A few days ago, I wrote a post called Defining relevance in search. In it, I suggested that relevance is not a single thing, but a set of qualities:

1. The human connection: relevant results connect to the searcher.
2. The discovery angle: relevance can be expected or unexpected.
3. The subjective nature: the degree of relevance changes from person to person and moment to moment.
4. The measurement conundrum: the degree of relevance occurs along a spectrum that makes it impossible to achieve 100%.

My blogbuddy David Berkowitz responded with some astute elaborations:

With the human connection, I think it’d be interesting to divide that into other categories. For instance, there’s the human connection of explicit interest - “I know I want a new car, so ads about cars are relevant to me.” Then there’s the connection of assumed interest - “I know I want a new minivan, so it’s true that I’m also in the market for baby furniture.” Then there’s the interest that technology unearths - “I know I want a new car, but since I happened to visit a number of sites for vacation packages to South America, then I also am in the market for deals on hotels in Ecuador.” In each instance, it’s relevant to the human, but for different reasons.

I agree with David, and I think it’s worth taking the opportunity to further elaborate on each of these qualities and their potential manifestations. When I started writing this piece, I realized that we can go into a fair bit of detail here, so I’ve decided to devote a post to each; this topic will continue in a multi-part series.

The human connection
As I stated in the earlier post, relevance isn’t relevance unless it’s connected to a person. It only exists in the eyes of the beholder. A search result might match every keyword, but it doesn’t become relevant until the searcher says, ‘Yes! That’s what I want!’

To me, this is the first and foremost principle of relevance. If we don’t get this bit right, nothing else matters. If we get this bit right, we can tolerate a whole lot of ambiguity in the rest of what we do.

1. Core interest
   Core interest is what VortexDNA is about. This is the stuff you are interested in because it connects with who you are at the deepest level. This is the stuff you find online that makes you feel like somebody knows you inside and out, or the stuff that makes you want to forward it to everyone you know.

   The premise behind VortexDNA is that this type of connection can lead to all of the others. If you can tap into core interest, every other effort will be more powerful because it will be based on the person, the individual.

   If you don’t mind, David, I’ll use your list as a starting point for the other ways in which the human connection can play out.

2. Explicit interest
   Explicit interest is search at its most basic and its most obvious. I want a book. I know its title. I go to Amazon and I enter the title of the book. It’s completely overt and, therefore, the easiest to deliver.

   Just because it’s the easiest, though, doesn’t mean it’s easy. Included in the category of explicit search is search where the user knows exactly what she wants but doesn’t know how to ask for it. Maybe you heard a song and remember just a snatch of lyrics, no idea of the title or the artist. Maybe you’re trying to find who said a particular quote, but don’t realize that you’ve got the words the wrong way around. There are all sorts of challenges with explicit interest, but this has been the first line of attack for search engines and, therefore, the area in which they’ve gotten the most skilled.

3. Assumed interest
   Assumed interest has its roots in brick-and-mortar merchandising efforts. Shopping for beer? You’ll be happy to find potato chips in the same aisle. Buying a new printer? Surely you’ll want to stock up on ink cartridges at the same time. Complementary offerings are the manifestation of assumed interest. Like explicit interest, these connections tend to be evident or become evident through user behavior.

4. Suggested interest
   I’m referring to David’s ‘interest that technology unearths’ as ’suggested interest’. This is interest that the user was unaware of, but that can be extrapolated and considered likely based on other behavior.

   Suggested interest is where Amazon excels—offering connections between things that don’t seem interrelated but that experience has shown are. Customers who bought this book also bought that one. Customers who like popcorn makers also like feather duvets. (I don’t know if that last one’s true in general terms—it’s certainly true for me!)

   This type of interest is of huge importance to recommendation technologists and ecommerce sites. It is one of the ways in which vendors can begin to tap into the buried desires of the public. The market for stuff that people know they want is yay big; the market for stuff that people don’t yet know they want is many times larger. As my design guru friend Dorenda Britten says, you have to be able to look into the future if you want to satisfy your customers.

I’ll discuss discovery, subjectivity and measurement over the rest of the week, but I’d also be delighted to delve further into the human connection: what have we missed? What have we overstated? Do you agree that the human connection is the first principle of search, or do you think the primary focus should be on something else? I’ll be looking forward to hearing from you.

This entry was posted on Sunday, July 29th, 2007 at 9:17 am and is filed under Relevance, Search. You can follow any responses to this entry through the RSS 2.0 feed. You can leave a response, or trackback from your own site.