The Timeflow Platform is our pre-configured and fully managed solution for event stream processing.
Though the data and analytics market is crowded, we do believe that in many ways the platform is a new class of analytics product which stands apart from other approaches.
As this is a very common discussion point for us, we wanted to summarise our core value propositions here, and describe why they are not currently well served by other analytics tools on the market.
1) Improve Time To Resolution
Timeflow is useful in any business situation that benefits from early identification. This could be a “real time” aim, where we need to identify and act instantly, or something on a slightly longer time horizon such as identifying something in hours which could have previously taken days.
Most analytics platforms are all about integrating data, and then surfacing it on a dashboard for review by a human, potentially days later. The focus is not really on speed of identification.
Companies are trying to tackle this and move towards more real time analytics, but they are often restricted by their data architecture and the design of products on the market which are often based around batch data ingestion and then giving humans visualisation tools.
Timeflow on the other hand is all about processing data in flight in order to reduce the time to identification.
2) Automating Resolution
Timeflow is useful where we not only want to identify a situation early, but where we then want to have some change automatically actioned when the situation occurs.
This change could be something as simple as informing a human via some alert process, through to executing an API call to change how an IOT device is operating in real time or how a website is presented.
Most analytics tools are powerful with regards to understanding and analysing data in a database, but they are not setup to trigger events, workflow and API calls when an identified situation occurs in an event stream. At best, they might provide some alerting but generally nothing more sophisticated than that.
3) Intelligent Processing Of Events
Timeflow is useful where you need to process your incoming event streams in an intelligent way in order to identify patterns and situations of interest.
This analysis could be based on simple business rules (e.g. an order that has shipped late), or more sophisticated statistical or machine learning models (e.g. an order more than 25% likely to ship late).
Other analytics tools will provide simple searching and reporting capabilities over captured data, but will not have stream processing or signal processing capabilities which can be used for processing events in flight. The products which are moving in this direction, are generally very IT and security focussed.
4) Embedded Analytics
Timeflow is useful where you want to move beyond reports and dashboards, and have your analytics and insights integrated into other endpoints such as applications, websites, devices. The aim of this is to get the insights into your employees and customers hands so that they can inform their next best action.
For further information on this idea, please visit this blog entry.
Other analytics tools main outputs are reports and dashboards, and opportunities to get at the underlying data for integration purposes are limited.
5) Citizen Developers
Timeflow is very focussed on making available powerful event stream processing capabilities to citizen developers or data analysts.
Previously, writing real time stream processing processors has only been possible for developers who can work with fairly complex frameworks. Other analytics and reporting tools simply do not have the concept of real time stream processing.
The differentiators above are sometimes subtle to understand and decide how to apply, but we do believe that they open up incredibly powerful tools by which companies can improve their customer experience and interal efficiency. We would welcome a discussion with anyone who is challenged with their current analytics solutions to see if real time event processing could add value.