Another approach is to do a full category only search and filter out the unwanted items after retrieval. Budweiser is a small percentage (under 2%), so this could work well here. For some words, it would result in high exclusion rates.
I have a tool for determining best keywords for getting widest range of items from a small category. Here's a keyword list for 563:
non-cumulative frequency table: (matches X in 1000)
bottle: 495
bottles: 146
vintage: 116
brewing: 111
unopened: 109
glass: 88
stout: 86
brewery: 80
stone: 70
dogfish: 65
collectible: 60
empty: 56
limited: 48
barrel: 47
three: 47
black: 45
floyds: 45
adams: 44
bruery: 42
...
budweiser: 33
other bud* variants would be <20 if present.
And cumulative title match percentages. Use like this:
bottle* matches 63%
(bottle*,vintage) matches 64.9%
(bottle*,vintage,brewing) matches 68.6%
bottle* 0.63
vintage 0.649
brewing 0.686
unopened 0.709
glass 0.731
stout 0.77
brewery 0.789
stone 0.804
dogfish 0.824
collectible 0.826
empty 0.82
limited 0.831
barrel 0.855
three 0.868
black 0.88
floyds 0.881
adams 0.887
bruery 0.905
Snapshot based on small sample (1000 items), default sort, not representative of entire category, and may vary with items.
Original Post
)
)
)
Use 'the' instead then.
...at it a little.