If you tell anyone in the LGBTQI+ community that you are into dates, you will probably hear “Are you on Grindr?”.
This geosocial network has become one of the dating apps with the most users worldwide, and a pioneer in its class. By 2022 it already had 11 M monthly users and was listed on Wall Street with a valuation of $2.1 billion.
Devlane has been closely associated with Grindr from 2015, as a key player on a defining era. Between 2015 and 2018, as the app faced increasing challenges, Devlane built a team of 30+ engineers including QA engineers, Mobile developers, Backend engineers, and product managers, among other tech specialists involved in the process.
Proudly building the #1 dating app for the LGBTQI+ Community
Find out how an efficient QA process can be key to the growth of a world-class company.
Not just a client, but a standard for the people
The team we’ve built for Grindr over the years
Tech and Frameworks
Our Laners adapted to the technologies previously used by Grindr, and carried out the process of researching, analyzing, choosing and developing more efficient and less expensive tech stacks.
RoR
Jenkins
Java
TestNG
Groovy
Cucumber
Selenium
TestRail
Appium
JUnit
BlazeMeter
Appium
Jmeter
Datadog
The challenges faced by QA team
At the starting point, Grindr had not yet implemented any QA processes. They were at the time carrying out QA manual and automated client unit tests, but this only accounted for 15% test coverage.
While traffic was growing steadily, the photo moderation process was slow and inefficient, and the problem of spambot attacks was becoming more serious. Fake profiles tried to send ads, obtain sensitive information or convince the users to migrate to the competitors.
When we started working together, Grindr was on its way to stabilize the app with a team of 10 developers. At the time they were working on a rewrite of the backend’s code, but failures and crashes regularly occurred when attempting to deploy it to production.
The backend deployment process used to be traumatic because automated tests were unreliable. So time spent on manual testing was delaying releases, increasing the number of changes in each release.
There was also a structural problem: the code hadn’t aged well, so it was hard to add new features and updates.
Improvements made to their QA process
Frontend/Mobile
We were involved in the development of Grindr 2.0, the new API, and deployments to production.
Development of an automation framework adapted for both Android and iOS.
We used Appium to implement a framework that allowed us to run simultaneous automated tests and optimize the manual ones, saving time and reducing maintenance costs. This framework was integrated with a distributed testing solution (made with Appium Test Distribution)
The mobile automation framework was also integrated with the backend one, allowing us to save a lot of time in operations such as creating users. Instead of doing it manually with the UI, it was done with an automated instruction.
Backend
We built from scratch a solution to the spambot problem. This Intrusion Detection subsystem was able to detect and block users displaying some specified unusual behavioral patterns based on big data.
Creation of a framework using Selenium and Java to automate tests on administration panel and data pipeline.
Migration from performance framework BlazeMeter to a custom solution using JMeter, reducing the maintenance costs.
Creation of a framework with JMeter to create profiles without using the UI. We have managed to simulate 50,000 requests, 1 million active users and 1000 conversations at the same time, as recorded in Datadog – the dashboard we implemented.
QA team built not just individual solutions, but parts of an ecosystem that could complement each other. It should be noted that the implementation would never have been possible without the agile and precise operation that we built between our and Grindr’s engineering team.
How QA led a sea change
Stability
Confidence
Safety
Faster growth
Negative reviews
During those years, we built Grindr to be a much more stable app. At the end of 2017, the uptime surpassed 99.98% for the first time. This means the system was only down for 17 hours in an entire year, instead of the initial record of 174 hours.
Having built a reliable automated testing suite allowed weekly releases without risk, and in some components we achieved continuous deployments with automatic and scheduled releases without human intervention.
Market Cap Value up to 158%
(From 155M to 400M)
Daily Active Users Growth by 33%
(3M to 4M)
By 2017, Grindr surpassed Meta in time spent on the app.
Our Engineers
Efficient time management means a more profitable investment.
The QA team understood this deeply, and their decisions reduced maintenance costs and testing time from hours to just minutes.
