MOVING FORWARD: NEW METHODS FOR CUSTOMER … › wp-content › uploads › ...MOVING FORWARD: NEW...

MOVING FORWARD: NEW METHODS FOR CUSTOMER ENGAGEMENT

Anne E. Dougherty Katherine Randazzo

Amanda Dwelley

We require more of end users as energy and climate goals

increase

The energy landscape is changing dramatically

To rise to these challenges, we need more sophisticated research techniques

Increasingly, we need to move to a more one-to-one understanding of end-users

Learnings from other fields can help get us there

Why Look Further?

2

Consumer Product Marketing: Latent Class Discrete Choice Analysis (LCDC)

Political Science: Micro-Targeting

Human Development: Latent Growth Curve Analysis

New Approaches for New Insight

3

Consumer Product Marketing: LCDC

4


5

Latent Class Discrete Choice Modeling generates customer segments by identifying product attribute preferences through trade-off analysis


6

LCDC: A-Line Segments

7

No Spend, No LED

On-the-Go Shoppers (14%)

Tech Lovers (25%)

Frugal DIY-ers (31%)

Conventional Shoppers (30%)

No Spend, LED

Spend, No LED

Spend, LED

Willing to Spend

Sample Size: 252

LED Interest

LCDC: Reflector Segments

8

Deal-Driven (13%)

Energy Enthusiasts ( 49 %)

Value-Seekers (14%)

No Spend, LED

Spend, No LED

Spend, LED

No Spend, No LED

Willing to Spend

Sample Size: 224

LED Interest

Product Experimenters (24%)

LCDC: Short-term Targets

9

On-the-Go Shoppers (14%)

Deal-Driven (13%)

Tech Lovers (25%)

Frugal DIY-ers (31%)

Energy Enthusiasts ( 49 %)

Value - Seekers (14%)

Conventional Shoppers (30%)

No Spend, LED

Spend, No LED

Spend, LED

No Spend, No LED

Willing to Spend

Reflector groups A - line groups

LED Interest

Product Experimenters (24%)

It’s Best to Use it When: You have a new product or offering to take to market You need clear direction for design and merchandising You need to prioritize marketing efforts in competitive

channels to target customers

Latent Class Discrete Choice Modeling

10


11


12

Traditional segmentation defines and divide a large homogenous population into identifiable groups based on similar characteristics

Micro-targeting identifies individual household propensities to act


13


14

The Obama 2012 campaign was the first in history to successfully leverage Facebook as a tool for mobilizing voter turnout on a mass scale

Applying Micro-Targeting to Program Offers

15

1. Develop Core Database

2. Surveys to Obtain APS-specific Behavioral & Psychographic Data

3. Intermediate Segment Variables

4. Populate Customer Database with New Variables

5. Develop Program-specific Propensity Models

Applying Micro-Targeting to Program Offers

16

1% 4%

5.7%

It’s Best to Use it When: You have access to multiple data sources Transactional or behavioral data is available You have teams willing and able to collaborate and to share

knowledge

Micro-targeting

17

LGCA uses cross-sectional, time-series analysis Identifies consumption curves and how they differ Predicts and describes the shape of consumption curves Identifies which factors mediate consumption

growth (or decay)


18


19

25th Percentile

50th Percentile

75th Percentile 95th Percentile

4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

Cogn

ition

Age (Years)


20 Intelligence volume 40, issue 1, January ti February 2012 "Is age kinder to the initially more able?: Yes, and no“ J. Gowa, Wendy Johnsona, Gita Mishrab, HALCyon Study Teamb, Marcus Richardsb, Diana Kuhb, Ian J. Dearya


21

1 2 3 4 5 6 7 8 9 10 11 12

Usag

e

Month Mean Growth Smith Jones Taylor Anderson


22

Control Group

Treatment A

Treatment B

Treatment C

-1 -2 -3 -4 -5 -6 -7 -8 -9 -10 -11 -12 0 1 2 3 4 5 6 7 8 9 10 11 12

Usag

e

Month

It’s Best to Use it When: You are working with a heterogeneous population You expect wide variation in response to stimulus You have to identify customer groups for targeting

Latent Growth Curve Analysis

23

Market Segmentation 24

Other disciplines offer field-tested methods Greater granularity provides a finer point for targeting Expand programs by aligning offerings with: Customer needs and wants Customer-specific messaging

Closing Points

25

QUESTIONS? Anne E. Dougherty

Director of Social and Behavioral Research [email protected]

REFINING AN AKA-B MODEL FOR GREATER BEHAVIOR CHANGE

Katherine V. Randazzo, Ph.D., Opinion Dynamics Jane S. Peters, Ph.D., Research Into Action

Caroline Chen, Statwizards Brian Smith, PG&E; Andrew Fessel, PG&E

Introduction

Refining the AKA-B Model 2

This presentation is about going beyond the low-hanging fruit in promoting energy-efficient behavior For many years the dominant “model” of how to get customers to save

energy was the PTEM (Physical-Technical-Economic Model) Incent people to purchase technologies and the technologies will save

the energy This was enough for a while, but we have less low-hanging fruit now

I’m going to talk about using a conceptual model of change to help us think about how to go beyond where we have been How do we go beyond installing lighting and certain one-time purchases? How do we get to the human element and to durable changes in the use of

the energy-efficient products? How do we avoid take-back?

California’s IOUs have long used a simple model of behavior change to think about influencing customers to change their energy efficiency behavior AKA or AKA-B Awareness, Knowledge, Attitudes, Behavior Shows up in most program logic models and statewide program planning metrics

Background

3 Refining the AKA-B Model

The CA IOUs wanted an assessment of the currency & usefulness of AKA AKA toward what? Programs? Energy efficiency? A technology? Attitudes are not enough to change behavior

Subsequently, jointly we developed a revised model and generated generic questionnaire items to track movement of customers based on that model and describe participants and non-participants at any point in time to see if

the programs were reaching the right customers

Revised Model (synthesized from a variety of literatures)


Stage Model—not causal Can be used for different motivational

domains Environmental Financial Comfort

Addresses durable change-not one-time or immediate decisions

Can be used for general or specific objects of awareness etc Program Technology/appliance, etc

(these aren’t actual interview questions)

Mapping Revised Model to Original AKA-B


Awareness and Knowledge Collapsed Attitude expanded into

Concern Personal Responsibility

Intention Added Behavior Expanded into

Behavior Change Maintenance

ak A B

Some Uses of the Model


Describes where customers are in the process of changing behavior in a sustained way

Knowing this helps with messaging and targeting If target customers are already far along, then information will not be

effective If customer is not knowledgeable, then information may be most effective

Provides new insight into the decision drivers for participants and nonparticipants, possibly a better estimate of true readiness to adopt technology and change habits on their own

Can be combined with segmentation to even better describe customers/participants/non-participants

Provides guidance on what to measure and at what point

Causal Model with Intervention Points


What We Did Next


Developed questionnaire items to measure each construct Reliability and validity good—following two pre tests testing nearly 100 items

Some insights from our initial uses of the model: It is essential to distinguish between influencing someone to make an energy-

efficient purchase (or behavior) on one occasion versus influencing sustained changes (we knew this before but testing the model reminded us)

It is easier to predict/influence specific attitudes and behaviors than general ones

But general attitudes may be important as the context in which to influence specific behaviors at any given moment Triggers Making new behaviors convenient

It is a lot to ask for a specific utility/third party program to change customers’ environmental awareness or convince them to take personal responsibility for it!

Refined Model (representing insights from prior slide)


But programs can build on a foundation of environmental awareness/concern that was built by mass media campaigns

/

Awareness/ Concern Personal Responsibility Intention Behavior

Change Maintenance

Awareness/Knowledge Intention

Program Makes it Convenient to Change

Habit

Program Information

Behavior Change

Program Messaging

Provides Decision Heuristics

Program Provides Trigger

Specific

Overcoming Barriers: Examples

General Knowledge


Examples: General akAB Messaging

Making the connection between energy and the environment (AK)


Examples: General akAB Messaging

Concern, Personal Responsibility Message

Examples: Specific Programs


Back to Some Uses of the Model


If we can identify customers are the general Intention stage, we can focus efforts/messaging to such things as: Awareness and knowledge of program Convenience Decision aids Triggers

If they are at the general Awareness/Knowledge stage Convenience isn’t enough Information about the program isn’t enough

If have some concern about the environment but have low self-efficacy about doing anything about it, apply different kinds of messaging, e.g.: Value of collective action Examples of the power of the collective: if ½ Californians replace 5

incandescent bulbs with CFLs, it is like taking 200,000 cars off the road

Back to Some Uses of the Model (cont)


The model can be used as the basis for experimentation Identify groups of customers at different stages Randomly assign them to different messaging conditions Assign to different program interventions Then test the relative efficacy of different interventions on different groups

(at different stages) This will help you fine tune messages but also test whether you are thinking

correctly about your customers and what will influence them

Summary


If there is already a foundation at the general level, programs can provide such things as: Information about the program and its benefits, Triggers to take long-intended action, Ways to make taking program-recommended actions convenient, Help for customers to change habits

Mass media programs by utilities, commissions, etc. can work at the general level so that specific programs can be more effective at a level appropriate to them These efforts would clearly be more effective if coordinated

We have said that the model can be used for general or specific behavior and attitude changes, but this version of the model shows how these levels can be used together for greater effectiveness

However, that approach starts with using a conceptual model such as this one for directing both types of actions

SAVING WASTE: ENERGY USE AND WASTE ANALYSIS BY END-USE

Bill Norton Opinion Dynamics Corporation

November 13, 2012

Presentation Overview

Statement of research objectives

Review of traditional approach

Overview of usage and waste analysis

Application of results/implications for future research

2

Research Objectives

Identify gaps in program offerings by providing a more complete assessment of usage at the end-use level

Disaggregate electricity usage by end-use and segment

Develop energy use profiles by end-use and segment that quantify: Base case usage Excess energy use due to inefficient technologies Energy “waste” due to customer behaviors Efficient case usage

3

Research Objectives (cont.)

4

Plug Load %

HVAC %

Cooling %

Other %

Segment Usage by End Use (% kWh)

Ref. %

Lighting %

HVAC End Use Energy Profile (%kWh)

Efficient Case Use,

X %

Technology Waste,

Y %

Behavioral Waste,

Z %

Disaggregation of segment usage by end use

Disaggregation of end use specific usage into efficient case and waste components

Standard Approach

A traditional potential study quantifies available energy savings from DSM by segments

Forecasts are developed by: Understanding baseline energy usage and market

conditions Modeling/forecasting market response to DSM programs

5

Where Behavior Gets Lost

Traditional potential study produces results too blunt for strategic program planning and program gap analysis: Rely heavily on secondary data reflecting regional/national trends Results do not adequately reflect potential associated with behavior

change

Behavior is addressed: Embedded within engineering algorithms (e.g. Hours of Use) Reflected in assumptions regarding efficient measure adoption

However… Baseline assumptions used - no adjustment for “efficient” behaviors Engineering assumptions fail to capture all behavioral influences

6

Behavioral Component

Energy use is defined by the interaction between end users and technology – accurately assessing behavioral component is critical to quantifying savings potential EXAMPLE: ELECTRIC SAVINGS FOR HIGH SEER CAC

ΔkWH = (FLHcool * BtuH * (1/SEERbase - 1/SEERee))/1000

7

End user behavior is embedded here and represents assumed set points at given

outside air temps

A New Approach

Enhanced primary data collection to inform understanding of: Baseline equipment saturation and penetration Baseline building and equipment characteristics Customer equipment use and occupancy patterns

Determination of efficient case behaviors for end uses

Enhanced engineering assessment to more accurately reflect behavioral component of energy use and waste

8

Primary Data Collection: C&I

9

Telephone Survey: 1,600 completes Penetration/saturation 3 end uses; Behavioral/operational practices

Onsite Audits: 345 completed visits Penetration/saturation all end uses; Equipment technical specifications; Behavioral/operational practices

Monitoring: 70 audited sites Lighting, HVAC (limited ref); Occupancy

• Nested sample design 311 onsite audits • Audits completed at sites not nested in phone survey sample included collection of

segment & operational information during recruitment calls

Primary Data Collection: Residential

10

Mail Survey: 4,414 completes; Penetration/saturation; Behavioral/operational practices

On-Site Audits: 297 completes; Penetration/saturation all end uses; Equipment technical specifications; Behavioral/operational practices

Monitoring: 140 completes; Current logging on all circuits; Lighting / occupancy logging; Temperature and humidity

• Originally designed as fully nested sample • Audits completed at sites not nested in phone survey sample included collection of

segment & operational information during recruitment calls

Waste Definitions

11

Waste Type Examples

Equipment type is not high efficiency - incandescent instead of CFL; - standard instead of Energy Star; - regular furnace fan (not ECM); - regular showerheads - no faucet aerators

Equipment is left on or in standby modewhen not in useProgrammable thermostat not aligned withoccupancy hours Water temperate too highThermostat set points too low in summer Furnace fans always on, rather than auto

Maintenance HVAC tuned up regularly

Technological

Equipment Characteristics

Behavioral

Hours of Use

Performance/Temperature Settings

End Use Usage & Waste Definition

12

End Use Energy Consumption

kWh Technology Waste

kWh Behavioral

Waste

Hours

Wat

ts

kWh Minimum Usage

Actual Hours x (kWinstalled – kWeff)

Actual Run Time

Efficient Run Time

kW of installed equipment

kW of efficient equipment

(Hoursactual – Hourseff) x kWeff

Determine actual usage based on primary data: kWhactual = kW/Tontype X Tonsuser X EFLHuser

where: KW/Tontype = Power draw per ton of cooling, a function of SEER Tonsuser = User System capacity in tons EFLHuser = Equivalent full load hours

Analytic Example: Residential CAC

13

EFLHuser determined by CDD profile and equipment “design day”. Model user CDD for different time of day periods and occupancy conditions based on

customer behaviors (set points).


Sample Presentation Name 14

Time Set PointActual

CDD6am-9am 67 85.3

9am-12pm 67 201.312pm-4pm 67 368.54pm-7pm 67 235.6

7pm-10pm 67 131.210pm-6am 67 154.9

Total 1176.8

Time Set PointActual

CDD6am-9am 78.5 16.7

9am-12pm 82 41.212pm-4pm 82 117.64pm-7pm 76.5 135.8

7pm-10pm 76.5 59.110pm-6am 78.5 30.0

Total 400.4

Home A

Home B

EFLH = 320.3

EFLH = 941.5


Determine technology waste associated with three categories of potential efficiency upgrades that affect CAC usage: CAC unit efficiency Building shell Duct sealing

For each category waste calculated as: kWhwaste = kWhactual - kWhefficient

15


16

TimeOccu-pancy

Rec. Set Pt. Opt. CDD

6am-9am Home 78 19.29am-12pm Away 85 23.412pm-4pm Away 85 75.54pm-7pm Home 78 115.3

7pm-10pm Home 78 46.310pm-6am Asleep 82 12.1

Total 291.8

TimeOccu-pancy

Rec. Set Pt. Opt. CDD

6am-9am Home 78 19.29am-12pm Home 78 87.212pm-4pm Home 78 190.94pm-7pm Home 78 115.3

7pm-10pm Home 78 46.310pm-6am Asleep 82 12.1

Total 471.0

Home A Home B

http://www.energystar.gov/ia/partners/publications/pubdocs/HeatingCoolingGuide%20FINAL_9-4-09.pdf

EFLHopt behave= 376.8 EFLHopt behave= 233.4

Getting to Behavioral Waste

17

End Use Energy Consumption

kWh Technology Waste

kWh Behavioral

Waste kWh Minimum Usage

Efficient Technology Threshold

FLHuser FLHopt behav

Why This Matters

Benefits of the study: Improved understanding of current end use energy consumption – particularly behavioral drivers Measurement of the behavior savings potential by end-use and segment Enhanced primary data provides basis for other analyses and ability to address other research questions - stimulate new research objectives

18

Implications for Program Planning

Assess efficacy of technology and behavioral program options to optimize DSM investment

Identify and prioritize among “opportunity pockets” Customer segments or end uses where energy savings can be realized through behavioral program elements or messaging

Assess benefits of replacing widgets or attempting to change how widgets are used, or both

19

STRATEGIES FOR INCREASING ENERGY SMART ACTION IN

COMMERCIAL ORGANIZATIONS

Hannah Arnold

Equipment costs Availability of capital Budget constraints Lengthy internal decision-making processes The corporate approval process Lack of internal support for projects Inability to gain consensus among decision-makers Program awareness

STAFFING

Barriers to Energy Savings Actions

2 2012 BECC Conference

Staffing Grant Approach

2012 BECC Conference 3

RFP and Application

Staffing and Management

Changes

Complete EE Projects

Incentivizing Internal Resources

Pay Overtime to

current employees

Complete project

paperwork

Submit incentive

applications


Staffing Grant

Approval

Completed Motor and

Grocery Projects

Funding External Resources

Hire External

Consultants

Present findings to

internal stakeholders

Secure funding


Staffing Grant

Approval

Implemented Pump

Optimization

Overcoming Participation Barriers


Equipment Costs

Internal Decision-Making

Availability of Capital

Corporate Approval

Staff Resource Constraints

Thank you!

Hannah Arnold, Project Manager [email protected]

510-444-5050 Ext. 118


mailto:[email protected]

PUSH ME – PULL YOU

Positioning Emerging Technologies to Move Markets

Olivia Patterson

Mary Sutter

Jenn Mitchell-Jackson

Dimensions of Emerging

Technology Program Design

Dimension 1: Identify Placement in Technology

Development Continuum

Clearly defining a program’s

market position helps to:

1. Define a program’s role vis-à-

vis other market players;

2. Provide a framework for

tactics (such as development

assessment and introduction

support);

3. Guide selection of end-uses

and projects.

CA technologies:

- Advanced HVAC

- Smart Appliances and

Plug Loads

- Advanced Lighting

- Integrated Building

Designs and Operations

California ContextWhy Emerging Technologies?

Stages of Utility Intervention for Emerging Technologies

Dimension 2: Choose from a Suite of Tactics

Dimension 3: Select and Prioritize

End-Uses

Dimension 4: Collaborate to Identify Strategic

Market Position

Utility Emerging Technology Programs…

PUSH supply side

market actors to

provide innovations

and conduct targeted

R&D

PULL California end

users into adopting

technologies,

approaches &

practices

- Identify where the targeted technology is in

its development (in Stage 1 to 5) and

whether pushes or pulls are needed

- Utility Emerging Technology Programs are

well placed for Stages 2 to 4

Top-rated push tactics

in CA:

- Improving available

data on existing

systems’ performance

- Testing products in

advance of Codes and

Standards

development

Top-rated pull tactics

in CA:

- Verifying equipment

performance and

energy savings

- Increasing visibility of

energy efficient

technologies

LEGAL NOTICE

This poster was prepared as an account of work sponsored by the California Public Utilities Commission. It does not

necessarily represent the views of the Commission or any of its employees except to the extent, if any, that it has

formally been approved by the Commission at a public meeting. For information regarding any such action,

communicate directly with the Commission at 505 Van Ness Avenue, San Francisco, California 94102. Neither the

Commission nor the State of California, nor any officer, employee, or any of its contractors or subcontractors makes

any warrant, express or implied, or assumes any legal liability whatsoever for the contents of this document.

Policy Background Emerging

Technology

Program Goals

- All residential new

construction will be zero

net energy (ZNE) by

2020

- All commercial new

construction will be ZNE

by 2030

- Transform HVAC

market to ensure

optimal equipment

performance for CA

climate

Goal 1: Increase

adoption of energy

efficiency

measures

Goal 2: Increase

supply of energy

efficiency

technologies

Goal 3: Support CA

Strategic Plan and

related solutions

including zero net

energy

CA Long-Term

Strategic Plan lists

four “Big Bold”

Energy Efficiency

Strategies through

2030:

- All eligible low-income

customers have

opportunities to

participate in

efficiency programs by

2020Stage 1

Research and

Development

Stage 2

Development

Support

Stage 3

Assessment

Support

Stage 4

Deployment

Support

Stage 5

Energy Efficiency

Program Support

Bring

entrepreneurs

and investors

together (do not

need to conduct

own R&D)

Support for

specifications,

Codes and

Standards

Prove technical

validity in multiple

settings and for

integrated suite

of measures

Assess market

barriers, enhance

visibility, and

increase market

traction

Incentives for

selling and

distributing

technology; also

support through

outreach,

education, and

training programs

More Pushes More Pulls

Validate savings from new

technologies

Identify new energy efficiency

technologies

Provide knowledge about how

customers interact with new

technologies

Demonstrate to the

marketplace what works

As energy efficiency goals increase, Emerging Technology Programs help:

MOVING FORWARD: NEW METHODS FOR CUSTOMER … › wp-content › uploads › ...MOVING FORWARD: NEW...

Documents

Transcript of MOVING FORWARD: NEW METHODS FOR CUSTOMER … › wp-content › uploads › ...MOVING FORWARD: NEW...