Crittercism couldn’t have existed 10 years ago.
The startup is on the frontlines of the mobile economy. It analyzes in real time billions of bits every day for the most popular apps around, including those of Pinterest, eBay and Paypal, rooting out when problems happen—or are about to—and why.
Crittercism, through a piece of code embedded in each app it’s tracking, is constantly monitoring and analyzing the breadcrumbs of information spit out by its customers’ apps. In total, Crittercism collects information from apps used by more than one billion people. This means 45,000 user interactions per second, and 4 billion requests each day. The startup tracks crucial mobile metrics, including app crashes, transactions gone awry, and slow response times. By analyzing that data in real time, it helps its customers quickly understand why problems are happening and fix them. ASAP. So mobile customers stay happy — and loyal.
Crittercism simply wouldn't work using a relational database approach. The company’s business is built on analyzing in real time a broad, constantly morphing variety of data from different devices, apps, and operating systems. Doing that in real time and modifying on the fly how it collects and stores that data, all without database downtime, would have been impossible with a relational approach. Just as crucial to its customers is that Crittercism can scale to meet their demands, a proposition that would have been too costly with expensive relational databases and their required, proprietary and expensive hardware.
The Crittercism founders knew that a relational database wouldn't work. They needed a new approach. They found it in MongoDB.
As a key part of today’s mobile ecosystem, Crittercism is an interpreter of problems. This year, it’s expected that more people will access the internet via smartphones than any other device. Consumers depend on apps to run their lives. So companies running those apps depend on Crittercism to ferret out and understand problems, looming or actual, before they cause real damage.
But data on its own doesn’t help solve or stem off a problem. Crittercism doesn’t just monitor apps or gather information. Using MongoDB’s powerful built in query functions, it analyzes avalanches of unstructured and non-uniform data in real time. It recognizes patterns, identifies trends, and diagnoses problems. That means that Cirttercism’s customers immediately understand the root cause of problems and the impact they’re having on business. So they know how to prioritize and correct the problems they’re facing and improve performance.
Crittercism gathers a hugely diverse collection of data from different devices, apps, operating systems, and customers. Relational databases and their rigid structure of columns and rows simply can’t handle this variety of data. MongoDB can. Using MongoDB’s flexible data model, Crittercism can easily add any kind of new data without having to follow strict schemas. And also unlike relational databases, MongoDB lets Crittercism quickly make schema modifications to constantly changing mobile data, without any downtime.
The kind of real time analysis that Crittercism provides customers would also be impossible with traditional databases. Crittercism is using MongoDB’s powerful query functions to analyze the broad variety of data it collects, in real time, within the database. A more traditional data warehouse approach, with ETLs and long loading times, can’t match this type of speed.
At the same time, MongoDB lets Crittercism efficiently handle the tons of data it’s collecting. During the past two years, the number of requests that Crittercism gathers and analyzes has jumped from 700 to 45,000 per second. Relational databases have a hard time scaling to meet these kinds of demands, typically requiring expensive add-on software, or additional layers of proprietary code, to keep up. With MongoDB, horizontal scalability across multiple data centers is a native function.
Finally, MongoDB saved Crittercism a lot of money. Relational databases are expensive, and require expensive, proprietary hardware. In contrast, MongoDB runs commodity hardware.
Crittercism is a creature of the mobile world. It exists to make sense of the sheer breadth of unstructured and real time data that mobile apps spit out, at enormous scale. Relational databases simply can’t do this. MongoDB can.