Reassemble the Lego Set

Create entirely new experiences rather than digitizing what came before

Bret Taylor
Inside the expert network training every frontier AI model

Reassemble the Lego Set

"Rather than literally digitizing what came before, if you can create an entirely new experience, it answers the question for a new customer, why should I give this the time of day?" - Bret Taylor

What It Is

A product design philosophy that says transformative products don't copy existing solutions into new mediums—they take apart the component pieces and reassemble them into something native to the new platform that couldn't have existed before.

The framework emerged from Bret Taylor's experience with Google Maps. The first version (Google Local) was essentially a digital Yellow Pages—a "me too" version of Yahoo Yellow Pages grafted onto Google Search. Despite having a link from the Google homepage, it didn't succeed because it didn't answer the question: why use this instead of what already exists?

The breakthrough came when they "disassembled the Lego set" and reassembled the pieces differently—making the map the canvas instead of a small sidebar, integrating local search with driving directions and mapping (previously separate product categories), and later adding satellite imagery. The result was something native to the platform that couldn't exist in the old form.

How It Works

1. Identify the Lego pieces What are the component elements of the existing solution? For Yellow Pages: business listings, categories, addresses, phone numbers, maps.

2. Question the assembly Why were they assembled this way? Usually because of constraints in the old medium (paper size, printing costs, distribution logistics).

3. Remove old constraints What constraints no longer apply in the new medium? Paper can't be interactive, but screens can. Paper maps are static, but digital maps can be draggable and zoomable.

4. Add new capabilities What new capabilities does the medium enable? Real-time updates, satellite imagery, integration with other data sources, personalization.

5. Reassemble for the new medium Build something that couldn't have existed before—native to the platform, not a simulation of the old thing.

How to Apply It

  1. Study the incumbent - Understand deeply what exists and why it's built that way

  2. List the constraints - What limitations shaped the current solution? (Physical distribution, manufacturing costs, bandwidth, storage, processing power)

  3. Identify removed constraints - Which of those constraints no longer apply with new technology?

  4. Imagine from scratch - If you were solving this problem for the first time with current technology, what would you build?

  5. Test for nativeness - Could this product exist in the old form? If yes, you probably haven't reassembled enough.

  6. Find the sizzle - What's the "viral moment" that demonstrates the new capability? (Satellite imagery was the sizzle for Google Maps—everyone wanted to see the top of their house)

Examples

Google Maps (success)

  • Old assembly: Paper map + Yellow Pages directory + printed driving directions
  • New assembly: Interactive map canvas as primary interface, with integrated local search, real-time directions, and satellite imagery
  • Why it worked: Native to the platform—couldn't exist on paper

Google Local (failure)

  • Old assembly: Yellow Pages search on top of Google Search
  • Why it failed: "A digital version of something that had come before"—just digitization, not transformation

The Sizzle vs. The Steak

A key insight from the Google Maps story: the sizzle (satellite imagery) wasn't the most important feature, but it created the viral moment. 90 million people used it the first day just to look at the top of their houses.

When reassembling, identify:

  • The steak: The enduring value proposition (integrated mapping + directions + local search)
  • The sizzle: The attention-grabbing feature that gets people to try it (satellite imagery)

These are "deeply related but not all the same thing."

Source

  • Guest: Bret Taylor
  • Episode: "Inside the expert network training every frontier AI model"
  • Key Discussion: (00:08:49) - On inverting the hierarchy and making the map the canvas
  • YouTube: Watch on YouTube

Related Frameworks

Back to Product Strategy