined lifecycle with event-driven notifications on every transition.

• Discoverable Products — Products are searchable via hybrid search and expose their vector embeddings so client applications can overlay their own taxonomy.

Product Models

graph TD
    subgraph Legacy
        A[Existing Asset] --> B[Upload / Attach]
        B --> C[Product Record]
    end
    subgraph Content Lake-backed
        D[Content Lake Documents] --> E[Select and Assemble]
        F[Cover Images] --> E
        E --> G[Product Record]
        G --> H[On-the-fly Format Generation]
    end

Legacy products have a pre-existing distributable asset — a video file, an ePub, a PDF. The Product Management service manages the product metadata and lifecycle, but the content itself is opaque. Legacy products are the compatibility path for existing Packt catalogue items.

Content Lake-backed products do not contain content. They hold product metadata, cover images, and an ordered list of references to Content Lake documents. When a client requests a format, it is generated on the fly by streaming content from the Content Lake. See Creation for details on the format generation pipeline.