OPW INTERVIEW – Aug 9 – believe that 'people will tell you everything if you only observe.' So I was quite interested to learn of the work Jonathan Mall was doing in using language patterns to match people. If you'd like to learn more I recommend the seminal work 'The Secret Life of Pronouns.' – Mark Brooks
What's your background and founding story of Mailmindmatch?
While doing my PhD in Psychology, I read the literature demonstrating that language synchronization patterns can be indicative of effective communication, mutual rapport and coordination. Various papers report that people who matched their language style were more likely to start a relationship or have a higher chance to prolong their existing relationship. I immediately wanted to calculate match scores for my own communication partners and found that most recorded conversations accumulated in my Facebook inbox. Previous work, by Pennebaker's group and others (e.g. Ireland et al., 2010; Taylor & Thomas, 2008) >conversation transcripts, thus a thorough investigation and optimization of the language style matching approach was lacking for online communication. Therefore, I build a Facebook app to analyze all my messages and visualize the language style match. The language style match scores obtained for my friends on Facebook largely coincided with my feelings towards them. Ex-girlfriends and even good male friends were on top and people I cared less about were at the bottom. I conducted a large study, asking 300 people to rate up to 20 of their Facebook friends and then calculated which language characteristics would predict how much people valued each other. Over 2.5 million individual messages were analyzed and the predictive power of the resulting algorithms was high. Thus, language style matching in online communication can reliably estimate how effective two people communicate and how much mutual rapport they are feeling.
Now that I've graduated I am looking for partners to apply these algorithms in an environment where a good fit between people is important and has commercial relevance.
How does Mailmindmatch help iDaters make better matches?
iDating has long been trying to create algorithms that match people before they ever made contact. It neglects a large part of the courting process, the actual conversation. Effective communication means listening and reacting "on the same wavelength", which is what mailmindmatch.com measures. While language style matching can also be used to propose prospective dates, based on the content of previous conversations, ongoing conversations offer the most relevant information to determine whether two people match. For example, when an iDater is chatting with 5 prospective dates, the algorithms can indicate which one appears to be the best match.
Where can we see it in action?
The Demo is available at http://www.mailmindmatch.com.
How can an iDating site work with you to integrate the test into their site?
The infrastructure to do the processing of messages is in place. Chat protocols with 300+ words exchanged are enough to calculate language style match scores. The next challenge will be to integrate the technology in the most beneficial manner. While matching based on language style can improve existing matching algorithms that find prospective dates, given the individual nature of interpersonal communication, the algorithms are particularly powerful to evaluate starting conversations.
What would the ideal implementation look like?
An overall match score could be displayed next to a conversation. iDaters can be getting advice on how to communicate more effectively, i.e. achieve higher rapport. At the back-end, language style match scores can be used to operationalize the actual fit of two people once they started chatting, allowing to optimize existing match algorithms without the need of iDaters rating each conversation.
How much would it cost the iDating site to use this test?
The exclusive license costs 80k/year. If IP rights are warranted, other cooperations can be negotiated. Please contact me if you like to know more.
(References:
Ireland, M. E., Slatcher, R. B., Eastwick, P. W., Scissors, L. E., Finkel, E. J., & Pennebaker, J. W. (2010). Language Style Matching Predicts Relationship Initiation and Stability. Psychological Science, 22(1), 39–44.)