Why 99% of product selectors filter out your truly best choices most of the time

You’re shopping for a new home in a city with 10,000 properties for sale. You visit a popular real estate website and, not wanting to sift through all 10,000 listings, you enter some criteria to narrow things down:

• $400,000 to $600,000
• 3 or 4 bedrooms
• In any of 5 neighborhoods

500 results. Better than 10,000, but still too many to review without going bonkers, so you change your criteria from “3 or 4 bedrooms” to “just 4 bedrooms”.

21 results… manageable. You sift through the 21 listings and identify the three most promising-looking houses, then go visit them. On your way back home, however, you happen to drive by an open house and, on a whim, decide to visit that one as well. Beautiful area, 3 bedrooms, $420,000 (the lower end of your price range), and a much better overall choice for you than any of the other three homes that you just saw. You make an offer on the spot.

The open house was clearly an excellent choice, so why was it excluded from the 21 “best matches” from your real estate website search? The answer, of course, is that it had 3 bedrooms, and you were forced to exclude 3-bedroom homes from your search in order to cut down the number of results -even though you were just fine with a 3-bedroom home. The net result? Perfectly good choices were eliminated for no good reason. In fact, in the above scenario, there’s actually a 50%+ chance that not a single one of your top 10 truly best choices would make it through the random culling from 500 results down to 21.

Product selection tools that are based on the concept of “matches” and “non-matches” work fine for eliminating obviously unacceptable choices (as in the above scenario when it got you from 10,000 total listings to 500 acceptable listings), but after that point, they break down. Once all of the remaining choices are basically acceptable, if you continue to narrow your criteria using these tools, they are just as likely to filter out your overall best choices as they are to filter out the barely-acceptable ones.

Essentially, they become random.

The underlying problem is that your truly best choices always represent some kind of feature trade-off. The single best choice is the one that has the best strengths (e.g., lower end of your price range, nice neighborhood) and the least serious flaws (e.g., 3 bedrooms instead of 4). “Match/non-match” tools don’t understand trade-offs, however, and therefore often eliminate choices with great overall combinations of features simply because of one less-than-ideal feature.

The bottom line:
The product selection tools on most e-commerce sites frequently eliminate your overall best choices before you’ve even seen them.