8 common misconceptions about ChoiceBot
ChoiceBot is based on a new concept called "emotional trade-off processing", which is fundamentally different than the concepts on which conventional product selection tools are built. Because of this, people sometimes make incorrect assumptions about how ChoiceBot works "under the hood". What follow are eight of the most common of these misconceptions, along with some clarifications.
8 common misconceptions, with clarifications
"ChoiceBot isn't fundamentally different than [insert name of other product selection technology here]."
We've been immersed in product selection and decision-assisting technologies for over 4 years, and have seen literally thousands of such tools. We know about collaborative filtering, and we've seen iterative filtering tools. We've seen pairwise comparison, weighted average schemes, decision trees and "wizards". We know about Daniel Bernoulli, utility theory and hedonic regression. We probably even know about the special product scoring system that you invented in Excel. So far, none have demonstrated the unique properties and capabilities of ChoiceBot.
The key concepts on which ChoiceBot is based are claimed in a United States patent that was granted after several years of intense scrutiny and research by the U.S. Patent and Trademark Office. If ChoiceBot were not fundamentally different than previous technologies, it almost certainly wouldn't have made it past the first stage of the rigorous patent granting process.
"ChoiceBot is just another 'filter products by features' tool."
ChoiceBot doesn't filter products. If there are 250 products in a database, ChoiceBot always returns all 250 (as opposed to "matches" and "non-matches"). ChoiceBot also puts the overall best choices based on all of your criteria at the top of the list, which product filtering tools don't do. No more sifting through dozens or hundreds of "good matches", no more "Your search returned zero results", no more "There are more than 500 products that match your criteria, please narrow your search", etc.
While very common on e-commerce sites, the "filter products by features" approach can only eliminate obviously bad choices, and is not helpful at all when it comes to zeroing in on the single best choice, once all of the remaining choices are basically acceptable. Worse yet, these tools routinely filter out the user's truly best choices (see "Why 99% of product selectors filter out your truly best choices" for more information on this).
"ChoiceBot is just another 'weight the importance of each feature' tool."
The term "importance" doesn't appear anywhere in the ChoiceBot interface. The ChoiceBot user interface also does not allow users to say that an entire feature (e.g. "Price") is always important, or always unimportant. While some other tools are based on this approach, it doesn't work well because the real answer to a question such as "How important is price to you?" is always "It depends (are we talking about $100, or $500, or $1,000?)".
The question that ChoiceBot asks users is, "How do you feel about the following 4 or 5 sample options for this feature?", and the available responses are all emotion-based (not importance-based). These emotional responses incorporate notions of importance, but also notions of tolerability, expectation and other dimensions that aren't encapsulated by the limited notion of "importance".
"ChoiceBot is just another weighted average scheme."
The algorithm behind ChoiceBot that ranks choices from best to worst is over three pages long, and is not a simple weighted average calculation. It reflects the way that people trade off decision factors in their head when making complex choices based on large amounts of information, a process that is not based on weighted averages.
"ChoiceBot is a rational tool, but people also consider 'irrational factors' when making decisions."
ChoiceBot supports features such as "Cool factor", "Aesthetic design", "Brand loyalty", "Look and feel" and any other reason why someone may prefer one choice over another. It then takes these "soft" factors into account in exactly the same way that it factors in "hard" features (mileage, megapixels, etc.).
ChoiceBot also allows users to enter "irrational" criteria, such as preferring higher prices (for a prestige purchase, for example).
"ChoiceBot is fine for people who base their decisions on emotions, but not suitable for people who prefer to make decisions based on rational, objective criteria."
When choosing the "best" something, there's no such thing as an "objective decision". While this may sound like a new-age platitude, it's widely agreed upon by expert researchers, and can be logically proven through simple thought experiments:
Imagine that two digital camera experts do the same amount of research about which is the best digital camera on the market. After careful consideration and looking over all the feature data, the first expert determines that camera X is the objectively best choice. The second expert, however, determines that camera Y is the objectively best choice (as routinely happens in the real world). After a lengthy, reasonable discussion, neither expert is able to convince the other that their choice is the "objectively best" (again, pretty typical of real-world situations).
In this scenario, what test, methodology or measurement could be used to determine which camera is the "real" (i.e. most objective) best choice? Such tests, methodologies and measurements don't exist (or don't work), because "objective goodness" isn't an actual property if things in the real world, and therefore can't be measured. "Objective goodness" is merely your brain's unconscious assessment of how much it likes something, and the tricky part is that your brain tricks you into thinking that the "goodness" or "badness" that it perceives is actually a real, measurable property of that thing that exists outside of your head. But it doesn't.
Simple psychological experiments show that what people assume to be "objectively good" and "objectively bad" features depend on a cocktail of factors -including genetics and life experiences- which can have little to do with the actual thing being considered. This is why perceptions what the "objectively best" choice is vary from person to person, with no single person ever being any "more objective" than anyone else. Everyone, however, tends to view their own assessments and decisions as "reflections of the way things really are", and the decisions of others as "emotionally biased" –especially if their decisions are different. This well-documented psychological phenomenon is sometimes referred to as the "bias blind spot".
This isn't to say that there's no difference between people who tend to consider "sensory" features (appearance, handling, look and feel, smell, cool factor, etc.) more than people who tend to ignore these kinds of features. The people who ignore them, however, can't claim to be "more objective". They're just people who tend to ignore certain kinds of features…
Choosing a "best choice" is always an emotional decision because, without emotion, such decisions would be impossible to make. If you really could take all emotion out of such a decision, you wouldn't prefer any feature over any other feature, and every choice would just seem equally "neutral" (in fact, you wouldn't even care about the outcome of the decision in the first place). This is why even CEO's making multi-billion-dollar financial decisions use emotion to make these decisions –there's simply no other way to make them.
This also isn't to say people don't consider data (sometimes a lot of data) when making "best choice" decisions, however the process of assessing the data and coming to a final choice is always entirely emotional. Even complex decision-support software programs require an emotional input at some stage of the game (usually in the form of "weights" that are entered by an emotionally-equipped human user) in order to be effective.
ChoiceBot recognizes this reality, and helps users zero in on the choice that they will perceive as being their objectively best choice, because that's what "best choice" actually means.
"How can you prove that ChoiceBot recommends the best possible choice?"
As discussed in the previous FAQ item, it's necessary to clarify what the phrase "best possible choice" actually means, because many people misunderstand what it actually means.
Many people think of the "best possible choice" as a choice that would be made without any bias or emotion, using some kind of scientifically validated decision-making methodology, and that any reasonable person would agree is the best choice. This "objectively best" choice is often contrasted with an "irrational" choice, which is a different choice that a "less rational" individual may prefer for "emotional reasons".
When it comes to "best choice" decisions, however, there are several problems with this view:
- How would the scientific decision-making methodology be tested, to verify that it really was recommending the "objectively best" choice?
- If the decision-maker doesn't agree that the "objectively best" choice really is their best choice, then the decision methodology would be claiming to know the decision-maker better than they know themselves. In the absence of mind-reading technology, this seems unlikely…
The underlying problem here is that preference decisions are made for the purpose of maximizing the individual decision-maker's happiness and well-being. Period. They are, by definition, subjective.
If, after the fact, a person feels that a particular decision was a bad one, then, by definition, it was a bad one -regardless of what any decision-recommendation methodology has to say about it. If, on the other hand, they feel confident that it was the best possible choice they could have made based on the information available at the time, then by definition, what they chose was the best possible choice.
The vast majority of ChoiceBot users intuitively agree that their top-ranked choice in ChoiceBot is, indeed, their truly, genuinely best choice based on all available information. By definition, this means that ChoiceBot recommends the best possible choices -even though it recommends different "top choices" for different people.
"People are lazy, and won't to expend even a few minutes of effort to make a purchase decision."
Our (speculative) opinion on this is that this is a hangover effect of past generations of product selection tools (and most current ones), where the effort required to use the tool was not rewarded by receiving useful recommendations. Several minutes of work to get recommendations that are clearly not your best choices (or to get zero results, or to get 107 "good matches") is a bad bargain, and one from which most people learned to shy away -even if that meant slogging through every single available choice "by hand".
On the other hand, spending a few minutes to zero in on what are clearly your best choices is a much more attractive proposition…