quantitative methods for sizing markets
DESCRIPTION
http://www.ilyamirman.com/bid/37982/Quantitative-Methods-for-Sizing-MarketsTRANSCRIPT
Quantitative Methods for Sizing Markets
6/19/2003
Ilya MirmanVice President, MarketingSolidWorks Corporation
Agenda
• Some considerations on market sizing• Cook-book approach• Available resources• International market sizing• Finding sweetspots
Readily-available Market DataMay Be Hard to Find• “The data available about market
characteristics, competitors, and so on, is frequently inversely related to the real potential of an opportunity; that is, if market data are readily available and if the data clearly shows significant potential, then a large number of competitors will enter the market and the opportunity will diminish.”
Jeffrey TimmonsNew Venture Creation(©1994 McGraw-Hill)
Sizing Techniques
• The golden rule: no one source of info is that reliable, or complete!
• So it’s better to use multiple approaches to “triangulate” on the real numbers
• Metrics that correlate with usage– Government databases– Relevant publications’ reader studies– Commercial databases
Some key questions to ask…
• Is this the primary product produced at the customer? Or is it one of the things that facilitates their business.– Communications Equipment at Lucent…vs– Packaging Machinery at Procter & Gamble
• What industries are the sweet-spots for this product/segment?
• Is this a MARKET or a POTENTIAL MARKET?• Do I want a worldwide overall assessment, or at
the local level?
Several Approaches• Top-down• Bottom-up• Published reports• Which metrics matter?
– Revenue of current competitors– Firms– Size of problem
• “Total Available Market” vs “Annual Spending”• Today’s talk is focused on Total Available Market
Building the “Total Available Market” (TAM) Model
1. Find a widely-available metric that correlateswith demand for your product
2. Carry out statistical analysis to identify the scaling factor between demand and this metric
3. Get database with the metric by industry and geography
4. Build the model5. Verify model’s predictive power
1. Find a widely-available metric that correlates with demand for your product
• Firms in certain industries• Employment in certain industries• Particular occupations in certain industries• Spending on certain categories• Subscription bases of certain magazines• Etc.
Seats VS Customers(for 50 U.S. States)
10
100
1,000
100 1,000 10,000Mechanical Engineers in State (#)
Solid
Wor
ks S
eats
in
Stat
e (#
)
Example of metrics that correlate with demand
Seats VS Readers(for 50 U.S. States)
1
10
100
1000
0.10% 1.00% 10.00%
Trade Magazine Subscribers in State (#)
Solid
Wor
ks S
eats
in S
tate
(#)
Example of metrics that correlate with demand
Example of metrics that correlate with demand
Subscriptions VS MechEng's in Industry
10
100
1,000
10,000
10 100 1,000 10,000
AVG Subscription in 3-digit SIC
ME'
s Em
ploy
ed in
3-d
igit
SIC
Example of metrics that correlate with demand
Trade Mags' Subscribers VS ME's in State(for 50 U.S. States)
0.10%
1.00%
10.00%
100 1,000 10,000
Mechanical Engineers in State
% o
f Tra
de M
agaz
ine'
s Su
bscr
iber
s (%
)
2. Carry out statistical analysis to identify the scaling factor between demand and this metric• Need not be exact• A top-down approximation may be enough• Primary use will be to compare relative opportunities of
various industries and geographies• Examples:
– 1 software seat per each firm of 50+ employees– 1 software seat per 20 workers– 5 software seats for each mechanical engineer– $1,000 for each $100,000 spent on fabricated metal
structures
Scaling Factors May BeIndustry-Specific• Example: want to use “Employees” as the metric• Two industries with different employment
characteristics may have 2 different scaling factors• Example:
– Target customer: IT administrator– Manufacturing sector: ~1 IT admin / 100 workers– Financial Services sector: ~1 IT admin / 25 workers
3. Get database with the metric by industry and geography
• Not always obvious• Be creative• Be resourceful
• Here are some ideas…
Government Databases
• Fairly comprehensive:– Industry– Zip code– State– Firm size
• The caveats:– Don’t distinguish between type of location (sales,
design, manufacturing, etc.);– Don’t provide names of firms;
Commercial Databases
• Harris Infosource http://www.harrisinfo.com/
• OneSource http://www.onesource.com/
• Corptech http://www.corptech.com/
• Thomas Register http://www.thomasregister.com/
• InfoUSA http://www.infousa.com/
• The Problem: incomplete databases (not all firms are included). Nonetheless, good for relative market sizing.
Useful Websites• U.S. Census Bureau http://www.census.gov/
• The 1997 Economic Census http://www.census.gov/prod/www/abs/97ecmani.html
• 1997 Economic Census Reports http://www.census.gov/epcd/www/ec97stat.htm#SUBJECT
• Industries Ranked by Growth http://www.census.gov/epcd/ec97sic/RANKUSD.HTM
• Statistics by US State http://www.census.gov/epcd/www/97EC31.HTM
• Massachusetts statistics http://www.census.gov/epcd/www/97EC_MA.HTM
• U.S. Bureau of Labor Statistics http://www.bls.gov/oes/1998/oessrch.htm
• Industry Quick Reports http://factfinder.census.gov/servlet/IQRBrowseServlet?_ts=73639935829
• Bureau of Economic Analysis http://www.bea.gov/bea/uguide.htm
• Census Bureau’s CD/ROMs http://www.census.gov/mp/www/rom/msrom.html#County
• County Business Patterns http://censtats.census.gov/cbpsic/cbpsic.shtml
• International Statistical Agencies http://www.census.gov/main/www/stat_int.html
• Department of Commerce http://www.commerce.gov/economic_analysis.html
• Economic Census Reports http://www.census.gov/epcd/www/econ97.html
• Related Census Sites http://www.census.gov/main/www/stat_fed.html
• State of the Nation http://www.stat-usa.gov/econtest.nsf
• Trade Data http://www.stat-usa.gov/tradtest.nsf
• Census Bureau Access Tools http://www.census.gov/main/www/access.html
• Business Counts in a Particular SIC Code http://www.melissadata.com/Lookups/sic.asp?
• Audits of Business Publications http://www.bpai.com/index.htm
Published reports (hardcopy, PDF)Structured databases (.XLS and .DBF)
Published reports (hardcopy, PDF)Structured databases (.XLS and .DBF)
Free from the Government:Amazingly Detailed Data
Free from the Government:Amazingly Detailed Data
Free from the Government:Amazingly Detailed Data
Getting Demographics fromIndustry Publications’ Auditshttp://www.bpai.com/index.htm
Subscribers: By Industry
By Geography
Simple Database Manipulation
4. Build the Model
• Could be a simple 1-parameter formula, or a more complex multi-factor model
• Driven by:– Who will use it?– How?– Required accuracy?– Relative market assessment?– By geography?– By industry?– Sales territory assignment?– Product development/investment decisions?
Building the Model: Which Industries Matter?• Need to filter out irrelevant industries!• A good starting point is your current customers or
prospects• Dun & Bradstreet can append company
demographic information to your customer data:– Size– Revenue– Standardized categories of line of business, SIC
codes– Etc.
The Categories
• No one universal way to categorize companies, markets, segments
• Most companies do more than one thing!• Categorization approaches
– Text descriptions– Structured codes (SIC, NAICS, others)– SIC is most popular, being gradually replaced
SIC Code Primer• SIC Codes are 2-, 3-, and 4-digit codes that refer to progressively
detailed industry classifications.• In the example below, we see how the 2-digit SIC Code #35 can be split
into several 3-digit categories, and how one of them (#356) can be further split into seven more – each one with more detail.
• Further still, 3565 (Packaging Machinery) can be broken down further, by product code; Examples:
– 35651-23: Cartoning & Multipacking Machinery– 35651-27: Paper, Film & Foil Wrapping Machinery– 35651-45: Capping, Sealing & Lidding Machinery– Etc.
351 Engines and Turbines352 Farm and Garden Machinery353 Construction and Related Machinery354 Metalworking Machinery355 Special Industry Machinery356 General Industrial Machinery357 Computer and Office Equipment358 Refrigeration and Service Machinery
30 Rubber And Misc. Plastics Products31 Leather And Leather Products32 Stone, Clay, And Glass Products33 Primary Metal Industries34 Fabricated Metal Products35 Industrial Machinery And Equipment36 Electronic & Other Electric Equipment37 Transportation Equipment38 Instruments And Related Products
3561 Pumps and pumping equipment3562 Ball and roller bearings3563 Air and gas compressors3564 Blowers and fans3565 Packaging machinery3566 Speed changers, drives, and gears3567 Industrial furnaces and ovens
Sample Industry Breakdown ofCustomer Database
354 - Metalworking Machinery
356 - General Industrial
Machinery
367 - Electronic Components and
Accessories
359 - Industrial Machinery
382 - Measuring and Controlling
Devices
355 - Special Industry Machinery
353 - Construction and Related Machinery
372 - Aircraft and Parts
371 - Motor Vehicles and Equipment
366 - Communications
Equipment
362 - Electrical Industrial
Apparatus
358 - Refrigeration and Service Machinery Other
• Typically, just a few SIC codes are relevant, making the database work simpler
Example 1
• Opportunity Predictor at the 5-digit ZIP level
• Works for single ZIP code, or a range
Example 2
Employment Across Six Industries
0%
20%
40%
60%
80%
100%
MotorVehicles
PlasticsProducts
Electronics FabricatedStructural
MetalProducts
Aircraftand Parts
MedicalProducts
Empl
oyee
sMachinists
Maintenance and RepairWorkers
Mold Makers and Operators
Cutting, Punching, and PressMachine Operators
Electronic Equipment Assemblers
Welders, Cutters, Solderers, andBrazers
Inspectors, Testers, Sorters,Samplers, and Weighers
Engineering Managers
First-Line Supervisors/Managers
Team Assemblers
• Labor workforce comparison across 6 industries
Example 3• Engineers by state
5. Verify model’s predictive power• Don’t release unproven data!• Find opportunities predicted by the model• Real Example (medical product manufacturers):
1. Predicted opportunity within 3-digit ZIP codes (based on government database)
2. Randomly picked several locations with high predicted opportunity, verified presence of medical manufacturers with Harris Infosource
3. Overlayed current medical customers as double-check
Let’s take a look…
International Market Sizing
• Rule of thumb: USA is 30-40% of world market (Europe is 40%, Japan+Asia/Pacific~20%)
• Alternative approach:1. Find widely-available metric that drives market
opportunity2. May need to adjust by country-specific factors
(e.g., per-capita GDP)
Looking for sweet spots in your base
Question:• After being in business for 5 years, you have:
– 1,000 customers in industry A– 2,000 customers in industry B
• A new telesales person is starting tomorrow. Which industry should she call into?
…It depends…Several considerations:• Product/industry fit• Current industry spending• Industry size• Channel fit• Geographic channel performance