Criteria-Weights

This was the feature that really got me thinking. The majority of candidate-matching engines have criteria-weights, unlike the shopping engines who rarely use it.*

When I had to use this feature on the candidate-matching sites it felt... incredibly annoying. The weights were in a constant state of flux. On the initial questions, weighting often felt like a coin-flip decision. And future questions made me re-think the relationship with prior weights! Overall, I did not feel comfortable with this feature.

There was one glorious exception, the sliding-scales method of USAToday. It was optional. It was after-the-fact. It was interactive. It just felt right.

* Actually, progressive-elimination attribute engines do have a subtle form of criteria-weighting. The first attribute you select is the most important. The next attribute is the next most important, and so on.

A La Cart vs. Guided Process

Many of the candidate-matching engines require their participants to answer every question. Or if they allow non-responses, they do not draw attention to it.

In contrast, shopping engines have gravitated towards the a-la-cart model of comparison. Select an attribute, see updated results, repeat, or stop anytime...

It's been awhile, but I thought more shopping engines employed wizards in the dot-com days. If so many analytics-conscious sites use one method, does that mean the market has spoken? Or is guided-process simply more resource-intensive to setup across categories?

To me, a-la-cart comparison feels more natural, more empowering. Yet I wonder if some people would prefer the guidance of the "wizard" approach, if done correctly?

Best Match vs 100% Match

Another interesting dynamic... the "wizards" rarely find results that match your criteria exactly. As opposed to most shopping engines, where every result must match your criteria exactly.

Because the wizards seek the best match, instead of a perfect match, they are able to juggle more criteria. This could help when you have trouble prioritizing or narrowing your criteria.

Put another way, I think the exact-match drilldown is at its best when every criteria is critical. But when there many nice-to-have criteria, you have to be very careful with "sudden death" attribute filtering.

Miscellanous...

Did you notice that PriceGrabber's (excellent) Candidate Match tool uses a start-to-finish format that is different than the default comparison shopping format?

* It is because the answers are too long and too numerous?

* It is because the results are not intended to be actionable?

* Is it needed for aggregating results and sharing w/ friends?

Summary

I do not think the "wizard" format is any threat to the default comparison shopping format, but it could be a useful in certain situations. I think anyone wanting to create a comparison engine should have a full set of tools in their "tool belt" and keep those tools as sharp as possible.