Analyze

Products

Solutions

Services

Partners

Resources

Company

Contact

Sitemap

Home

 

 

 

Models are the core of any analytical system, and Performance Chain uses the most sophisticated models available.

 

Key performance indicators (KPIs) are time series. "What gets measured gets managed," so to manage key performance indicators you need to analyze time series. But time series analysis is complicated. That's why until now no vendor "offers significant ability to model, analyze, and forecast time series." (Seth Grimes, Intelligent Enterprise magazine, August 10, 2003, page 17)

 

Performance Chain, for the first time ever, brings sophisticated time series analysis to the desktop of everyday managers. Its unique models and user interface allow managers to understand past and current performance and to identify performance gaps. It provides all the information needed to form realistic, meaningful plans.

 

Performance Chain models both individual account behavior and how different accounts are related, for both financial and operational measures. It uses advanced statistical techniques to isolate key performance indicators by identifying both the effect of natural random variation and the impact of unusual events and exceptions. The result is in-depth knowledge of the current state of your business:

 

         Current trends, relationships, and run rates

         How trends, relationships, and run rates have changed over time in the past

         Effect of past initiatives and other actions

         Benchmarks, standards, and best practices

 

The goal of modeling is to separate the systematic, predictable, and controllable effects from random variation. There are two major techniques for modeling the systematic effects:

         Decomposition, and

         Regression

 

Both provide powerful insight into performance.

 

 

Decomposition Model

 

Commonly thought of as "deseasonalization," this technique breaks down, or "decomposes," complex time series into the individual factors that affect it: trend, seasonality, events, and random variation.

 

 

TREND is the slowly-changing movement, up or down, of a time series over time. This movement is usually due to factors such as population change and economic cycles. Some of these factors are under the control of or can be influenced by managers.

 

 

SEASONALITY is the pattern of variation during a year that repeats each year.

 

 

EVENTS are unusual or non-repeating occurrences that have a significant effect. Events can be planned or controlled, such are marketing promotions, or unplanned, such as power failures or floods.

 

 

RANDOM VARIATION is the unpredictable variation in each time period. This variation does not have a systematic cause. It can be positive or negative in any one period, but over time it averages zero.

 

 

Isolating each of these factors enables us to describe these complex data series in simple terms. We learn how and why the trend is changing, what impact events have on performance, and how much the data varies from time period to time period due to random effects. This analysis makes it simple to plan and evaluate our future performance.

 

Performance Chain offers unique decomposition modeling features, with robust time period configuration and advanced event analysis.

 

Regression Model

 

This technique models cause and effect relationships between accounts. It is used when one account is directly influenced by another account (the "driver"). We can isolate the effect of the driver account, to understand the relationship and how it has changed over time. This technique also tells us how much the data varies from time period to time period due to random effects.

Regression analysis isolates "fixed" values (the amount that is not related to the driver) and provides a true "marginal" rate of change - how much the account will change given a change in the driver. Both of these are key performance indicators and are available only by using this type of analysis.

 

Once we know the relationship, we can plan and evaluate the relationship instead of the complex time series. We have a meaningful, unambiguous, and understandable measure of performance.

 

Performance Chain offers unique regression modeling features, with interactive relationship modeling and user-controlled event analysis.

 

 

This combination of modeling ability is unparalleled in the Performance Management and Business Intelligence industry. No other vendor can match this analytical functionality. Using the wrong model to analyze an account's behavior will produce misleading and incorrect results. Performance Chain shows you the right model to use and ensures you will get accurate modeling information for all of your accounts.

 

 

3Overview Plan4

 

Product Overview

Analyze

Plan

Perform

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

5TOP5

Products

Solutions

Services

Partners

Resources

Company

Contact

Sitemap

Home

2004 Thinking Links, LLC  All rights reserved.