Clarification & Understanding

What is our rationale for doing this project? Is there some sort of strategic imperative we need to align with or a particular objective we should focus on? Let’s just focus on providing a really good experience to our users and helping them find a compatible roommate.

Is it safe to assume that we should worry solely about the act of finding a suitable roommate and not all of the logistics involved with the actual moving process? Yes.

I’m assuming this won’t be restricted to an individual country.

 

Users

It’s hard to find a magic bullet that solves everything for everyone, so let’s narrow our scope by talking about the different users and choose one to focus on:

1. Young professionals: just got a job in a new city, probably a recent college grad
2. Fresh starts: Want to try something new and different from their current situation, maybe a recent divorcee
3. International Explorers: adventurous and want to experience different cultures and ways of living

Out of the above user groups, I’m going to suggest we focus on the young professionals. For me this is the most personally relatable, but more importantly this is the largest of the three groups so by focusing on it we’ll be able to provide a better experience to more people.

 

Pain Points / Needs:

1. It’s tough to figure out which section of the new city you want to live in
2. You don’t know what your potential new roommate is like personality and interests wise and whether or not you two will get a long. Are we going to be friends or is this person a psycho?
3. It’s hard to know what the other person’s standards / habits are when it comes to the household. For example, the definition of clean can mean two very different things to two different people.

Solutions

I think at the crux of it, this is really a matchmaking problem. We’re trying to find 2 or more people who are the best fit for each other. Two other very common match making problems are dating and job recruiting so I think it would be a smart idea to draw inspiration from some of the platforms that address those needs. Here are some following solutions:

1. Tinder for roommates –  Fill out a profile, upload some pictures and swipe left or right on potential roommates based upon who you think you could be compatible with. Instead of describing what you’re looking for in a significant other you would describe what you’re looking for in a roommate. When matched a chat dialogue would open up between the two users.
2. Background Checker – Users could ask us to perform a roommate specific background check on any perspective roommates. Instead of professional references, users could supply personal references or past roommates they’ve lived with and we would send them a survey or call them up with questions about how the individual was as a roommate. Furthermore, we could and should conduct a standard criminal background check. You don’t wanter to enter a 12 month legal contract to live with an individual and then find out that they are on parole for attempted murder 0_0.
3. Compatability Score –  We could have users fill out surveys about their living habits and interests then create a new roommate compatability metric from the results.  To make this more time efficient we could ingest user data from other platforms. For example, seeing that both users subscribe to the same video game subreddit would increase their compatability score. We could even display some of he common interests the two prospective roommates have in common.

 

Prioritization

I’m assuming we don’t have the resources to build out all three solutions in parallel so let’s choose one to prioritize. Remember our goal here is to provide a really good user experience and actually help people moving to a new city find a roommate so in order to help guide our thinking I’m going to use the following comparison matrix:

Ease of Implementation, User Satisfaction

1. B-, B -> While this does a good job of addressing pain points #2 and #3, I think the downside here is that it’s going to be laborious to find and make a match. I mean there are a lot more people in cities open to dating than finding roommates and getting matches on Tinder is hard enough already. While useful this places too much of a burden on the user.

2. B+, C+ -> I think ensuring your roommate isn’t a serial killer is definitely a good thing, but this is really addressing an edge case. Checking out references is a good idea on paper, but realistically no one would provide the contact details for anyone who was going to say anything bad about them.

3. B-, A -> I like this solution a lot. It isn’t difficult to answer a quick survey about your living habits and to grant the Compatability Score algorithm access to your data on other platforms would only be a few clicks of a button. While like most algorithms it will by no means be perfect, it will still definitely save users time by knowing which prospective roommates to focus on vetting and which to ignore.

Based upon the discussion above, I am going to suggest we prioritize building out solution #3 the Compatability Score based upon its blend of usefulness and practicality.

 

Compatability Score Details

I’m viewing this compatability score as being an additional feature and less of a standalone product. One of the existing ways people find a roommate in a new city is by posting on different housing boards like craiglists or FB marketplace or groups. People are generally descriptive about what they are looking for in their post, but the issue with this category of solution is that it is very time consuming to make your way through all of the posts. Some of these platforms offer the ability to filter and sory by things like price and location but that doesn’t really get to the crux of roommate compatibility, which is where our compatability score would come in.

Our Compatability Score would assess how compatabile of roommates the user would be with the poster looking for a roommate. When looking at roommate postings, users could sort by or filter by our Compatability Score. When clicking into a post, users could see what common interests and living habits they have with the poster as well as their incompatible areas as well. For example, Jane likes to go to bed early but you like to stay up late.

One of the challenges with an amorphous algorithmic score like this is that you need to get users to trust and believe in your algorithm otherwise they won’t use it and get any value out of it. I think by showing specific common interests or compatable areas we will help elucidate some of the underpinnings of the algorithm to the users and help build their trust in the feature.

 

Summary

To recap, we are going to provide a really good user experience for young professionals trying to find roommates in a new city by creating an algorithm that will give them a compatability score with others looking for roommates based upon their survey answers and ingested data from other platforms. This compatability score will help users get a better sense for who their potential roommate is both in terms of living habits and interests.

 

Metrics

Ultimately we wanted to provide a really good experience to our users so in this context I think that means we’re trying to figure out if our Compatability Score feature actaully helped users find roommates. I think a good way of examning this would be looking at how much users engage with the feature. Still envisioning our Compatability Score as part of a FB marketplace or craigslist like platform, we could measure how often users Sort by or Filter by our Compatability Score.

It is going to be hard to determine if two people end up living together, but as a weak proxxy we could determine whether or not a user sorting or filtering by our compatability score opens up a dialogue with a potential roommate. A dialogue could be defined as at least 5 messages exchanged between users.