---
title: "It's Not Time to Build"
date: 2026-08-24T09:40
author: Julien Reszka
description: "Manufacturing is the least profitable stage of the value chain. Building only pays off when it starts from a real UX insight and ends in something sellable."
keywords: ["startups", "design", "marketing", "product", "strategy"]
canonical: https://julienreszka.com/blog/it-s-not-time-to-build/
---

# It's Not Time to Build

Manufacturing is the least profitable stage of the value chain. Building only pays off when it starts from a real UX insight and ends in something sellable.

Marc Andreessen's [2020 essay](https://a16z.com/its-time-to-build/) argued that the refusal to build, houses, vaccines, factories, trains, was the real failure behind a stagnant decade, and it became a rallying cry for founders: stop planning, start building. The advice is not wrong about effort, but it treats building as valuable on its own terms. It usually is not.

In 1992, Acer founder Stan Shih drew a chart to explain why his own company kept making PCs on thin margins while other firms in the same industry did not. Plot value added against the stages of a product's life, from concept to sale, and the line dips in the middle and rises at both ends, forming a smile. Assembly and manufacturing sit at the bottom. Research, design, and branding sit at the peaks on either side of it. The depth of that dip is not fixed: semiconductor fabrication and other precision manufacturing niches sit closer to the peaks than the trough, because technical barriers to entry are the same kind of moat that a strong brand or a sharp UX insight provides everywhere else on the curve. A 2021 analysis of decades of international trade data found the same shape at the country level: developing economies have taken on more of the world's assembly work while developed economies hold onto a larger share of the design and R&D end.

- Apple's estimated gross profit on a $299 30GB Video iPod in 2005 was $80, a 36 percent margin, captured almost entirely through design, in-house software, and marketing
- The workers who did the final assembly of that same device were paid roughly $100 a month, under a penny a minute, for the physical act of building it
- A 2008 study of value chains across industries found value added concentrating at the design and R&D end and the marketing and sales end, with production capturing less throughout
- Contract assemblers in the electronics industry reported gross margins around 9 percent in the mid-2000s, well below the roughly 33 percent average across the industry's leading firms
- The pattern held nearly two decades later: in fiscal 2023 Apple's gross margin was about 39 percent, while Foxconn, which assembles most iPhones, posted a net margin near 2 percent

Building is not easy, and none of this argues that it is. What it argues is that the hours spent building are worthless unless two conditions hold: the thing being built has to answer an actual insight about how a user behaves, not a guess dressed up as one, and once it exists, people have to recommend it to each other without being paid or begged to, because that is what marketing actually is. A working prototype built on the wrong assumption about user behavior is just a well built mistake. A working prototype nobody talks about on their own is a well built secret.

A UX insight that gets fifty people to preorder something that does not exist yet only clears the first bar, before a line of code is written. It proves the behavior is real, not assumed. It does not clear the second one. A preorder can be bought with ads or pulled out of your own network by asking; word of mouth is what happens after, when the people who used the thing describe it to somebody else because they wanted to, not because you asked them to. Get the insight wrong and you build the wrong thing well. Get it right and skip earning that recommendation and you build the right thing that nobody hears about. Either way, the manufacturing you were so eager to start turns out to be the cheapest part of what you end up paying for. Andreessen's own essay names the same bottleneck under a different word: the problem, he wrote, is desire, we need to want these things. Desire is not generated by shipping faster. It has to be earned first, by an insight worth wanting and a recommendation worth repeating.

---

**Actionable insight:** Before you build anything, name the specific user behavior your idea is based on, and get a real preorder or waitlist signup to confirm it. That validates the insight, nothing more. Marketing happens later, when people describe it to someone else unasked.

## Key figure

**$80 vs $100/month** — Apple's estimated gross profit on each $299 30GB Video iPod in 2005, versus the roughly $100 monthly wage paid to the workers who assembled it

*Source: Dedrick, J., Kraemer, K.L., Linden, G., Who Profits from Innovation in Global Value Chains? A Study of the iPod and Notebook PCs, Sloan Industry Studies Annual Conference, Personal Computing Industry Center, UC Irvine, 2008*

## Myth vs reality

**Myth:** Building the product is where the value comes from. Effort put into building is valuable on its own, and design and marketing are secondary polish you add once the thing exists.

**Reality:** A 2008 study of value chains across industries found value concentrates at the design and R&D end and the marketing and sales end, with production capturing the least. Apple's own iPod teardown shows the same pattern.

*Source: Mudambi, R., Location, control and innovation in knowledge-intensive industries, Journal of Economic Geography, 2008*

## Quote

> "The problem is desire. We need to want these things."
>
> — Marc Andreessen, cofounder of Andreessen Horowitz, whose own essay ends up naming the same bottleneck this post argues for

*Source: Andreessen, M., It's Time to Build, a16z.com, April 2020*