September 9–11, 2013
Anaheim, California
Automating FP&A Analytics Using SAP Visual Intelligence and Predictive Analysis
Learning Points
Create management insight tool using SAP Visual Intelligence
Develop visualizations that facilitate decision making and analysis
Agenda
Business Context
FP&A Margin Analytics
Linking Analytics to Value Creation
Developing visualizations for decision making and analysis
Data progression
Waterfall charts
Scatter plots and Bubble Charts
Box and Whiskers Plot & Stacked charts
Management Insight Tool :-Automation Approach
Short term vs. Long Term
Mathematical Forecasting using historical trends
Forecasting using Multiple Linear Regression and ARIMA
Steps in SAP Predictive Analytics – DEMO
Business Context
Management Insight Tool
• Different systems for business consolidation,
planning and forecasting and financial
reporting.
• The Data required for analysis is at a lot of
different places and its highly detailed.
• Desire for more commentary around drivers of
margins and variance with focus on what
numbers mean vs. what they are.
Automation of P&L Forecasting
• Fields forecasting process takes significant time
and effort to update after the financials close.
• Business Units use different standards to
update forecasting in the planning and
forecasting application.
• Provide mathematical baseline to increase
accuracy of the field forecasting function.
Linking analytics to value creation – Key Metrics for Margin Analysis
Value
Creation
Growth
Net Sales
Cases
(volume)
Rate
($/Case)
Margin
Gross Profit
Rate
($/Case)
Cause of
Change
Operating
Leverage
Expenses
from Ops
Corporate
expenses
Developing displays that facilitates decision making and analysis
Seq.
Key Business Questions :
-Margin Analysis
Metrics
Visualizations
1. Are Sales / Volumes / Margins Growing ? Volume, Rate of change.
Data progression and possible patterns, 12 months trend chart
2. What is driving the growth/change in Sales/ Volumes/ Margins compared to Last year, Plan, forecasts?
Change by customers, products
Waterfall Charts , TY vs. LY, TY vs. Plan, TY vs. forecast
3. Are we realizing higher margins compared to Plan, LY and forecasts ?
% change, point change
Vertical bar chart, horizontal bar chart
4. What aspects of our operations are contributing to improving/lowering margins?
Cause of Change Percentage of parts or as ratios to a whole represented using a pie chart. Waterfall Chart, cause of change analysis – cost, price , Volume, mix calculations 5. Are we meeting our customers expectations across
distribution centers ?
Service level %, volume & sales
Box and whiskers Plot
6.
Are expenses reducing ? Are we making progress in leveraging our scale?Expenses, TY , LY, Plan, Forecasts
Vertical bar chart, horizontal bar chart, bar chart with two y axis, Stacked Column charts and 12 months trend charts
7. How is our portfolio doing in terms of delivering profitable growth? Are we Growing at the expense of pricing ? Are we growing with the customer we have ?
% Change , point change, ratios
Scatter Plots/ Bubble Charts Operating income Point change vs. Net Sales Growth
Scatter Plots/ Bubble Charts Gross Margin Point change vs. Net Sales Growth Scatter Plots/Bubble Charts New Lost business Ratio vs. growth from
Are sales, margins & volumes growing?
Data Progression, Trends
Are we realizing higher margins compared to
Last year, Plan, forecasts?
What is driving this growth/change in volume, sales,
margins compared to Last year, Plan, forecasts?
Waterfall Charts & CVP Analysis
What aspects of our operations (cost basis, pricing,
volume, mix change etc.) are contributing to
improving/lowering margins?
Scatter Plots and Bubble Charts
Are we growing profitably ?
Are we growing at the expense of pricing ?
Box and Whiskers Plots & Stacked Columns Charts
Are expenses growing ? Are we making progress in
leveraging our scale?
Are we meeting our customers expectations across
distribution centers ?
Management insight Tool : Automation Approach
Short Term solution
:-streamline data gathering to
feed into reporting process
Set up retrieves to streamline data gathering from readily accessible systems.
Identify alternative approaches to access data not in current systems (e.g. emailed
spreadsheets, csv files, freehand SQL)
Create consolidated
database/spreadsheets to drive reporting and visualizations
Longer term solution
:-pull data from transaction
systems and warehouse
Identify alternative approaches to pull data for a more
automated solution
Online Report generation and delivery with click thru
capabilities and collaboration options.
Identify alternative warehousing systems/architecture to support solution development and deployment.
Scope of Today’s
Discussion
Short Term Solution Cont.…
Gain a broader understanding of the various
source system and requirements for data
munging & custom calculations.
All financial reporting systems have an
interface to Excel, to run retrieves.
Use freehand SQL or CSV files for various
database sources that do not support
retrieves.
Financial
Consolidation
Financial
Reporting
Data Mart
Planning &
Forecasting
Others
Retrieves Retrieves Retrieves Flat FilesSource
Data files
Consolidated reporting
backup
Generated and distributed manually
SAP Predictive
Analytics
Spreadsheets organized for reporting
Automatically refreshes when new retrieves are run
Short Term approach detailed
Identify granularity of data
needed for each metric
Investigate sources for each
metric at the desired level of
granularity
Estimate effort needed to
streamline gathering of data
available in readily accessible
systems
Catalogue data not available in
readily accessible systems for
further investigation
Automation of P&L Forecasting
Mathematical Forecasting using historical trends
Forecasting using Multiple Linear Regression and ARIMA
Mathematical Forecasting using Historical trends
Forecasting
Causal
Multiple
Linear
Regression
Time Series
Moving
Average
Exponential
Smoothing
ARIMA
Field Forecast
Planning and
Forecasting
System
Native algorithms and desktop R
algorithms can be used.
P&L Forecasting solution works
with limited data volumes and
hence usage of HANA PAL
algorithms or HANA integration
with R may not be required.
Native algorithms: which logic is
implemented natively in PA's core.
Desktop R algorithms:
implemented with R scripts that
are run against a local (desktop) R
installation
R Desktop
Customization
Native SAP-PA
Algorithm
Desktop R
algorithms
‘math models’ to predict performance
16
Forecast Approach
Goal
•
Create base mathematic model to predict the
future trend of Sales and P&L with minimal
variance
Drivers
•
External variables: Inflation (PPI)
•
Internal variables (input and output):
–
Input drivers: cases sold, margin
compression, expenses per case
–
Output variables: Sales ($), Net COGS ($)
and net Opex ($)
Key Assumptions
•
Exclude impact of Strategic initiatives
•
Trend model based on 5 years of historical data
Cases Sold Inflation (PPI) Inflation Pass Thru Expenses Per Case Net COGS ($) Sales ($) Net OPEX ($)