ARLIER this month, Marriott International unveiled new designs for the rooms in its various hotel chains. These are not routine updates to replace worn-out furniture or carpets. They represent a significant shift from the cookie-cutter standardization that built Marriott into one of the world's largest hotel companies, to a new emphasis on aesthetics and personalization.

"It's a makeover," said a company spokesman, John Wolf.

From granite countertops and aromatherapy shampoos in the bathrooms to piles of pillows on the beds, the rooms aim to please a style-conscious new generation of travelers. The floral bedspreads once ubiquitous in hotel rooms are out. After all, says Mr. Wolf, "Who has one of those at home?"

Marriott is following an industry trend. Over the last several years, hotel chains have been competing ever more intensely to upgrade the look and feel of their rooms. That competition was set off by Starwood Hotels and Resorts with its stylish redesigns of Westin and Sheraton rooms beginning in 1999.

This emphasis on aesthetics goes beyond hotels to all sorts of products and commercial environments, from the design of cellphones and trash cans to the look and feel of stores and restaurants. The quality of goods and services is increasingly judged not only by function but also by style.

These upgrades present real problems for economists charged with tracking inflation. Hotels hope to sell their redesigned rooms at higher rates. If they succeed, should economists count those higher rates as contributing to inflation or simply as consumers' paying for more to get more value?

Hotel rooms thus provide a relatively simple example of the complex judgments economists have to make when they try to track inflation in a dynamic economy.

The quality of goods and services is always changing, often for the better and often in intangible ways. If those changes take place at the same time that prices go up, it is hard to separate paying for greater value - for, in effect, a different good - from paying for inflation.

"Quality change has typically been considered the least tractable problem associated with the Consumer Price Index," the National Research Council said in 2002 in a report on cost-of-living and price indexes.

Though businesses and government treat it as a hard number, the Consumer Price Index is not a fact of nature like the atomic weight of lead or the speed of light. It is a statistical construction that requires a lot of judgment calls.

In theory, "inflation" represents an increase in the overall price level, the numbers on every price tag, not in the cost of any particular good. It is independent of such underlying relative values as how many bananas equal a shirt or how many gallons of gas equal a new laptop computer.

The index tries to measure the net increase or decrease in prices across all consumer goods. Separating the overall price level from relative shifts in the prices of specific goods is difficult, in part because it is so hard to define those goods consistently over time. A banana may be only a banana, but a hotel room may be something more.

To compile the index, surveyors for the Bureau of Labor Statistics go back to the same businesses month after month and collect prices on specific goods and services. The composite of all this information forms the index.

But the right way to select and define the items to track is not obvious.

When, for instance, should the index add completely new products like Botox or cellular phones? New technologies are often very expensive initially but then drop rapidly in price. If the bureau waits too long to add a product category to the index, as some critics say it did for cellphones, it can miss the big price drop and inadvertently overstate the rate of inflation.

And, as hotel rooms illustrate, even well-established products and services are always changing.

When hotels do routine renovations, like replacement of worn carpet, the bureau simply ignores the change and treats the room as the same good before and after. If the price rises after the maintenance, that increase raises the index.

Sometimes, however, a hotel adds specific, measurable new attributes found in other, higher-priced hotel rooms. Then the bureau treats the room as a new product.

"Suppose a hotel room selected previously now includes complimentary breakfast and dinner, has been renovated, and now includes a VCR," the bureau explains on its Web site. "The price of the room has increased, and the price change is strictly due to these enhancements. This change in price, due to the change in quality of the services provided, would not be reflected in the index."

The "old" room disappears from the price index and the "new" room takes its place. The new, higher rate is treated as the inflation-neutral price of a more luxurious good.

These procedures are more straightforward in theory than in real hotel rooms. With their emphasis on aesthetic upgrades, most of the recent hotel improvements do not fall easily into either of these two categories. They are neither easily definable new attributes nor routine maintenance.

With hotels and many other goods and services, the question economists have to answer, or at least to guess at, is what customers think of the changes. It is not enough that hotels are spending a lot of money ($190 million for linens alone by Marriott) or that they believe the rooms are better. Value is in the eye of the consumer.

"Maybe the average hotel guest wouldn't care whether there are 300-count or 200-count sheets. Maybe they just want a bed that's not caved in," says Dan Ginsburg, a supervisory economist at the Bureau of Labor Statistics.

If the guests do not see the difference as more valuable, then the higher charge should be factored into inflation measures. If the redesigned room costs $15 a night more, the bureau's statistics should count the whole $15 toward an increase in the index.

If, on the other hand, guests care as much about aesthetics as hoteliers believe they do, it would be irresponsible to treat the $15 as a true price increase.

Depending on how much guests value the change - suppose they would have willingly paid $25 more for those amenities - the higher rate could even represent an effective price cut.

Bureau economists, Mr. Ginsburg says, are "caught in this conundrum." Underestimate quality improvements, and they make inflation sound too high. Overestimate improvements, and they make it sound too low. And there is a reasonable case to be made for whatever they do.

Measuring inflation, he acknowledges, "is more of an art than a science, unfortunately."

Virginia Postrel (www.dynamist.com) is the author of "The Substance of Style: How the Rise of Aesthetic Value Is Remaking Commerce, Culture and Consciousness."