Filter vs. Matching
For those of you who know what Red Radius is about (hopefully there’s more that do than don’t), it’ll be clear to you that the core value of our business is providing the right matches to attendees. A conference attendee can only get value out of the platform if and only if a good number of connections recommended to him or her are relevant to them in one way or another. In this regard, the team at Red Radius has been asking one question: What is the difference between five great matches from five random matches?
In order to answer this question, we first need to decipher between filtering algorithms and matching algorithms. There are a lot of software businesses out there that claim a great “matching” system but instead use a filtering system. Let’s quickly define these two. Filter Algorithm: Based on certain input criteria, a filter algorithm matches a user with the optimal choice from a group of options. Here’s an example of a filtered algorithm. Based on projects completed in the past, and other search criteria, a software recommends employees that should staff an upcoming similar project. Essentially comparing apples to apples. According to Dr. Tony Karrer (former CTO of eHarmony), there’s nothing wrong with a filteration algorithm, however using a filtration algorithm is not proprietary and can be replicated by anyone.”If you are going to be building the best matching algorithm for high school students looking to find the right college, but it is based on criteria that are a search (geographical location, majors offered), no one is going to ascribe greater value. You may have a perfectly fine business, but it’s not going to be differentiated based on that simple algorithm”.
Matching Algorithm: Based on industry specific data, a matching algorithm finds other users or services that will compliment and add value to those of the user. This is not just comparing apples to apples. Ideally, what we want for Red Radius is a matching algorithm that bases all matches on industry specific research and data. Providing recommendations that are not only relevant to a user professionally but recommendations that are also based on a user’s “personality profile” as we term it. Let’s give a quick example. Say that you’re looking to buy some type of aroma for your home or office. This website, allows you to create your own “aroma personality profile”. Based on your choices such as Ikea over Crate and Barrel, or an antique couch vs. a modern couch, their algorithm can make some specific assumptions about the user and their preference on aromas. Now this is not random. What they’ve found is a correlation between people that like Ikea, antique furniture etc. and what aromas these type of people like. All based on industry specific research and data. Whether it’s matching an unemployed individual with a firm based on company culture or finding the right date for a single on Match.com, finding a service or individual that compliments a user’s attributes are issues that need industry specific data and research.
This is what the team at Red Radius has been working on in the past weeks. Collecting some industry related research and discovering correlation on which our matches will be based on. In short, we’re using both a Filter algorithm and matching algorithm. We first find all attendees that you should ideally meet at a conference based on conference objective and who you would like to meet. However we then go through these first tier of matches and sort through to find a second tier of matches that are based on our industry specific research algorithm.
