Blending is an important skill for correcting minor problems or imbalances in beers. It has produced more than one Ninkasi winner in the past and you can too...if you have the palate and skill.
So true. Over the years I've had batches that didn't end up as intended, and I was always able to salvage them by blending with other brews I had on hand (some of my favorite standard 'house' recipes were actually the result of some of these blends). Taking good notes on brew day and careful measuring at blending time allowed pretty accurate reproduction of the blends and repeatability by subsequently using some 'reverse engineering'.
The blending idea has certainly been used in commercial settings as well, both in actual production as well as experimentation/formulation. It's standard procedure in some UK breweries, and it has also been suggested (whether it's true or not) that many of Pabst's contract brewed 'legacy' brands actually may be a variety of blends using a core set of standard brews.
In the early '80s, I attended a beer talk/tasting in New York conducted by Matt Reich, the guy behind the excellent New Amsterdam Lager. During a chat with Matt afterwards, I asked how he and his consultant (the esteemed Joe Owades) came up with the flavor profile and formula for New Amsterdam (which was a very good beer, and unsurprisingly in many ways quite similar to Sam Adams Boston Lager, which hit the market around the same time). His answer was that one of the primary things they did to arrive at a prototype for the final product was to concoct and taste various blends of a large selection of commercial products (both domestic and imported) which were available at the time.