Innovation cuts both ways, of course. Atari was an early innovator with its video-game technology. Today its machines are yard-sale material, victims of superior innovations from gamemakers like Nintendo and Sony. VisiCalc touched off a business revolution with its spreadsheet software in the late 1970s, but that market now belongs to the Microsoft juggernaut with Excel. It’s not enough to innovate once. You have to keep it up. Apple Computer turned the PC industry around when it rolled out the Macintosh in 1984, and then gave in to complacency.

For the investor, too, taking innovation into account isn’t enough. Paying attention to valuation, or the price being paid for a stock, should always be an element of an investment strategy. While it may be true that an investor can “pay up” somewhat for a high-quality business and still come out all right in the end, there are limits to this logic.

“Buy and hold,” for example, does not mean buy at any price, even for a tech stock on the verge of setting the world on fire. If stock prices reflect future expectations (which they do), there must be room in the stock price for when these expectations become a reality. And the more room, the better for the buy-and-hold investor in technology stocks.

To use Cisco Systems as an example again, the company famously reached a peak market valuation of $592 billion in the spring of 2000, when its stock price briefly reached $80 a share. At that price, the company’s valuation was equivalent to more than 5 percent of U.S. gross domestic product (GDP) the previous year. Meanwhile the company’s actual sales year amounted to merely .15 percent of GDP.

By contrast, the amount of money Americans spent on motor-vehicle, furniture and household equipment during all of 1999 worked out to just over 6 percent of GDP. Where was the room for Cisco’s market valuation to grow? Was it rational to expect that the company was worth the value of all of the cars, couches and dishwashers sold in the country during a given year? Eventually, investors determined the answer was “no.”

In the end, investors need to remember that the technology sector is inherently risky, and not just because it’s subject to enthusiasms that can drive a stock past rational valuations. Companies that live by constant innovation can die by it, too. An investment strategy must account for this possibility. One tactic: include more innovation-resistant stocks in a portfolio to offset some of the risks inherent in technology stocks. That’s probably a lesson that more than a few investors have figured out on their own by now, and in a particularly painful way.

For example, how much can an industry such as manufacturing potato chips change from year to year, or even from decade to decade? Human taste buds and snacking habits will not change with the next iteration of the Windows operating system or be greatly affected by the geometric growth of the Internet. There are only so many ways to skin and cook a potato. This homely reality ensures a level of business predictability that a computer-chip company will never enjoy.