analytics for the midmarket and smaller companies

18
A Guide for Small and Midsize Companies Metric-X, LLC 2692 American Drive, Troy, Michigan 48083 www.metricx.com Five Steps for Beating Your Competition With Analytics

Upload: metric-x

Post on 01-Nov-2014

57 views

Category:

Data & Analytics


2 download

DESCRIPTION

How to get started with analytics.

TRANSCRIPT

Page 1: Analytics for the Midmarket and Smaller Companies

A Guide for Small and Midsize Companies

Metric-X, LLC 2692 American Drive, Troy, Michigan 48083

www.metricx.com

Five Steps for Beating Your Competition With Analytics

Page 2: Analytics for the Midmarket and Smaller Companies

Slide 2 metric-x 12/20/2012

Contents

Encouragement

All You Need Is …

Levels of Analytical Capability

The Five Steps

How to Get Started

Page 3: Analytics for the Midmarket and Smaller Companies

Slide 3 metric-x 12/20/2012

Encouragement

Competing on analytics is not only for large enterprises

You can get started with a zero budget: – All you need is a data dump and Microsoft Excel

Much can be achieved with little knowledge of statistics or computer programming

The analytics capability is built gradually by testing theories and gaining knowledge about your data

Page 4: Analytics for the Midmarket and Smaller Companies

Slide 4 metric-x 12/20/2012

All You Need Is …

Three components are needed to compete on analytics:

– Skills – Tools – Data

Page 5: Analytics for the Midmarket and Smaller Companies

Slide 5 metric-x 12/20/2012

Levels of Analytics Capability

We define four levels of capability

Moving to next level involves investment in skills, tools and data organization

Each level involves exploration until a ceiling is reached

Level 3:

Sweet

Spot

Level 2:

Basic

Level 4:

Formidable

Level 1:

Starting

Point

Page 6: Analytics for the Midmarket and Smaller Companies

Slide 6 metric-x 12/20/2012

Levels of Analytics Capability

Level Characteristics

Level 1: Starting Point

• Excel and an extract from the accounting or other system • Use Excel pivot tables to understand data • Ceiling is reached when data from multiple sources needs to be combined

Level 2: Basic • Data from multiple sources is combined by using Excel formulas or simple SQL • Pivot tables or a visualization tool is used • Analytical capability is concentrated among few power users • Ceiling reached when preparing data for analysis becomes complex or time-consuming

Level 3: Sweet Spot • Data is managed in a database system, and is well-designed for analysis • Fresh data is added daily or more frequently • Data is cleaned and transformed to be usable by non-programmers • Visualization tools are widely used, and reports and dashboards can be published to

other users • Ceiling is reached when the team lacks the statistical knowledge or tools to build models

Level 4: Formidable • Proprietary models are developed that enable continuous improvement and prediction • Highly skilled “data scientists” have access to r

Page 7: Analytics for the Midmarket and Smaller Companies

Slide 7 metric-x 12/20/2012

Levels of Analytics Capability

Level Characteristics Ceiling

Level 1: Starting Point

• Excel and an extract from the accounting or other system • Use Excel pivot tables to understand data

• Ceiling is reached when data from multiple sources needs to be combined

Level 2: Basic • Data from multiple sources is combined by using Excel formulas or simple SQL

• Pivot tables or a visualization tool is used • Analytical capability is concentrated among few power users

• Ceiling reached when preparing data for analysis becomes complex or time-consuming

Level 3: Sweet Spot

• Data is managed in a database system, and is well-designed for analysis

• Fresh data is added daily or more frequently • Data is cleaned and transformed to be usable by non-programmers • Visualization tools are widely used, and reports and dashboards can

be published to other users

• Ceiling is reached when the team lacks the statistical knowledge or tools to build models

Level 4: Formidable

• Proprietary models are developed that enable continuous improvement and prediction

• Highly skilled “data scientists” have access to r

Page 8: Analytics for the Midmarket and Smaller Companies

Slide 8 metric-x 12/20/2012

Five Steps

Iterate at each level using these five steps

Eventually, limitations of data, tools or skills will be realized

Invest in getting to the next level when ceiling is reached

2. List Your

Questions

3. Gather the Data

1. Select a Topic

5. Review, Validate the Discoveries

4. Answer

Questions With Data

Page 9: Analytics for the Midmarket and Smaller Companies

Slide 9 metric-x 12/20/2012

Topic: Marketing Questions: • Which regions are we most profitable in? • How does customer satisfaction correlate with

profitability? • What is the ROI on our marketing campaigns? • Which market segment is the most loyal?

Topic: Sales Questions: • Is there a correlation between commissions earned

and profit generated? • Can we predict which customers will be

uncollectable? • What factors impact the length of the sales cycle? • What

Examples of Topics and Questions

Topic: Production Questions: • Which plants / shifts have the highest scrap rates? • What is the variance in quality by time of day? • How long does it take us to set up for a new order? • Do we order too much /too little raw material for

each production run?

Topic: Finance Questions: • What was our work in process (WIP) over the last

twelve months?

• Are we paying too much for express shipping

production parts?

• What is the cost of sales for product X?

• How do discounts affect our profitability?

Page 10: Analytics for the Midmarket and Smaller Companies

Slide 10 metric-x 12/20/2012

Selecting a Topic

Decide which processes will give you a competitive advantage

Examples: – Customer Service – Job Costing? – Technician Productivity – Flexible Pricing – Advertising and Promotions

Identify the metrics that matter

Ensure data is available for analysis

Page 11: Analytics for the Midmarket and Smaller Companies

Slide 11 metric-x 12/20/2012

Excel Pivot Tables

Visualization Tools

Statistical Analysis Tools

Modeling, Simulation Tools

High

Complexity

Low

Complexity

Less Specialized

Analytical Skillset

(“Anybody”)

Highly Specialized

Analytical Skillset

(“Data Scientist”)

Skillset

To

ol C

om

ple

xit

y

Analytical Tools Complexity vs. Skillset

Page 12: Analytics for the Midmarket and Smaller Companies

Slide 12 metric-x 12/20/2012

Data Dump Into Excel

Multiple, Linked Excel Files or Queries

Departmental Data Warehouse

(Database tables with auto-refresh)

Enterprise Data Warehouse

(Multi-department, automated, high volume, “single source of truth”)

High

Invest-

ment

Low

Invest-

ment Less

Sophisticated

(manual, single

user)

Highly Sophisticated

(Automated, clean,

multi-user)

Data Organization

Tim

e &

Mo

ne

y

Data Organization Investment vs. Sophistication

Page 13: Analytics for the Midmarket and Smaller Companies

Slide 13 metric-x 12/20/2012

Modeling,

Simulation

Tools

Excel

Pivot

Tables

Data Dump

Into Excel

Enterprise Data

Warehouse

Data Organization

To

ols

Analytics Maturity

Companies Advance to Next Level After Hitting Ceiling

Statistical

Analysis

Tools

Visualization

Tools

Linked Excel

Files

Departmental

Data Warehouse

Level 1:

Starting

Point

Level 2:

Basic

Capability

Level 4:

Formidable

Analytical

Capability

Level 3:

Sweet Spot

(for Most Companies)

Page 14: Analytics for the Midmarket and Smaller Companies

Slide 14 metric-x 12/20/2012

Step 2: Understand Your Data

Start analyzing your data – Pull data into Excel and create pivot tables, cross tabs, summaries

Identify unknowns – What questions are you unable to answer? – Which metrics cannot be computed?

Define requirements – What targets should be set? – What information does your team need to monitor actual vs target? – What knowledge will give you an edge in your industry?

Page 15: Analytics for the Midmarket and Smaller Companies

Slide 15 metric-x 12/20/2012

Step 3: Organize, Clean Your Data “Flatten” the data

– Create common codes, descriptions regardless of system of origin: • Customer = Cust = Client = Account = Company • General Motors = GM = General Motors Corp • Ohio is in Midwest Territory? Central Region? USA? Americas?

– Make data easy to fetch, avoiding lookups

Clean the Data – Deal with missing values, data entry errors, outliers

Create a Common Repository – Pull the data from multiple systems into a common database – Regularly refresh the data

Page 16: Analytics for the Midmarket and Smaller Companies

Slide 16 metric-x 12/20/2012

Step 4: Learn to Use Analytical Tools

Try several tools – Data exploration, visualization, drill-down, statistical etc.

Get trained, develop expertise in using one or more tools

Create analysis models – Reports, workbooks, models

Publish your findings, invite conversations from co-workers

Promote an analytical culture

Page 17: Analytics for the Midmarket and Smaller Companies

Slide 17 metric-x 12/20/2012

Step 5: Test Your Theories Put your theories on the table. Examples:

– Sales of part X spike in the winter – Our profit margins on service type Y is low – We lose jobs because our estimates are too high – Technician utilization on small jobs is too low and cuts into margins

Validate the theories with historical data

Identify missing information

Update business processes so that new data can be collected to test theories

Page 18: Analytics for the Midmarket and Smaller Companies

Slide 18 metric-x

• Thank you!

• Saad Shah

• 1-248-495-4925

[email protected]