Matching vs Recommendations

I spoke to someone who described themselves as a CFO-for-hire a while back. After hearing my description of Trybe’s mission, she commented that “there are a ton of recommendation sites out there.” I couldn’t deny that but felt that this didn’t adequately describe our concept. It inspired some thinking about how we are different. I looked up “match” and “recommend” in Webster’s but found the definitions to be somewhat circular. To recommend is to “suggest.” Duh! A match is a good or suitable fit.

But, as I pondered this, it occurred that there is a difference between the two and, of course, I bring it up because it illustrates why Trybe’s solution is better than the alternatives. Our model is based on matching technology.

Let’s take recommendations first. To recommend something is an action. You suggest something to someone, presumably based on some measure of its suitability or relevance. But your assumptions about their tastes or preferences could be flawed. I find this with Amazon. I’ve bought hundreds of books and CDs from Amazon and occasionally their recommendations for a musical artist or author are useful. But, in general, I don’t look to them for recommendations and generally ignore their emails. Why? Because I feel they don’t know me. This was the case even before last summer when I bought a power saw through them, (because of the convenience of OneClick). Very soon, I began getting recommendations like: “Special on Power Tools this weekend.” And “20% off Work Boots.” I kid you not. They were inferring from one purchase an interest in or need for power tools that didn’t exist. They don’t know me. They continue to recommend power tools to me to this day. A+ for persistence.

What about matching? A match is a good fit between two entities (sometimes potential spouses). Understanding what makes a good match implies a knowledge of BOTH entities. Amazon would stand a better chance of matching me with music and books I might like if they knew me better. Of course, even then, matches or recommendations can seem off base. No one can predict with 100% accuracy whether someone else will like something or not.

Trybe’s strategy is to know you by…wait, we’re still in stealth mode. Suffice it to say, we’re going to need some information about you in order to fulfill the promise of offering relevance. One of the challenges is how to glean that information while keeping you engaged. There has to be a better way than subjecting people to something like eHarmony’s endless series of questions. We’re not the only ones going down this path. And clearly doing it well is a real challenge. Because it hasn’t been done yet. Yet…