Monte Carlo wants to do for data what application performance management did for enterprise software uptime.
The startup launched the Monte Carlo Data Observability Platform, which aims prevent bad data pipelines. Monte Carlo CEO Barr Moses likened data as the new software for companies. “What New Relic does for microservices, Monte Carlo will do for data,” she said. “Data is everything to strategic decisions. It’s the new software.”
Of course, there’s a catch. If the data is wrong and can’t be trusted, enterprises face a vicious flywheel of bad decisions based on gut feel. Customers are lost and CXOs can lose their jobs. “The worst thing to happen is if the data is wrong or can’t be trusted,” said Moses.
Monte Carlo’s platform uses machine learning to enable analytics team to resolve data issues faster. Today, data flows often break and the data science team is often the last to know. With an approach that allows for quick implementations, Monte Carlo aims to show value quickly by solving for the dirty data issue.
By using a data observability approach, Monte Carlo is preaching a concept technology leaders will get quickly. After all, technology buyers know DataDog, AppDynamics, New Relic and the application performance monitoring (APM) space well. The return on investment case is straightforward: Companies spend a ton of time trying to clean up bad data and problems compound along with exponential growth and complexity.
Here’s the problem in a nutshell:
How Monte Carlo Data Observability Platform works
Monte Carlo’s platform uses machine learning to infer and learn a company’s data by using APIs to tap into data warehouses, data lakes, data operations and business intelligence. Monte Carlo’s Data Observability Platform works with on-premise and cloud services including Snowflake, Google Cloud BigQuery and Amazon Redshift to name a few.
Once the platform is hooked up via no-code onboarding, it can identify data issues, assess impact and send alerts. Teams can then find root causes faster. “Teams are spending too much time on data fire drills and data downtime,” said Moses.
Moses noted that the Monte Carlo platform only accesses data on a read-only basis and extracts metadata and logs. The platform doesn’t extract data and monitors it at rest.
Machine learning is used to understand the data architecture and what smooth operations look like based on five core elements:
- Freshness and recency.
- Volume.
- Schema.
- Distribution including things like fields, duplicates and null values.
- And lineage to understand how data connects upstream and downstream.
Customers don’t have to configure or set thresholds.
Monte Carlo, which raised $16 million in venture funding in September, charges subscriptions on a software as a service basis, but uses a virtual private cloud architecture to run locally in a hybrid model. The company never stores or processes customer data.
Moses added that it takes about a week to learn the data environment and benchmark. Monte Carlo’s platform is available.
