Experience you can trust. Market Transformation Programs: What they are. How to evaluate them. Kansas Corporation Commission Mitchell Rosenberg, Vice President Topeka, Kansas March 26, 2008
Mar 27, 2015
Experience you can trust.
Market Transformation Programs: What they are. How to evaluate them.
Kansas Corporation CommissionMitchell Rosenberg, Vice President
Topeka, Kansas
March 26, 2008
2
Overview
Market Transformation: Definitions and Reality Importance of MT Concepts in Energy Efficiency
Program Design and Evaluation Key Challenges in MT Evaluation Steps in Meeting MT Evaluation Challenges:
examples from actual programs & studies Lessons Learned
3
Definition of Market Transformation
Market Effect: “a change in the structure or functioning of a market or the behavior of participants in a market that results from one or more energy efficiency program efforts.”
– Typically, these efforts are designed to increase the adoption of energy-efficient products, services, or practices, and are causally related to market interventions.
Market Transformation: Market effects that persist once supporting programs are terminated
4
MT References the Product Life Cycle
0%
10%
20%
30%
40%
50%
60%
70%
80%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Year
Ma
rk
et
Sh
are
Introduction Early Acceptance Take Off Maturity
SUPPLY CHAIN
Smaller competitors or new entrants producing No standardization Very high prices Performance problems Distribution mostly in non-standard channels Limited product lines Little marketing support
1 – 2 larger competitors enter Product standards develop Prices high relative to standard Improved performance Limited distribution Limited model line build out Some marketing support
Most large competitors offer product Product standards adopted voluntarily Prices are higher but in line with standard models Product lines built out Good distribution Manu. Marketing support
All major competitors in market Mandatory standards Prices approach those of standard Producers compete on price and features Distribution via all channels Retailer marketing support
CUSTOMERS
Low level of awareness Skepticism of product claims Only early adopters buying
More customers aware Continued skepticism of product claims Small expansion of market beyond early adopters
Strong demand in advanced segments Some demand in all segments
Strong demand in all segments
5
Policy & Programs to Accelerate MT
0%
10%
20%
30%
40%
50%
60%
70%
80%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Year
Ma
rk
et
Sh
are
Introduction Early Acceptance Take Off Maturity
SUPPLY CHAIN ORIENTED Government lab R&D Sponsored corporate R&D Technology road mapping Mediate technology standard setting Development of performance metrics and testing protocols
Vendor technical and sales training Co-advertising Vendor merchandising support Development & promotion of voluntary product efficiency standards Product testing
Vendor technical and sales training Co-advertising Vendor merchandising support Upstream product subsidies Initiate consideration of higher product standards Develop common service specifications
Mandatory codes and standards Promulgate higher voluntary standards
DEMAND ORIENTED
Purchase of prototypes or early models Develop and publicize case studies of applications
Bulk purchase Customer education Rebate programs General EE public relations
Customer education Rebate programs General EE public relations
Continued customer education Rebate programs for higher efficiency units only
6
EE Programs do help transform markets: The case of electronic ballasts
0
10
20
30
40
50
60
70
80
90
1981
1982
1983
1984
1985
1986
1987
1988
1989
1990
1991
1992
1993
1994
1995
1996
1997
An
nu
al B
allast
Sh
ipm
en
ts, M
illio
ns o
f U
nit
s
0
50
100
150
200
250
300
An
nu
al V
alu
e o
f C
on
str
ucti
on
, 1992 D
ollars, $ B
il.
Electronic
Magnetic
Non-residentialConstructionValue
R&D: LBL & CA utilities
Government Procurement
DSM Programs @ $50m/yr
DSM Programs @ $150m/year
Programs Underway
7
The case of resource-efficient clothes washers
1998 2000 2002 2004 2006
Local Programs in Effect
12 50 70 100+ 90
ENERGY STAR & Other Federal Events
1st ENERGY
STAR
specification (1997)
DOE announces new min. standard
National promotions initiated
Federal min standard increased
(2007) Fed min standard and ENERGY STAR
increased
Manufacturers producing ENERGY STAR models
8 14 17 21 24
Number of ENERGY STAR models
18 35 84 125 212*
* Includes only those that meet the revised 2007 specification.
0%
5%
10%
15%
20%
25%
30%
35%
40%
1996 1998 1999 2000 2001 2002 2003 2004 2005 2006
Mar
ket
Sh
are
8
Why MT Concepts are Important in Program Design & Evaluation Program designs must be matched to current stage of
development to be cost effective Harnessing supply side motivations enhances cost
effectiveness High GHG reduction goals
– Need continuous pipeline of energy-efficient products, services, design practices
– Need to extend adoption of established products to all relevant market segments
In some cases, market research provides most fruitful approach to NTG evaluation
9
Key Challenges in MT Evaluation Effects occur over a long time frame
– Events of interest span decades– Program planning & evaluation cycles typically 3 years or less
Multiple program sponsors and policy makers are involved
– Often coordinate activities among jurisdictions– How tease out effects of one state’s programs?
Data are difficult to obtain
– Manufacturer shipment data often proprietary– Distributors, contractors, retailers, designers tend to be small,
scattered, disorganized
10
Steps in Addressing the Challenges of MT Evaluation Market Assessment/Program Logic Model Program History Development: Market Presence
Indicators Selection of analysis objectives and strategy (ies) Appraisal of results/Adjustments for next rounds
Close parallels to steps in program design and revision
11
Program Logic Model: Home Performance
12
Market Assessment/Program Theory Key Objectives
– Location on the product life cycle
– Identification of key market actors & segments, motivations, barriers
– Assess data availability for program management and evaluation
– Get program planners, implementers, regulators on same page re:
short & long term goals and expectations Timing: Should begin in program design phase
Resources: Avoid over-investment
– Secondary data & reports useful
– Selected local primary research often very useful
13
Program History/Market Presence Important to build reference for market effects and
attribution analysis Should include national as well as local events &
trends over time Data to capture
– Program activities: marketing, training, incentives– Participation: customer and vendor, numbers and
characteristics of participants and measures, timing, region
– Efficiency levels supported v. baseline practice
14
Example: EVT clothes washer programs
1998 2000 2002 2004
NATIONAL PROGRAMS AND MARKETS
Local Programs in Effect 12 50 70 100+
ENERGY STAR & Other Federal Events
1st ENERGY STAR
specification DOE announces new min. standard
National promotions initiated
Federal min standard increased
Manufacturers producing ENERGY STAR models
8 14 17 21
Number of ENERGY STAR models
18 35 84 125
ENERGY STAR Market Share 6.2% 9.3% 16.0% 27.2%
LOCAL PROGRAMS AND MARKETS
ENERGY STAR Market Share 24.9% 27.3% 35.7% 49.6%
Number of Partic. Retailers 40 58 70 75
Number of Incentives Paid 1,950 2,476 2,699 4,129
Number of ENERGY STAR units sold without rebates
7 938 1,401 2,131
15
Analysis Strategies Market Effects Indicators
– Market share: sales, prevalence of practices, fleet efficiency indicators (e.g. mean EERs, MPG)
– Price trends– Changes in codes and standards
Attribution Analysis
– Cross sectional– Time Series– Self-reporting– Historical
Best to attempt multiple indicators and attribution analysis approaches
16
Net Effects of Vermont CW programs:Cross Sectional & Historical Approach
Results of cross-sectional modeling
– Dependent variable: State level Energy Star CW market share
– Independent variables:
Program indicators
Customer Demographics & Energy Prices
Change indicator
Model Estimated Mkt. Share % of E Star Sales
Attributed to Program
Year Actual
Market Share No
Program No Program
+ No Cum. Effect Program
Only Program +
Cum Effect Variable
2001 22.6% 12.9% n/a 34% n/a
2003 52.6% 46.8% 43.4% 7% 26%
2004 48.9% 46.1% 43.4% 6% 12%
17
VT clothes washer programs: What’s happening to local program influence
Methods
– ‘No-program Area’ market shares rising faster than market areas
Market conditions
– Impending federal minimum standard changes
– Profitable product for manufacturers
Fair representation of local program effects?
1999 2000 2001 2002 2003 2004
0.00%
10.00%
20.00%
30.00%
40.00%
50.00%
60.00%
No Program
US Average
Program
VT
18
CA commercial lighting programs: presence in the market
Sponsor Electronic Ballasts Rebated US Utility Ballasts Rebated
US EB Shipments
mm Units mm Units
As % of Program
Area EB Sales
As % of all
US EB Shipments mm Units
As % of US
EB Shipments
1992 13.3 0.9 90% 7% 9.4 71%
1993 24.5 1.2 90% 5% 13.2 54%
1994 24.6 2.0 75% 8% 12.2 50%
1995 32.9 1.9 67% 6% 11.1 34%
1996 30.3 1.3 57% 4% 10.5 35%
Total 125.6 7.3 68% 6% 56.4 45%
19
General Challenges from C&I Programs Little available market share data Complex ‘measures’ Greater customer market segmentation Multiple levels of supply chain with direct influence on
project-level equipment selection and design Multiple decision makers and criteria in one customer
Generally greater reliance on ‘weight of evidence’ and self-reported approaches than in residential.
20
Other Potential Approaches to Attribution or Baseline Development Diffusion modeling Delphi or other expert judging Conjoint analysis of preferences for efficient
substitutes for current products
21
Lessons Learned
Generate as detailed a story as possible
– Know the history of the program and its relatives– Know the history of the market and technology– Know what other programs are doing– Develop the program logic with local stakeholders
Know the available data resources
– Sampling, contents, collection methods– Know what others have done
Design data collection to the attribution strategy or strategies
22
Lessons Learned: Maximizing data opportunities
Quality of Indicators
– Accuracy, face validity, bias (lack of same) Replicability (ability to support historical or time
series analysis)
– Sample frame: captures full population, updated regularly, documented compilation
– Sampling: keep it kosher and document it Comparability (supports cross-sectional analysis)
– Same definitions as data collected elsewhere – Capitalize on channels to reach national markets
23
The Bigger Picture
Play well with others
– Markets addressed by local programs are national and international
– Many program operators are heavily involved already – get to know them
– Initiatives under way to procure sales & shipment data Independent program influence?
– Many local programs already coordinate operations, or are developed in explicit reference to each other (e.g.) codes & standards
– Why try to tease apart effects?