First, saving a few bottles (or cases) of a wine has an important research application for a winery. All too often I see wineries with less than a handful of released vintages claiming their wine will drink well for 20-25 years. This may be the case, but there is no actual evidence that it will be the case. Sure, a winery can't wait 25 years to make the claim that their wines will actually last a generation as a serviceable beverage, but at least tasting a few wines after they've been laid down for a few years will provide some evidence of such claims.
Second, wineries need to document the results of the choices they made between bud break and the time the finished wine is bottled. If the entire production is sold, a winery will have no way of knowing how the acid adjustment or the spinning cones (or the non-interventionist approach) turned out. Holding back a few samples is necessary for finding out how the wine changes with age. Some people call it drinking, but in the wine industry it should be called research and development. Wineries should always try to be improving on their most recent vintage and there might be no better way than drinking their wines when they have aged.
Third, and maybe most important for consumers, library wines means library releases. Reserving more than a few cases allows a winery to re-release wines that may be their customers weren't able to buy on release. Many consumers don't have the patience or the ability to hold on to wines until they're at their peak, so being able to buy directly from the winery a few years down the line has its advantages. Knowing how a wine ages is often a key piece of information consumers use when trying to decide on how much of a currently release to buy. One of a winery's advantages is charging a premium for library wine. As long as a the premium isn't outrageous, it's a win-win situation for the winery and the consumer.
|Scherrer "Shale Terrace" Zinfandel, Alexander Valley AVA|