My adventures at the dating sensation.
IT Specialist and Ops Engineer, stop-gap solution with Data Science and Product Design teams
Jira, Confluence, Interana, Sketch
March 2015 - October 2015
I started with the Ops department, handling local infrastructure and writing documentation for internal processes, and eventually started assisting Ops with systems monitoring of the production application. During this time I approached the Data Science and Product Design departments and asked to get involved there in order to learn more and develop my skills, and was allowed to help with demographic research and the Super Like implementation.
I joined Tinder as employee #37, hot off the trail of Tinder Plus, and they were growing rapidly. As such, they were looking to institute internal processes such as ticketing systems for local support (e.g. a broken laptop) and documentation on how to handle broken laptops, onboarding of new employees, provisioning of new machines, and other IT services. I was brought on as a contractor in order to assist in this front.
Additionally, Tinder was in the beginning stages of designing and implementing Super Like and growing their data science operations, so I used this as an opportunity to engage with new skills and practices.
For the process around IT, we wanted to stick wIth an existing system unless absolutely necessary. Considering Tinder already had the Atlassian Suite for development and DevOps, this meant there was not only an effective ticketing and documentation system setup, but also a system that people were already familiar with, so this made it an easy choice.
For the documentation, we setup an IT space within Confluence and Jira to isolate its tasks from the rest of the ecosystem and got to work. We started by determining all the different documentation needed and assigned each of those a category. The categories were used to create a table of contents to easily organize all the documentation into easy to find places.
We took the time to make sure each process was well documented, taking you not only from start to finish, but if there was any additional knowledge needed beyond what was assumed to be standard background knowledge, we included links to external materials to cover those concepts. Additionally, if a process relied on knowing how another process works, they were linked together with not only a related field, but also linked within the relevant sections of the documentation.
For ticketing, we set up JIRA to have an IT type of ticket, and edited the type field for “hardware”, “software”, “accessory”, and “requires monitor” so we could act accordingly and see the problem at an immediate glance.
We wanted a way for users to boost their chances with potential matches they took a strong liking to. The name “Super Like” came easy as they wanted a way of saying “Hey, I really really like you!”. There was a lot of discussion around the best way to implement this and it was decided a few things would present this:
Because Tinder used a gesture system already (Swipe Right for liking, Swipe Left for disliking), a gesture for this action was important. We decided to stick with an easy cardinal direction by swiping up, as swiping down could potentially interfere with the buttons at the bottom and slightly resembled throwing something away.
Since super liking someone gave you priority over others, a star icon was chosen: since stars not only are a common icon for things that are important, but also because you were trying to make yourself seem like a star over others.
Super-like is currently in use in the product today, still using all of the research and design decisions described here
Tinder was looking into gaining a better understanding of their demographics, usage patterns, and other data points as well as implement a ranking system to optimize matches. Having a background in doing data for political campaigns, this was an area I had familiarity with, but never in such a large-scale capacity. I assisted in determining what sets of data could provide insights into usage patterns and other selected items. I also assisted in running analytics and analyzing this data via Interana.
While I had no direct involvement in optimizing rankings, I did sit in numerous meetings discussing how this would be accomplished, learning a lot about how ELO ranking and other rankings methods work, alongside various methods of implementation.