Sisu Data adds predictive analytics to new offering

Sisu Data adds predictive analytics to new offeringSkip to main content

San Francisco-based Sisu Data, a company empowering enterprises with actionable insights on data, today announced an update introducing predictive analytics capabilities to its decision intelligence engine.

While most business intelligence tools focus just on the ‘what’ side of things and visualize events, Sisu has been working on both ‘what’ and ‘why’ aspects. The company uses its AI-driven decision intelligence engine to detail not only metric changes, but also the key drivers behind them. Now, with the addition of predictive analytics, it is moving to answer the next major question enterprises face – what to do next?

How will this work for Sisu Data users?

According to the company, the decision intelligence engine will be able to use machine learning (ML) to predict the future impact of a metric change or future outcomes for specific entities. It will then visualize this data through smart charts, ranked results and crisp dashboards, helping teams confidently plan their future actions regarding the entities in question.

For instance, the engine could leverage data to highlight customers who have not churned but are likely to churn in the future, giving teams a direction to work in.

“Sisu’s analysis provides a predictive value (expressed as a percentage) for each entity in the data, such as customer or transaction. This value is a measure of how likely the characteristics of that entity indicate it should belong to a specific subgroup (churn or fraud),” said Peter Bailis, the founder and CEO of Sisu Data. “Sisu customers can quickly identify entities with a high predictive index as likely to be part of a defined category (churned customers, fraudulent transactions, etc.) and take immediate action based on the key drivers of the prediction.”

The results are returned as a table with the entity as the primary key and a dimension that shows the predictive value, he explained.

Trend and anomaly detection

In addition to predictive capabilities, the update also brings the ability to detect trends and anomalies in data over time. As part of this, the decision intelligence engine will automatically track metrics and visualize not only trends and inflection points between trends, but also a range of expected values for a given metric and any data points that fall outside of that range.

“Unlike conventional trend and anomaly detection systems that focus on one variable at a time, Sisu’s engine applies analyses across many cuts of the data simultaneously – looking at tens of thousands of combinations of relevant variables at once,” Bailis noted. “This means that a user can monitor the joint impact of variables such as customer segment, product, and marketing channel all at once – and be notified when a significant change occurs in any one (or a combination) of these variables. Analysts and viewers automatically know if a change is statistically significant and if action is required.”

With this capability, users can even conduct first, second and third order analyses of the drivers behind the trends to learn more.

Availability

The trend and anomaly detection capabilities are now available to all Sisu Data customers, while predictive analytics is rolling out only to select customers as part of a closed beta. The company did not share when exactly the capability would be widely available.

Notably, this update also brings a few key integrations, including one that allows customers to be notified via email or Slack when a metric falls outside pre-defined parameters. Others enable connection to platforms such as Looker, Tableau and Embedly.

The features come nearly a year after Sisu Data’s series C funding of $62 million. Overall, the company has raised close to $130 million with the backing of Green Bay Ventures, New Enterprise Associates (NEA) and Andreessen Horowitz. Other players in the decision intelligence space are Tellius, Domo, Cloverpop and Diwo.

According to Gartner, by 2023, more than 33% of large organizations will have analysts practicing decision intelligence, including decision modeling.