SeeWhy is the first real time Business Intelligence platform for the event driven enterprise. SeeWhy continuously analyzes and interprets streams of individual business events, to alert you immediately to opportunities and risks and enable everyday decisions to be automated. Scalable for the most demanding applications, SeeWhy puts ‘instant insight‘ into your business processes, giving you real time visibility into current and future performance.

At a very high level, the system will enable you to generate:

• Real time metrics
• Real time alerts
• Real time actions

All driven by the business events streaming through your network, middleware or service oriented architecture (SOA).

By analysing this data as it happens, the system continuously monitors the business without requiring a data warehouse, or the running of database queries. Whenever it is needed, by an application or by a person, the data is always up to date.

This intelligence can be built into business processes, or can alert operations staff automatically to changing business trends, providing proactive notification of critical operational conditions that require action. This enables processes to be tuned, customers contacted, risks averted and opportunities taken which otherwise would have been missed.

The SeeWhy Adaptive Intepretation Cycle

The SeeWhysystem uses an adaptive approach to interpreting real time data. What this means in practice is that the system automatically adapts the rules where appropriate to take account of changing behaviour.

This is a significant innovation compared with other approaches: traditional business intelligence or rules based approaches require IT to set the level of rules and maintain them over time. The problem is that in order to maintain the rules effectively, you need a very good understanding of the data, and how the business is changing at a microscopic level of detail. Consequently these rules generally do not get maintained adequately, and performance of the system gradually deteriorates over time.

Rules that adapt automatically

The solution to this is to use an intelligent system which recognises that the business is changing, and adapts the rules automatically to suit. For example, after many months of relatively static equities volumes on the London and New York stock markets, in October 2003, volumes rose significantly reflecting a more bullish environment. Most simple rules systems would not have recognised this up-tick as a change in the normal trend, and consequently the thresholds would be out of date rapidly.

SeeWhy adapts to changes in the baseline, as thresholds will automatically adjust, so that the monitoring rules keep pace with changing business conditions.

SeeWhy - The first of a new generation of BI tools

The main differences between SeeWhy and traditional query based Business Intelligence products lies in its event driven architecture. Products of this type are sometimes referred to as BI 2.0, reflecting that they represent the next generation of Business Intelligence. The main characteristics of these product generations are summarised in the table below below:

BI 2.0 Evolution

The SeeWhy Implementation Cycle

The SeeWhy Interpretation Cycle is also mirrored in the typical process involved in applying SeeWhy to a specific business problem. Once the technical installation of the system is complete, the usual first step is to prime the system with historical business events. These can be used as the basis for developing the calculations and rules that will be used to monitor future events as they happen. Operating the system in ‘replay’ mode then allows the application of these calculations and rules to the historical events as if they had been monitored in real time, including the creation of alerts and their associated notifications.

‘replay’ mode then allows the application of these calculations and rules to the historical events as if they had been monitored in real time, including the creation of alerts and their associated notifications.

Metrics tuned for best results

This process provides the baseline for determining what is normal, and also allows the calculations, rules and notifications to be tuned before they are applied to new events in a production environment. This involves representatives of the business community, who will be the ultimate consumers of the information generated, reviewing the outputs from the system and the impact this will have on business process. Several tuning cycles are likely to be required before the solution and its associated business process is ready for release to a wider audience.

An Example Application

Here is a simple, but real, example of how SeeWhy might be applied. The marketing manager responsible for an eCommerce website needs to use dynamic pricing, automatically adjusting the price of products as demand changes, in the same way that airlines set seat pricing. To tackle this problem you would set up a metric in SeeWhy to monitor in real time the sales volume for each product being sold. By comparing each product’s sales volume to what would normally be sold at this point in time, very subtle rates of change can be detected.

Analyses every business event

In operation, events stream into the SeeWhy system for every product as sales are made. As each of these events is received the system calculates the number of sales in the last hour and compares this with the average sales of this product on this day of the week, and hour of the day, over the last 12 weeks. For example, if our event was received at 10:45 on Tuesday, the system would calculate the total sales for this product since 9:45 and compare that to the average number of sales for this hour over the last 12 Tuesdays.

Once it is detected that a certain product is selling significantly faster than normal then a notification would be generated and sent to the pricing manager suggesting a price adjustment. Equally, the price change action could be automated to increase the price of the product by, say, 10% until additional stocks of the product arrive.

When instead the system detects a product that is not selling, price adjustments in the opposite direction could be initiated. Note that in a physical store environment this later situation could also be indicative of an availability problem, for example where the product is not on the shelf, but there is stock in the back room.

High volume, highly scalable

This type of scenario is straightforward to set up in SeeWhy, even if you have thousands of products to track, and you’re making a million sales per day. Doing this type of real time tracking in a database is difficult due to the nature of the comparisons of real time sales volume to what is expected, yet it is this comparison that gives you the visibility of changing trends, which in turn enables you to adjust the price with confidence.

The details of the calculations and comparisons described in our example are fully definable and easily tailored to specific requirements. The richness of the SeeWhy product also makes it equally applicable to many more varied and complex monitoring requirements.