The Digital Affairs Minister on Projects and Products: Why Taiwan’s Software Struggles to Go Global

The new digital affairs minister’s distinction between projects and products is correct. For a software company, owning a product is better than taking projects wherever they arise. Yet behind that apparent common sense are practical considerations and challenges. Does the approach suit small Taiwan and its current circumstances, or how should the environment change? Taiwan’s most successful software company is still Trend Micro, founded in 1988, just one year after TSMC. Founder Steve Chang’s vision and outlook mattered, but so did the historical setting. In the late 1980s and 1990s, Taiwan was the world’s computer-production powerhouse. Trend could sell bundled products on that foundation. Nearly forty years later, it remains the software company Taiwan can point to. This returns to output per employee. Semiconductors’ output, profit contribution, and compensation per employee are remarkable. ASPEED’s average employee compensation expense is 5.5 million, almost matching Silicon Valley giants. Taiwan’s software companies compete for talent with NVIDIA and Google, semiconductor companies TSMC, MediaTek, and ASPEED, and electronics companies ASUS and Wistron. Without Taiwan’s best talent, how can they build world-class products? Software is almost a winner-take-all market. Another question: at the Olympics, should we expect Taiwan to win gold in every event, or focus resources on a few priorities? Can Taiwan’s population, given its physical characteristics, win in traditional major track-and-field events? It is not impossible, but requires resources and long-term planning. Would that crowd out other events? Or should Taiwan follow South Korea in focusing on medals in shooting and archery? Standardized products are also constrained by domestic market size. The US market is enormous. Software made there can serve not only Americans but people worldwide. Imagine a Taiwanese company called TaiHard hiring 100 engineers to make word-processing software, while Microsoft in the United States assigns 1,000 to Word. Assume Taiwan’s engineers are ten times as capable. The US product still costs much more to develop, with more, higher-paid engineers. But Microsoft can recover those costs and make a profit in the US market alone, then take the product worldwide. Having already recovered costs at home, it treats other countries as peripheral markets, where whatever it charges is additional profit. TaiHard might not even recover development costs from Taiwan alone. Suppose it charges 3,000 per copy and Microsoft also charges 3,000. Microsoft’s product has no global compatibility problem. And do you really think software made by 1,000 American engineers would be harder to use than software made by 100 Taiwanese engineers? A more likely scenario is that Microsoft has recovered all development costs in America and made substantial profits, so sales in other peripheral markets add profit at any price. The Taiwanese software company TaiSoft calculates it needs to charge 3,000 to break even. But Microsoft’s software works well, is widely compatible, has a recognized brand, and sells for just 300: one-tenth TaiHard’s price. Microsoft’s word processor then sweeps the world. Ultimately, this is about market size, especially when software works globally and American products are almost universally treated as the standard. These views may be unwelcome, but I offer them to digital affairs officials for consideration. --- On going global, Taiwan’s electronics hardware and key-component companies do so by finding a distinctive place in a supply chain. Satisfy the next customer with cooperation, quality, pricing, and delivery, and global reach follows naturally. Volumes for each product start in the millions, then feed into final products sold worldwide. Software lacks this supply-chain structure and must actively consider how to serve customers globally. TSMC leads the semiconductor supply chain: adoption by TSMC means serving global customers. Electronics companies also keep spawning new branches. Consider the many companies emerging from Acer; even Trend Micro’s current CEO worked there. The Kinpo–Compal group and Foxconn are other examples. Software has yet to produce a major company with comparable influence. Many founders did come from Trend Micro, but their impact remains some distance behind.