Detecting Data Anomalies with Anodot

Prevent an incident from becoming a crisis

Let's talk

Our focus is using data and analytics to help businesses make smart, data-driven decisions. Because we recognise the importance of being able to make decisions in real time, one of the technologies we are expert in is Anodot.

Anodot relies on AI to detect real time data anomalies that might otherwise go unnoticed. With its ability to detect data irregularities as they happen, we implement it as a tool to empower our clients to react faster to either solve problems or, even better, exploit new opportunities.

The importance of anomaly detection

We now have more data at our disposal than ever before. With so many metrics available, companies can develop an impressive dataset to measure performance. Within this dataset there are patterns that basically represent “business as usual”. An unexpected change within these data patterns, or an event that does not conform to the expected data pattern, is considered an anomaly. A deviation can be a good thing, or a bad thing. But it is definitely a ‘thing’ worth investigating.

When something changes in the data stream, it often indicates a business disturbance. Real-time alerts are necessary to prevent negative impact on customer satisfaction/trust, revenue and most importantly your brand.

There are millions of data permutations, but only so much you can see on any given dashboard. Unsupervised anomaly detection can throw light on business blindspots, and generate real alerts for real issues.

‘The price of light is less than the cost of darkness” – Arthur Nielsen

Anomaly examples:

A sales boom for an ecommerce site on Singles’ Day is not an everyday occurrence, but a sales peak every November 11 would be a well-established peak for companies active in Asia. It would be business as usual. An anomaly would be recording slow sales on Singles’ Day, especially if there had been strong sales the year before. It would mean something different was happening. A break in the pattern. Something going wrong in need of urgent attention. Maybe your payment gateway has crashed? A price glitch? Whatever it is, it would be an issue that you would want to know about immediately so that you could fix it and prevent a problem from becoming a crisis.

An anomaly can also be premeditated. A boost in marketing spend for example, or the launch of a new and improved model. An increase in sales above the regular pattern could be an anomaly indicating that a new campaign is making a positive impact on sales. In this instance, anomaly detection can be used to measure the effectiveness of a new business initiative and help you better manage your resources.

In many real-life scenarios, business anomalies often require immediate action. If something has gone wrong with your ecommerce site on Singles’ Day, you need to fix it immediately, otherwise sales will be lost for every second an issue remains unsolved. Or if sales start to spike because Selena Gomez mentioned your brand on Instagram, you are going to want to know about it as soon as it happens so you can start converting likes and shares into sales.

You don’t know what you don’t know until you know it.

How Anodot works

Anodot offers automated anomaly detection that can detect real-time anomalies at large scale. It has the capability to handle millions of data points, and updates every second. It is powered by patented machine learning algorithms using the latest in AI research.

Anodot’s business incident detection solution automatically learns your streaming data’s normal behaviour, even seasonal or other complex patterns, and then alerts on any metric or combination of metrics that behave abnormally. This gives you the tools you need to detect and diagnose issues early, resolve them quickly, and take preemptive actions.

The advanced, robust algorithms can handle any number of data variables, intelligently correlating related alerts to avoid alert storms and enable faster root cause analysis. Anomalies correlated across multiple data sources tell a clear story of what is happening with the data, which can then either be investigated by human experts, or used as triggers for automated business actions.

By identifying the root cause of a problem you can immediately zero in on the issue. The end result is a powerful analytics solution based on AI, which enables us to help our clients react faster to solve problems or, even better, capitalise on new opportunities.

Let’s talk.

Sydney: +61 2 9299 4430

Melbourne: +61 3 8605 4880

  • This field is for validation purposes and should be left unchanged.


Level 4

350 Kent Street,

Sydney, NSW, 2000

+61 2 9299 4430


Level 8,

350 Collins Street,

Melbourne, VIC, 3000

+61 3 8605 4880