With AI analytics slated as the biggest disruptor to big data and analytics, data leaders are quickly integrating this capability into their data strategy.
Itzik Feldman, data engineering manager at Atlassian, the enterprise software company responsible for Jira and Trello, recently credited Anodot with helping keep the company’s 3,000 employees in touch with product performance and customer experience.
“I like to think of [Anodot] as our safety net. If everything else failed, if quality framework didn’t work, if there was no notification made, the anomaly detection system should catch all problems,” said Feldman of the AI analytics platform, while addressing an audience at the YOW! data tech conference in Sydney. With comprehensive and real-time coverage, Atlassian is using Anodot to focus on reaching new users with minimal effort.
“You just stream all your data into it and if there is some anomaly in the behavior of the data, it will alert you, will tell you there’s a problem, it will send a notification wherever it’s needed. It will also correlate between different data sets as well,” according to Feldman.
Anodot Autonomous Business Monitoring, which works across silos to correlated related events and KPIs, is essential for large, international companies like Atlassian, whose operational workflow is highly complex and involves multiple solutions. The company uses Parquet to store files, Spark and Hive technologies for data ingestion and preparation, Presto and Athena to query, and Tableau and Amplitude for visualization. Atlassian manages approximately two petabytes of data and two billion events daily. Anodot works across these data stacks to create a bird’s-eye perspective of the data, so that product managers can be sure that the products are working exactly as promised.
Anodot is “doing a lot of processing behind the scenes. It really increases the trust of your users,” Feldman said.
Manual monitoring holds data teams back
Many businesses as they scale struggle with managing enormous data sets, which can remain siloed and underutilized. Data leaders are coming to understand that there’s really no alternative to AI analytics for extracting value from metrics that often reach into the billions.
When Atlassian had only a few hundred employees in 2012, they appointed a team of analysts to generate insights. Once they became a corporation with employees across the world, there was simply too much data to monitor manually.
The project managers tracking the business funnel, conversion rates and other key KPIs simply couldn’t keep up. No matter how many more analysts they hired, it was never enough. There were too many dashboards to monitor manually. And attempts to “automate” with static thresholds triggered alert storms that made it easy to lose sight of the critical anomalies.
Incident management became reactive and, ultimately, untenable. Glitches would easily escalate into crises, customers got angry and revenue quickly bled out before teams even realized there were problems.
Atlassian saw the potential for proactive monitoring in AI analytics, and Anodot Autonomous Analytics became their platform of choice. Along with companies such as Wix, Overstock, Foursquare and LivePerson, Atlassian began using Anodot’s machine learning algorithms to continuously monitor millions of metrics and adapt to seasonality and influencing events and factors.
With Anodot’s real-time alerts, Atlassian was able to quickly zero in on glitches and avoid costly incidents.
Anodot monitors at scale and in real time
AI analytics solutions like Anodot go beyond the capacity of any one group of analysts. Millions of metrics are continuously monitored, at the most granular level, so that no area of the business goes unnoticed. This coverage enables big data companies to address all questions and issues stemming from their user base.
Anodot’s anomaly detection surfaces product quality issues. such as price glitches, quickly, far before customers are impacted. Its machine learning models work autonomously and continuously across all data, an incomparable alternative to the time-consuming, dashboard-based manual tools currently marketed for KPI analysis. Anodot effectively measures all aspects of business activity, including operational performance of applications and infrastructure components, to give a thorough understanding of root cause.
Anodot helps build trust in analytics and refocuses teams on mission-critical tasks
When an anomaly is identified or when a static threshold has been breached, Anodot’s anomaly detection alerts the appropriate owner, whether that be the engineer, PM, QA or all of the above. This effectively saves product managers countless hours of unnecessary monitoring, so that they can instead focus on moving the product forward in other ways. Instead of being flooded with messages, only highly accurate, critical and actionable alerts are sent to the PM’s workflow.
Altassian can not only keep tabs on technical glitches and incidents of fraud, but also understand when opportunities arise, such as using changes in consumer behavior or potential up-sells. They are smarter about the way they interact with users. Perhaps most importantly, said Feldman, “if there is a problem you will get alerted and people will start trusting the data more.”