You're in a pub looking at a beer menu. The publican has very thoughtfully given you some data on the beers he serves, including bitterness units and alcohol percentage. If you are an average beer geek, that IBU number is going to attract your eye, and you may choose the beer with the highest number. But does it mean anything?
Let's start with the process of bittering a beer. Hops are used to balance the natural sweetness that comes from fermenting sweet wort. They add bitterness, flavor, and aroma. The bittering agents in hops are known collectively as alpha acids, but in order to make a beer taste bitter, they need to go through a chemical change known as isomerization, which happens during the boil. The amount of alpha acids a hop contains (1%-20%) and the length of time hops are exposed to boiling wort determine how many "bitterness units" a beer contains (IBU stands for "international bittering unit"). Other factors affect the perception of bitterness, like the strength of the beer and the level of carbonation; 50 IBUs will leave a barleywine tasting sweet but make a pale ale puckeringly sharp.
Okay, now back to the pub. Brewers almost always calculate their beers' bitterness. Larger breweries with labs just run a chemical test on their beer. The results they get are measures of actual acids dissolved in the beer. Small breweries, lacking a lab, have to do it mathematically. Computer programs take the several variables (alpha acid content, length of time in the boil, type of hop product) and calculate an expected IBU. In my experience, these nearly always overestimate the actual bitterness a beer will have. This is especially (but not always) true when a beer is listed at having extravagant bitterness. Many times I've tried a beer that is supposed to be in the 80s or 90s and found it light and approachable. My theory is that these programs tend to over-attribute alpha contributions by later additions, so if a brewery puts in loads of hops between 30 minutes and the end of the boil, it will inflate the IBU number, but add little bitterness. (Holler if you have a better theory.)
Given how many people base so much of their decision on the IBU figure, this is worse than a shame--it's a bit of a catastrophe. Hops are misunderstood enough already. Many people like "hoppy" beers, but are at a loss to distinguish between hop flavor and pure bitterness. IBUs only measure bitterness, and are no guide to flavor and aroma. I suspect the average drinker also doesn't know that IBUs are a chemical measure, not a guide to perceived bitterness. So I wonder--does the thoughtful impulse to include IBUs actually confuse people more than it helps?