1. Scaling Up—From Vision
to Large-scale Change
A Management Framework
for Practitioners
March 2006
2. Scaling Up—From Vision to Large-scale Change
A Management Framework for Practitioners
March 2006
Authored by Larry Cooley and Richard Kohl
Edited by Rachel Glass
Design by Kris Humbert and Pamela Guandique
Management Systems International
600 Water Street, SW
Washington, DC 20024
Tel. 202-484-7170
Fax. 202-488-0754
www.msiworldwide.com
Funded in part by the John D. and Catherine T. MacArthur Foundation
5. Executive Summary
T he concept of “scaling up” has
become increasingly popular as
donors have acknowledged with
concern the relatively poor record of
innovative pilot projects in extending
This field-tested framework and set of
guidelines offer practical advice on how
their reach to large populations. to carry out each of ten key tasks needed
Recognizing this, in October 2003, the
for effective scaling up.
John D. and Catherine T. MacArthur
Foundation awarded a grant to
Management Systems International (MSI)
to develop a field-tested framework and Step 1: Develop a Scaling-up Plan
set of guidelines for improved
Task 1: Create a Vision
management of the scaling-up process.
This framework was intended to be of 1A. The Model: What Is Being
direct and immediate use to those Scaled Up?
planning, implementing, and funding 1B. The Methods: How Will Scaling Up
pilot projects and to those hoping to take Be Accomplished?
the results of such projects to scale. 1C. Organizational Roles: Who Performs
the Key Functions?
An earlier draft of the Scaling Up
1D. Dimensions of Scaling Up: Where
Management (SUM) Framework was
Does the Scaling Up Occur?
used in field tests with reproductive
health non-governmental organizations Task 2: Assess Scalability
(NGOs) in Nigeria and Mexico and as a 2A. Determining the Viability of the
basis for initial dissemination efforts. This Model for Scaling Up
revised version of the framework reflects 2B. Analyzing the Organizational and
that experience and incorporates the Social Context
feedback from initial dissemination.
Task 3: Fill Information Gaps
One significant finding emerged from Task 4: Prepare a Scaling-up Plan
this research: Few so-called “pilot
projects” take the steps needed to Step 2: Establish the Pre-conditions for
maximize their prospects for scaling up. Scaling Up
The framework and guidelines presented Task 5: Legitimize Change
in this document seek to improve this
Task 6: Build a Constituency
track record by offering practical advice
on a three-step process to carry out each Task 7: Realign and Mobilize
Resources
of ten key tasks needed for effective
scaling up. These tasks include:
From Vision to Large-Scale Change—A Management Framework for Practitioners
1
6. Scaling Up Begins with a Plan
Step 3: Implement the Scaling-up references tools and outside resources
Process that can be of help. Annexes to the
document summarize the field work
Task 8: Modify Organizational
Structures conducted by MSI in Mexico and
Nigeria (Annex 1); present a sequenced
Task 9: Coordinate Action list of questions to guide the scaling-up
Task 10: Track Performance and process (Annex 2); and provide a
Maintain Momentum consolidated list of some of the tools
most useful for each phase of the
For each Task, the document suggests
scaling-up effort (Annex 3).
actions that need to be taken, presents
alternative tactics that can be used, and
2 From Vision to Large-scale Change—A Management Framework for Practitioners
7. Scaling Up Begins
with a Plan
T he persistence of poverty and
preventable illness in low-income
countries after 30 years of
development efforts has drawn attention
to the relatively poor record of pilot and
The persistence of poverty and
preventable illness in low-income countries
demonstration projects in successfully after 30 years of development efforts has
stimulating systemic change and reaching
drawn attention to the relatively poor record
large populations. In rich and poor
countries alike, service providers and of pilot and demonstration projects in
funders find themselves under pressure
to reduce costs, improve social outcomes,
successfully stimulating systemic change
and explain why it has proven so difficult and reaching large populations.
to accelerate the spread of best practices.
Recognizing this, in October 2003, the
MacArthur Foundation awarded a grant To disseminate these findings, with
to Management Systems International particular emphasis on the
(MSI) to study the scaling up of small population and reproductive health
pilot and demonstration projects and to communities.
field test methods for improving the The initial version of the Scaling Up
scaling-up process. The grant called for Management (SUM) Framework
five activities: developed under this grant built on 11
To develop a framework that years of applied research carried out by
synthesized the existing state of MSI under the USAID-funded
knowledge on scaling up; Implementing Policy Change Program.1
To conduct field trials, using the It also drew heavily from the existing
framework as a general guideline; literature on scaling up, strategic
planning, and organizational
To assess the experience of the field
trials; development. This framework was used
in field tests with reproductive health
To revise the framework based on
non-governmental organizations (NGOs)
the experience of the field trials; and
1 See www.msiworldwide.com/ipc
From Vision to Large-Scale Change—A Management Framework for Practitioners
3
8. Scaling Up Begins with a Plan
in Nigeria and Mexico and as a basis for those wanting to extend these projects to
initial dissemination efforts with larger audiences. As such, it focuses on
academics and practitioners. The current the practical steps and concrete tasks
version of the framework has been involved in managing the scaling-up
revised to reflect the lessons from that process.
experience and to incorporate feedback
from initial dissemination.2 It is The remainder of this paper is organized
deliberately operational and is directed around the 3 Steps and 10 Tasks featured
primarily to those involved in funding in the SUM Framework.
and implementing pilot projects and to
2 See Annex 1 for a summary of actions taken in Mexico and Nigeria. An extensive bibliography, PowerPoint presentation, and
detailed case study on the CLP Project in Nigeria are available on request from Richard Kohl (rkohl@msi-inc.com).
4 From Vision to Large-scale Change—A Management Framework for Practitioners
9. Scaling Up Begins with a Plan
Step 1: Develop a Scaling-up Plan
S uccessful scaling up begins with
good planning. Ideally, that
planning starts during pilot project
design and should take place long before
implementation is completed. This
Successful scaling up begins with good
planning. Ideally, that planning starts
section of the Framework provides during pilot project design.
guidelines for Step 1 of the scaling-up
process, Develop a Scaling-up Plan. This
first step includes four distinct tasks,
Concrete results achieved during Step 1
namely:
include a realistic assessment of the
Task 1: Create a Vision; prospects and parameters for scaling up
Task 2: Assess Scalability; and a road map for getting to scale. This
Step also includes developing the
Task 3: Fill Information Gaps; and documentation and beginning to build
Task 4: Prepare a Scaling-up Plan the support that will be needed later in
the scaling-up process.
Task 1: Create a Vision
A lack of agreement on basic definitions studies makes scaling up more difficult.3
and a shortage of well-documented case The term “scaling up,” for example, is
3 Notable exceptions include:
Expand Net Scaling Up Health Interventions http://www.expandnet.net/;
Simmons, Ruth and Jeremy Shiffman. Scaling-up Reproductive Health Service Innovations: A Conceptual Framework. Paper prepared
for the Bellagio Conference: From Pilot Projects to Policies and Programs, March 21–April 5, 2003; November 15, 2002, and
updated February 2, 2005.
Shiffman, Jeremy. Generating political will for safe motherhood in Indonesia. Social Science and Medicine 56. 2003.
Shiffman, Jeremy, Cynthia Stanton and Patricia Salazar. The emergence of political priority for safe motherhood in
Honduras. Health Policy and Planning 19 (6). 2004.
Uvin, Peter, Pankaj S. Jain, and L. David Brown. Think Large and Act Small: Towards a New Paradigm for NGO Scaling Up.
World Development 28 (8). 2000.
Navrongo Health Research Center. What Works, What Fails (Vols. 1 and 2). Ghana Ministry of Health (www.navrongo.org).
Goff, B. Mastering the Craft of Scaling Up in Colombia. Grassroots Development 14 (1). 1990.
Haaga, John G. and Rushikesh Maru. The effects of operations research on program changes in Bangladesh. Studies in Family
Planning 27 (2). 1996.
DeJong, Jocelyn. NGO Experiences of Scaling Up HIV/AIDS Programs. Presentation at World Bank Workshop on Orphans and
Vulnerable Children. May 2003.
Ford Foundation. Asset Building for Social Change: Pathways to Large Scale Impact.
Levine, Ruth. Millions Saved: Proven Successes in Global Health. Washington, DC. Center for Global Development. November
2004.
World Bank. Scaling-up the Impact of Good Practices in Rural Development: A Working Paper to Support Implementation of the World
Bank’s Rural Development Strategy. Report Number 26031. Agriculture and Rural Development Department. June 2003.
From Vision to Large-scale Change—A Management Framework for Practitioners
5
10. Scaling Up Begins with a Plan
The following questions should be
Task 1 focuses on creating a concrete addressed when beginning Task 1:
vision of what scaling up would look like What organizational, process, and
technical factors were critical to
if it were successful. success on a pilot scale?
Can the model be simplified without
undermining its effectiveness? Is it
applied to several distinct strategies absolutely necessary to replicate all
including: the dissemination of a new elements of the model on a large
scale?
technique, prototype product, or process
innovation; “growing” an organization to Does the organization that carried
a new level; and translating a small-scale out the pilot project have the desire
initiative into a government policy. To and organizational capacity to
expand its operations and deliver
organize these issues and differentiate
services on a substantially larger
among strategies, Task 1 of the SUM scale?
Framework focuses on creating a concrete
vision of what scaling up would look like If not, which organization(s) are best
suited and motivated to implement
if it were successful. This vision becomes
the model on a scaled up basis or to
the yardstick for judging scalability (Task serve as partners in implementing
2) and for deciding what more needs to the model?
be done before embarking on the scaling-
Should the scaling-up effort include
up process. Task 1 includes the following
policy change by the government or
four elements: rely exclusively on voluntary
Clarify the model, innovation or adoption by private and non-
project to be scaled up—what is
w governmental organizations (NGOs)?
being scaled up; Is there a need for one or more
Identify the methods of going to intermediary organizations to
scale—the how of scaling up; support the scaling-up process? If so,
what help is needed and which
Determine the organizational roles organizations are best suited to
involved in scaling up—the who of performing these roles?
scaling up;
Establish the expected scope of
the scaling-up effort and the Step 1: Developing a Scaling-up Plan
dimension(s) along which
Task 1: Create a Vision
scaling up will occur—loosely
speaking, the where of scaling Task 2: Assess Scalability
up.
Task 3: Fill Information Gaps
Task 4: Prepare a Scaling-up Plan
6 From Vision to Large-scale Change—A Management Framework for Practitioners
11. Scaling Up Begins with a Plan
Along what dimension(s) should
scaling up take place?
What would scaling up look like if it Scaling up should begin by clarifying
were successful?
exactly what is to be scaled up.
The following paragraphs explore these
questions and more. They are organized
around the elements of what, how, who
purpose finding and testing new
and where the scaling-up effort will take
solutions to a particular problem. By
place.
definition, they include at least one
technical, process, or organizational
1A. THE MODEL: innovation. Examples of each would be: a
WHAT IS BEING SCALED UP? new technology (technical innovation); a Task 1: Create a Vision
novel service delivery approach (process 1A. The Model: What
Scaling up should begin by clarifying Is Being Scaled Up?
innovation); or creative use of a
exactly what is to be scaled up. In the 1B. The Methods: How Will
public/private/NGO partnership Scaling Up Be
discussion that follows, we refer to this Accomplished?
(organizational innovation). A pilot
as the “model.” This model is normally 1C. Organizational Roles:
project can also take a model that has Who Performs the Key
embedded, at least initially, in a project Functions?
worked successfully in one context or for 1D. Dimensions of Scaling
and can include technical, process, and
one problem and apply it to a new Up: Where Does the
organizational components. We refer to Scaling Up Occur?
context or problem.6
untested models or individual
components of models as “innovations.”4 Demonstration projects take an existing
model and raise awareness about its
In thinking about scaling up, it is useful
usefulness. The intention is to make
to distinguish five different types of
existing solutions better known and more
projects5—pilot (or research and
widely accepted by decision makers and
development [R&D]), demonstration,
potential users.
capacity building (or infrastructure),
policy, and service delivery. Capacity-building projects are intended
to create the institutions, skills, physical
Pilot projects, upon which this paper
infrastructure or systems needed to make
mainly focuses, have as their primary
4 The nascent scaling up literature borrows heavily from social science literature on the diffusion and adoption of technological
innovations. While in some sections of this paper we find the “innovation” to be useful, many small-scale pilot projects discussed
in the literature do not contain an obvious technical, process, or organizational innovation. We therefore prefer to use the term
“model” rather than “innovation” to refer to what is being scaled up.
5 Taken from Cooley, Larry. Uses and Limitations of Projects as Instruments of Change. Unpublished: Management Systems International,
1992.
6 The case studies cited above include numerous examples of each of these types of innovation.
From Vision to Large-scale Change—A Management Framework for Practitioners
7
12. Scaling Up Begins with a Plan
find a new and better solution to a
problem on the assumption that, if
successful, these innovative and novel
Each of these five types of projects—pilot,
features can and will be adopted by
demonstration, capacity building, policy, others. Using terminology discussed later
in this chapter, most pilot projects focus
and service delivery—has an internal logic
on “effectiveness” with the implicit
with respect to scaling up. assumption that “efficiency” and
“expansion” will be addressed at some
later date. Demonstration projects,
permanent changes in the level or quality capacity-building projects, and policy
of service delivery. These projects projects implicitly take a first step in
implicitly assume that the other elements scaling up by seeking to accomplish one
necessary for going to scale already exist or more of the tasks necessary to
or are being provided elsewhere. operating on a larger scale: creating
legitimacy and awareness, increasing
Policy projects emphasize advocacy and capacity, and mobilizing resources,
research and focus explicitly on bringing respectively. Service-delivery projects
about changes in public policy. These normally act as an alternative to
projects target policymakers as their government-provided services, dealing
intended audience and do not typically directly with the issue of scale by
include direct provision of services to the reaching as many people as resources
affected public. allow. In none of these cases does the
project design typically include a
Service-delivery projects attempt to complete strategy for reaching scale on a
“projectize” service delivery on the sustainable basis.
assumption that the benefits or services
provided are needed urgently and are not Many of the projects funded by
being provided by existing programs and foundations and other donors describe
institutions. National immunization themselves as pilot projects, as do most of
campaigns offer an obvious case in point. the examples cited in the scaling-up
These projects may or may not contain literature. It is important to note,
innovations. however, that relatively few of these
projects contain an obvious innovation or
Each of these five types of projects—pilot, a research component, and most include
demonstration, capacity building, policy, many elements that need not—or
and service delivery—has an internal cannot—be reproduced on a large scale.
logic with respect to scaling up. Pilot The majority of these projects are better
projects approach scaling up in two seen as service-delivery projects than as
stages—the initial project is intended to true pilot projects.
8 From Vision to Large-scale Change—A Management Framework for Practitioners
13. Scaling Up Begins with a Plan
Any serious effort to scale up a model
or pilot project should be preceded by
testing, clarifying, refining, and
Any serious effort to scale up a model or
simplifying the model to emphasize
those elements essential to its pilot project should be preceded by testing,
success.7 This process can take many
years. In reviewing a number of cases
clarifying, refining, and simplifying the
in the field of rural development, the model to emphasize those elements
World Bank recently observed that
successful cases “generally started
essential to its success.
with 10- to 15-year lead-up times,
during which locally effective and
appropriate technologies and objective evidence), a promising
processes were refined, often with practice (anecdotal reports and
subsidized donor support.”8 This testimonials); a model (positive
same evidence suggests, however, evidence in a few cases); a good
that many social entrepreneurs are practice (clear evidence from several
reluctant to consider simplifications settings/ evaluations); best practices
to their initial models or prototypes, (evidence of impact from multiple
despite the fact that evidence from a settings, meta-analyses, expert
variety of sources indicates that the reviews); or a
factors relevant to the success of policy principle
scaling-up efforts include determined (proven in Categories of “Best Practice”
efforts at simplification.9 multiple settings;
Innovation—minimal objective evidence
considered widely Promising practice—anecdotal reports and
While most models proposed for applicable ‘truism’ testimonials
scaling up are described by proponents essential for Model—positive evidence in a few cases
as “best practices,” few would meet success).” 10 Good practice—clear evidence from several
this standard. To this end, the World settings/evaluations
Bank publication cited above Best practices—evidence of impact from
This first element multiple settings, meta-analyses, expert
introduces a useful set of categories of Task 1 should reviews
originally developed by the U.S. result in a clear Policy principle—proven in multiple
Center for Drug Abuse Prevention settings; considered widely applicable
specification of the
“truism” essential for success
and drawn from the language of rationale for, and
scientific discourse. These distinguish substance of, what
between “an innovation (minimal is to be scaled up.
7 Simplification also implies an effort to identify in an objective way those elements of an intervention that are essential and cost-
effective for producing the desired results. In this regard, the randomized trials pioneered by the Poverty Action Lab are
instructive (see www.povertyactionlab.org).
8 World Bank, op. cit., p. 24.
9 World Bank, op. cit., p. xiv.
10 World Bank; June, 2003; p. xi.
From Vision to Large-scale Change—A Management Framework for Practitioners
9
14. Scaling Up Begins with a Plan
1B. THE METHODS: HOW WILL and expansion (developing a way to
provide the solution on a larger scale).11
Task 1: Create a Vision
SCALING UP BE ACCOMPLISHED? Figure 1 illustrates this progression.12
1A. The Model: What Is
Being Scaled Up? The second element in Task 1 involves
1B. The Methods: Scaling up focuses principally on the
articulating a strategy for how the model
How Will Scaling third stage—expansion—and assumes
Up Be
can best be extended to large numbers of
that workable solutions have been found
Accomplished? people. A good starting point is David
for the issues implied by the first and
1C. Organizational Roles: Korten’s classic depiction of the scaling-
Who Performs the Key second stages. However, literature and
Functions? up process as three successive stages—
1D. Dimensions of Scaling experience both suggest that, within this
Up: Where Does the
effectiveness (developing a solution that
third stage, it is useful to distinguish
Scaling Up Occur? works), efficiency (finding a way to
between several distinct approaches or
deliver the solution at an affordable cost),
methods of achieving
Figure 1. Summary Overview of Issues scale (see Table 1).13
for Analyzing Scaling Up Thus, the SUM
Framework groups
third-stage “expansion”
Transfer of Experience:
State of applying information in a
methods into three
practice: new setting Factors for categories—expansion,
e
evidence and Scaling Up
applicability replication, and
Factors for
Efficiency
Contextual collaboration—
Factors for Internal distinguished from one
Effective - Contextual
ness another by the degree
Internal Replicating, spreading
Contextual to which the
out institutionalizing for
wider impact originating
Internal Better use of Expansion
Context for
resources organization (i.e., the
Efficiency
innovation Local impact organization that
Effectiveness
source of idea, nce:
managed the initial
g experie
Innovations Expandin arning
Internal Le project) continues to
Scale: control implementation
numbers, area
as the model goes to
Source: Jim Hancock, World Bank Consultant scale.14
11 Korten, F. F. and R. Y. Siy, Jr. (eds.). “Sharing Experiences—Examples of Participatory Approaches in Philippines Communal
Irrigation Projects.” In Transforming Bureaucracy: the Experience of the Philippine National Irrigation Administration. West Hartford, CT:
Kumarian Press, 1988. http://www.worldbank.org/wbi/sourcebook/sb0215.htm
12 World Bank, op. cit., p. 14.
13 The following discussion parallels and integrates the distinctions outlined in a recent publication between “expansion of
experience” and “transfer of experience” each of which can be pursued through either organizational growth (“horizontal”) or
institutional and policy change (“vertical”) approaches, and each of which can engage the organization doing the scaling up on a
“direct” or “indirect” basis. (World Bank, op. cit., p. x)
14 Among other useful frameworks for distinguishing alternative methods of scaling up is the following taxonomy: developing
public policies, fostering communities of practice, influencing market forces, changing power relationships, and promoting social
learning. For a full discussion of this framework, see Ford Foundation. Asset Building for Social Change: Pathways to Large-Scale
Impact, 2004.
10 From Vision to Large-scale Change—A Management Framework for Practitioners
15. Scaling Up Begins with a Plan
Expansion here refers to taking a model
to scale by increasing the scope of
operations of the organization that
The SUM Framework groups third-stage
originally developed and piloted it.
Often, expansion occurs in cases of pilot “expansion” methods into three
or demonstration projects, where a
project fills a vacuum in terms of
categories—expansion, replication, and
e
delivering products or services, and collaboration—distinguished from one
where the model and the organization in
which it is embedded are either
another by the degree to which the
inextricably linked or the originating originating organization continues
organization is unwilling to relinquish
control. to control implementation as the model
goes to scale.
The most common form of expansion is
growth, which normally occurs by
branching out into new locations.
Sometimes this growth is accompanied and spinning off aspects or parts of the
by decentralization or restructuring, which originating organization to operate
we regard here as a distinct method of independently.
expansion because of the special
demands it places on the originating Replication involves increasing the use of
organization. Two other methods of a particular process, technology, or
expansion are franchising the model to model of service delivery by getting
organizations operating as agents or others, including the public sector, to
clones of the originating organization, take up and implement the model. In
these cases, an arms-length relationship
Table 1. Types and Methods of Scaling Up between the originating and “adopting”
organizations (defined below) exists.
Type How Replication can occur between
Expansion Growth organizations of the same type (e.g.,
Restructuring or
Decentralization NGO to NGO) or between organizations
Franchising of different types.
Spin-Off
Replication: Policy Adoption One of the most common types of
Grafting replication is policy adoption, when a
Diffusion and Spillover model is scaled up from a pilot run by an
Mass Media NGO to a program or practice mandated
Collaboration Formal Partnerships, Joint and often run by the public sector.
Ventures, and Strategic Alliances
Another common form of replication is
Networks and Coalitions
grafting, where a model—or one
From Vision to Large-scale Change—A Management Framework for Practitioners
11
16. Scaling Up Begins with a Plan
ventures and strategic alliances are
increasingly common methods for
organizing collaborative efforts, as are
Each of the ten scaling-up methods
less formal networks and coalitions based
has pros and cons, and choosing among on memoranda of understanding or
merely a handshake. Typically, these
alternatives involves balancing a number
arrangements include some division of
of considerations. responsibility among the collaborating
organizations.
Some of these arrangements include the
component of a model—is incorporated public sector as a key partner; many
into another organization’s array of others are agreements among civil society
services or methods of service delivery. groups and/or partnerships with private
Policy adoption and grafting can occur firms, such as an NGO involved in
together, as when a public sector agency education and awareness that partners
incorporates a technique innovated by with media organizations to co-create
NGOs into its services, such as a new methods of delivering products and
participatory approach to HIV education. services to an expanded audience.
Diffusion and spillover are other methods Recognition by private firms of
of replication. They tend to be commercial opportunities among the
spontaneous in nature and occur when a poor15 and a growing emphasis on
model spreads by informal networking corporate social responsibility have
with new or existing organizations or greatly expanded the opportunities for
through the use of more deliberate these types of partnerships.
dissemination efforts. Use of mass media is
a special case of diffusion that bypasses
Balancing the Pros and Cons
organizations altogether by marketing
new ideas directly to the affected public. Each of the ten methods listed in Table 1
has pros and cons, and choosing among
Collaboration, the third method for alternative scaling-up methods involves
scaling up, falls somewhere between the balancing a number of considerations.
expansion and replication approaches. Take, for example, the case of replication
Collaboration mechanisms run the gamut through policy adoption—the transfer of a
from formal partnerships to informal model from an NGO to public sector
networks and include a number of institutions. The clear advantages of policy
innovative structures and governance adoption are mandatory compliance and
arrangements. Formal partnerships, joint access to resources, as state and national
15 See, for example, the recent book: Prahalad, C. K. The Fortune at the Bottom of the Pyramid. Philadelphia: Wharton School of
Publishing, 2004.
12 From Vision to Large-scale Change—A Management Framework for Practitioners
17. Scaling Up Begins with a Plan
governments have greater financial
resources than most NGOs. Moreover,
governments generally have greater public
This second element of Task 1 entails
legitimacy, especially if they are
democratic; and donors and foundations weighing these options and arriving at
frequently view operating at scale on a
preliminary decisions about which scaling-
sustainable basis as a more appropriate
role for government than for NGOs. Policy up methods are to be used, with a clear
adoption also has the advantage that it can
occur fairly rapidly in a system where
understanding of the implications.
decision making is highly centralized and
can cover a large area quickly. On the
other hand, organizational congruence— scale. Expansion by NGOs across socially
the match between the skills, procedures, and political diverse regions and
and values of the NGO and those of the audiences is a particular challenge.
government—can be a serious problem.
For example, where the model being The pros and cons of collaboration
transferred involves a highly participatory depend on the nature of the
approach, adoption by bureaucratic public organizations, governance structures,
agencies may be impractical. For this and partnership model that is used.
reason, policy adoption is typically more Collaboration has greatest potential
effective when the model involved is where various organizations have
primarily technical than when process different and complementary skills or
sensitivity and community participation resources, have shared or overlapping
are key factors in its success.16 objectives, and have a high level of
mutual trust. For example, networks
The pros and cons of using expansion as between similar institutions, such as
a scaling-up method largely mirror those between NGOs or between public sector
associated with policy adoption. Major agencies, can be a powerful form of
challenges for expansion are the ability of scaling up. However, because networks
existing management to undertake and are voluntary and frequently lack
implement the necessary internal external resources, the rate of adoption
changes—and to secure sufficient and coverage of the program may be
financial resources—both to support the slower and less widespread, respectively,
scaling-up exercise and to operate at than with other strategies.
16 This corresponds to the distinction between the “hardware” (the technical components) and “software” (participation, quality of
service, and other less tangible components) aspects of change.
From Vision to Large-scale Change—A Management Framework for Practitioners
13
18. Scaling Up Begins with a Plan
Table 2 displays critical factors affecting This second element of Task 1 entails
the choice among alternative scaling-up weighing these options and arriving at
methods along with the implications of preliminary decisions about which
each of these factors for preferred scaling-up methods are to be used, with a
methods: clear understanding of the implications.
Table 2. Choosing a Scaling-up Method 1C. ORGANIZATIONAL ROLES:
Factors to Consider Method Preferred WHO PERFORMS THE
Type of Model KEY FUNCTIONS?
Technology Intensive Any
Process Intensive Expansion or Collaboration The third element of Task 1 is deciding
Comprehensiveness of Model who needs to do what in order for scaling
Specific Practice Any up—and operating at scale—to be
Complete Model Expansion successful.
Capacity of Originating Organization
Strong Expansion or Collaboration Drawing on a typology developed by
Weak Replication Simmons and Shiffman (2003),17 at least
Source of Financing two different organizational roles are
Internal Any involved in scaling up: the originating
External Replication or Collaboration organization that develops and pilots the
Availability of Formal Evaluation and model, and the adopting organization
Documentation of the Model
Yes Any that takes up the model. Adopting
No Expansion organizations may be newly created for
Observability of Results the purpose of taking up the model, or
High Any may be pre-existing. In the case of
Low Expansion collaborative strategies, the role of the
Ease of Transfer to Other Organizations adopting organization is sometimes
High Replication or Collaboration shared between the originating
Low Expansion
organization and one or more partners. In
Quality of Governance
cases where scaling up takes place
High Replication
Low Expansion or Collaboration
through expansion, the originating and
Presence of NGO Networks
adopting organizations are one and the
Strong Replication same. Nevertheless, the conceptual
Weak Expansion or Collaboration distinction is useful because significant
Social Homogeneity expansion will almost certainly require
High Any the originating organization to undergo
Low Replication major change.
17 Simmons, Ruth and Jeremy Shiffman. Scaling-up Reproductive Health Service Innovations: A Conceptual Framework. Paper prepared
for the Bellagio Conference: From Pilot Projects to Policies and Programs, March 21–April 5, 2003. November 15, 2002.
14 From Vision to Large-scale Change—A Management Framework for Practitioners
19. Scaling Up Begins with a Plan
Simmons and Shiffman have noted the
importance of the compatibility of a
model with the values, norms, and
The third element of Task 1 is deciding
systems of potential adopting
organizations.18 This means that, when who needs to do what in order for
considering potential adopting
organizations, it is important to look at
scaling up—and operating at scale—to
how similar the organizational contexts, be successful.
capacities, and values are; how much
adaptation and capacity building will be
needed; and what resources will be
which the originating and adopting
required.
organizations are embedded. If the vision
For example, community-based projects of scaling up involves a new population Task 1: Create a Vision
or location, this requires at least a 1A. The Model: What Is
frequently owe their success to Being Scaled Up?
participation by program participants preliminary assessment of the context 1B. The Methods: How Will
Scaling Up Be
and stakeholders, including local where scaling up will occur. The Accomplished?
ownership, volunteer labor, use of local objective of this assessment is to ensure 1C. Organizational
that the scaling-up strategy takes into Roles: Who
resources, and priorities determined by Performs the Key
community needs. Large public sector account opportunities and threats in the Functions?
bureaucracies embody, almost by new environment and adjusts to social 1D. Dimensions of Scaling
Up: Where Does the
definition, exactly the opposite conditions present in the new context. Scaling Up Occur?
characteristics. Supply-driven and less This should include an assessment of the
responsive to local needs, they are supply and demand for the services that
unlikely to make extensive use of local are to be provided and whether there are
resources or engender a strong sense of any “competitors” present who may be
local ownership. While public threatened by the scaling-up effort.
bureaucracies have compensating
Most discussions of scaling up assume
virtues—legitimacy, resources, and
that the originating organization is also
infrastructure—differences in their
the organization that does the work
“software” (processes) and values may
needed to transfer the model or take the
make transfer difficult and ultimately
model to scale. Experience and theory
jeopardize the viability of scaling up in
both suggest, however, that many of the
cases where these components are
tasks involved in successfully
fundamental to the success of the model.
transferring or expanding a model can
Considerations of compatibility apply best be done by, or with the assistance of,
equally to the social environment in a neutral third party or intermediary
18 Simmons, Ruth and Jeremy Shiffman. Op. cit.
From Vision to Large-scale Change—A Management Framework for Practitioners
15
20. Scaling Up Begins with a Plan
required in originating and adopting
organizations.19
This third element of Task 1 involves Figure 2 summarizes the organizational
identifying the organizations best suited to roles involved in scaling up. The roles of
the originating organization, adopting
perform each of these functions and the organization, and intermediary
major organizational changes that organization can each be performed by
public sector agencies, NGOs, private
scaling up will require of them. voluntary organizations, consulting
firms, and community-based
organizations.
organization charged specifically with
assisting the scaling-up process. The tasks This third element of Task 1 involves
these organizations perform can include identifying the organizations best suited
conducting visioning and planning to perform each of these functions and
exercises; project evaluation and process the major organizational changes that
documentation; political mapping and scaling up will require of them.
stakeholder assessment; coalition
building; design and conduct of advocacy 1D. DIMENSIONS OF SCALING
campaigns; and fundraising. In the case UP: WHERE DOES THE
of collaborative strategies for scaling up,
intermediary organizations can also be
SCALING UP OCCUR?
essential in designing and forming So far, we have reviewed three of the
innovative partnerships. In strategies that
four the key elements of scaling up: the
depend on expansion or replication, they
model, the types and methods of scaling
often play essential roles in assessing and
up, and organizational roles. The fourth
strengthening the internal capacity and final element needed to complete a
vision of the scaling-
up process is
Figure 2. Organizational Roles in Scaling Up
dimension—the size of
the intended scaling-
up effort and the
Intermediary
Originating Adopting vector along which the
Organization
Organization Organization model or project is to
(Process Facilitator)
be extended.
Or ig i na t ing A d opt i ng
S oc ia l S oc ia l
19 One of the major findings from MSI’s research and fieldwork is that there are few intermediary organizations in most
developing countries with the range of skills needed to support scaling-up. Also noteworthy is the absence of funding to cover
these services in most donor-assisted projects.
16 From Vision to Large-scale Change—A Management Framework for Practitioners
21. Scaling Up Begins with a Plan
Extension of a pilot project’s services or
benefits can be along any of the
following five vectors: The fourth and final element needed
Geographic coverage (extending to
new locations);
to complete a vision of the scaling-up
Breadth of coverage (extending to process is dimension—the size of the
more people in the currently served
categories and localities);
intended scaling-up effort and
Depth of services (extending the vector along which the model or
additional services to current
project is to be extended.
clients);
Client type (extending to new
categories of clients); and
along which of these directions the
Problem definition (extending
expansion will occur is the fourth and
current methods to new problems). Task 1: Create a Vision
final element in developing a vision and 1A. The Model: What Is
Being Scaled Up?
To clarify these distinctions, consider a broad strategy for scaling up.
1B. The Methods: How Will
project being piloted in a particular Scaling Up Be
Accomplished?
village and intended to lower infant PUTTING TASK 1 INTO ACTION 1C. Organizational Roles:
mortality by educating pregnant women Who Performs the Key
Functions?
about infant diseases and the need for The starting point for Task 1 depends on 1D. Dimensions of
ante- and post-natal care. When scaling who initiates it and at what stage of the Scaling Up: Where
pilot effort. Experience clearly Does the Scaling
up is discussed, frequently only the Up Occur?
geographic dimension is considered— demonstrates that scaling up is highly
expanding the project to reach more unlikely without some level of active
villages or a larger region, such as a support from the originating
district or the entire country. However, a organization. However, the focus on
number of other dimensions might also scale often begins with donor or host-
be considered. These could include government interest in fashioning a
extending services to more pregnant solution that can be provided to large
women within the original village segments of the affected population. In
(breadth of coverage); offering nutritional cases where the NGO responsible for the
information and/or access to ante- and pilot project does not share this emphasis
post-natal care (depth of services); on coverage, a mismatch of expectations
widening the target population to is likely. The most typical result is vague
include all women of childbearing age language about scaling up in the funding
(client type); or applying the approach to documents for the pilot project and
address other issues, such as maternal limited attention to scaling up during
mortality (problem definition). Deciding implementation. The best protection
on how many people will be served and against this is clear and candid
From Vision to Large-scale Change—A Management Framework for Practitioners
17
22. Scaling Up Begins with a Plan
In other cases, multiple candidates may
be appropriate for scaling up to fill a
Experience clearly demonstrates that larger need, and a selection must be
made. For these, prioritization of projects
scaling up is highly unlikely without for scaling up should be based on the
scalability assessment criteria
some level of active support from the
enumerated in Task 2 below.
originating organization.
In the best-case scenario, scaling up is
anticipated during the initial design of a
pilot project. In such cases, widely
communication early on about each
accepted best practices suggest that the
party’s vision for the future scale of the
following elements be incorporated into
effort, and their willingness to work
the original design and implementation
toward that future.
of the pilot project: doing a baseline
Timing is also critically important. Often, survey; documenting the model,
discussion of scaling up begins only after especially processes such as working
a pilot project is well underway or with local communities; building in an
completed. These projects are often ongoing method for monitoring,
assumed to be successful and ready for measuring, evaluating and publicizing
scaling up based on anecdotal evidence results; and building in mechanisms for
rather than a thorough, evidence-based gaining buy-in from policymakers and
evaluation of the extent and reasons for a other representatives of potential users or
model’s success; an assessment of the adopting organizations.
model’s strengths, weaknesses, and cost-
Regardless of entry point, scaling up
effectiveness; and a comparison with
depends on a shared vision incorporating
alternative models or mechanisms for
the elements described in Task 1 of the
achieving the same goals. Taking the time
SUM Framework. Much of the work
to do an evaluation, assessment, and
needed for ironing out these issues can be
comparison with alternatives is important
done in a workshop setting with officials
and, ideally, should be done by someone
from the originating organization,
who is detached and independent. Third-
potential donors, potential adopting
party assessments often provide elements
organizations, and selected other
essential to the scaling-up process,
stakeholders. Field-tested models and
including documentation or credible
materials have been developed to
verification of impact, which can be used
support this planning effort and are
to publicize and market the model.20
available upon request.21
20 Ashman, Darcy. Closing the Gap between Promise and Practice: A Framework for Planning, Monitoring and Evaluating Social
Development Networks. Unpublished, Mimeo.
21 For more information, contact Richard Kohl at Management Systems International (rkohl@msi-inc.com).
18 From Vision to Large-scale Change—A Management Framework for Practitioners
23. Scaling Up Begins with a Plan
Task 2: Assess Scalability
The second part of the planning process
involves reaching pragmatic judgments
about the scalability of the model or While some of the factors that affect
program in question. While some of the
factors that affect scalability relate to the
scalability relate to the model itself, many
model itself, many relate to the larger relate to the larger context in which
context in which scaling up would take
place. As a result, the task of assessing scaling up would take place.
scalability should usually be undertaken
at the same time as Task 1 to ensure that
the vision and plan are fully informed by need to be present for the model to
the realities of the situation. be replicated successfully?
Does the adopting organization have
The following questions can help to the appropriate organizational and
guide Task 2: implementation capacity, or the
Do relevant stakeholders, potential means to develop that capacity?
partners, and intended beneficiaries Does the needed funding exist for
perceive a need for this kind of replicating the model on a large
model? scale?
Has the model been documented, Are the central mission,
including the process component, organizational culture and values of
and has its cost-effectiveness been the proposed adopting organization
objectively assessed? Does evidence sufficiently compatible with those
indicate that the model is more cost- necessary to adopt and implement
effective than other approaches? the model successfully?
Are there obvious economies or
These issues are discussed in greater
diseconomies of scale?
detail below.
How easily can the institutional
characteristics that were key to the
outcomes achieved be replicated or 2A. DETERMINING THE VIABILITY
enlarged? OF THE MODEL FOR SCALING UP
Is there anything special or unique In analyzing the comparative scalability Task 2: Assess Scalability
about the social context, political 2A. Determining the
of various pilot projects, Ruth Simmons Viability of the
context, or general circumstances of
the pilot project (e.g., cultural, ethnic, enumerated seven useful criteria based Model for Scaling
on the characteristics of successful Up
or religious values/characteristics; 2B. Analyzing the
distribution of power; homogeneity; technological or economic innovations. Organizational and
Social Context
economic conditions) that would She summarizes them as follows:
From Vision to Large-scale Change—A Management Framework for Practitioners
19
24. Scaling Up Begins with a Plan
In addition, experience demonstrates that
the easiest pilot efforts to scale up are
those that involve a clear and replicable
Experience demonstrates that the easiest
technology and that self-generate financial
pilot efforts to scale up are those that resources needed for expansion. This
helps to explain why many of the most
involve a clear and replicable technology
common examples of scaling up are
and that self-generate financial resources commercial or fee-for-service products
such as micro-credit, and why it has
needed for expansion. generally been easier to scale up
innovations, such as new seeds or cell
phones, than models where process, values
Innovations [Models] must be: and organizational context are critical.23
(1) credible, based on sound evidence
or espoused by respected persons The checklist shown in Table 3 is a crude
or institutions; test of the scalability of pilot projects
based on the factors noted above. Every
(2) observable to ensure that potential
users can see the results in practice; check placed in Column A indicates a
factor that simplifies scaling up; and
(3) relevant for addressing persistent
every check in Column C represents a
or sharply felt problems;
complicating factor. A check placed in
(4) having a relative advantage over column B indicates an intermediate or
existing practices [positive cost- neutral situation with regard to a
benefit, including implementation
particular characteristic. By counting the
costs];
number of checks in Column A and
(5) easy to transfer and adopt; subtracting the number of checks in
(6) compatible with the existing users’ Column C one gets a rough measure of a
established values, norms, and model’s scalability. The higher the
facilities; and number, the easier it will normally be to
(7) able to be tested or tried without scale up the model.24
committing the potential user to
complete adoption when results The Scalability Checklist is intended to
have not yet been seen.22 stimulate, not substitute for, serious
22 Simmons et al. (2002), p. 63. Emphasis added.
23 The World Bank paper cited above (op. cit., p. 9) uses the terms “universalist” and “contextualist” to distinguish these two
broad approaches and provides the following definitions of each:
“Universalist approach to scaling up. In this approach, experience provides a set of universal generalizations that can be
U
replicated, directly expanded, or adopted elsewhere with a simple set of rules. This does not require identifying and dealing
with local variability. For that reason, it may take less time and effort than a contextualist approach to scaling up.
Contextualist approach to scaling up. In this approach, practices to be scaled-up are tailor-made at the outset to address
context-specific conditions”
24 Adapted from Implementing Policy Change Project Technical Note #3, Management Systems International
(www.msiworldwide.com/ipc).
20 From Vision to Large-scale Change—A Management Framework for Practitioners
25. Scaling Up Begins with a Plan
dialogue and analysis. It is best used not
as a scorecard to determine what can be
scaled up and what can’t, but as an aid
The Scalability Checklist is best used ... as
for prioritizing alternatives and a means
for identifying some of the actions that an aid for prioritizing alternatives and a
can be taken to simplify the scaling-up
means for identifying some of the
process. Every time it is possible to make
a change that removes a check from actions that can be taken to simplify
Column C (and, perhaps, changes it
sufficiently to replace it with a check in
the scaling-up process.
Column A), the task of implementation
has probably been made easier. For
example, any action that makes the Organizational culture and
benefits of adopting a model more values/principles,
apparent almost certainly increases the Staffing skills and requirements,
ease with which it can be scaled up. Management and leadership style,
Financial system and resources,
2B. ANALYZING THE Task 2: Assess Scalability
External partnerships, and
ORGANIZATIONAL AND 2A. Determining the
Viability of the Model
Monitoring and evaluation.
SOCIAL CONTEXT for Scaling Up
2B. Analyzing the
The broader social and political context Organizational and
In many types of development projects,
in which projects are located can also Social Context
organizational factors are most
exercise substantial impact on the
responsible for pilot-scale success. It is
scaling-up process. For this reason, it is
thus particularly important to identify
important to assess the external
the organizational features that need to
environment in which the pilot project
be retained, recreated, or substituted for
has been operating to identify contextual
scaling up the model successfully. In
factors that may have been essential to
identifying potentially unique or
the success of the model. Here again, the
distinguishing features of the
goal of the analysis is to identify features
organization that implemented the pilot
that need to be recreated or substituted
project—what we call elsewhere in this
for if the model is to be successfully
paper the “originating organization”—the
scaled up. This analysis of the social
categories and methodology used in the
context can be particularly important, as
Institutional Development Framework
these factors are often invisible to those
(IDF)25 are a useful guide. These are:
who—like fish unaware that they swim
25 For more information, tools and guidelines, contact Mark Renzi (mrenzi@msi-inc.com) or see Renzi, Mark. “An Integrated
TOOLKIT for Institutional Development.” Public Administration and Development. Vol. 16, pp. 469-483. 1996. Also available at
http://www.msiworldwide.com/gral/nwproductsinfo/institutional_dev.htm
From Vision to Large-scale Change—A Management Framework for Practitioners
21
26. Scaling Up Begins with a Plan
Table 3. Scalability Checklist
Characteristics A B C
of the Model Simplifying Factor Neutral Complicating Factor
✓ ✓ ✓
Is the model credible? Based on sound evidence Little or no solid evidence
Evaluated by independent Not evaluated by independent
sources sources
Supported and espoused Not supported or espoused
by respected individuals by respected individuals
and institutions and institutions
How observable are the Very visible to casual observation; Not very visible; not easily
model’s results? easily communicated to public communicated to public
Clearly associated with the Not clearly associated with the
intervention intervention
Has a clear emotional appeal Has little or no clear
emotional appeal
How relevant is the model? Addresses a persistent problem Addresses a temporary problem
Addresses a need that is sharply Addresses a need not sharply
felt by the target population felt by the target population
Addresses a need that is sharply Addresses a need that is not
felt by potential adopting sharply felt by potential
organization(s) adopting organization(s)
Does the model have relative Current solutions are Current solutions are
advantage over existing considered inadequate considered adequate
practices? Superior cost-effectiveness Little or no objective evidence of
clearly established superiority to current solutions
How easy is the model to Few decision makers are involved Many decision makers are
transfer and adopt? in adoption of model involved in adoption of model
Small departure from current Large departure from current
practices and behaviors for target practices and behaviors for
population target population
Small departure from current Large departure from current
practices and culture of potential practices and culture of potential
adopting organizations adopting organizations
Little emphasis on values and/or Significant emphasis on values
process and/or process
Model has low technical Model has high technical
sophistication sophistication
Includes a clear and easily Does not include a clear and
replicated technology easily replicated technology
Low complexity; simple with few High complexity; integrated
components package with many components
Able to use current Requires new infrastructure
infrastructure and facilities and facilities
How testable is the model? Able to be tested by users on Unable to be tested without
a limited scale complete adoption
Is funding likely to be Much less expensive than Much more expensive than
available and/or will current practice current practice
resources be saved? Fully funded by revenues or a No dedicated funding source;
dedicated funding source zero or low cost recovery
Total Number of Checks
22 From Vision to Large-scale Change—A Management Framework for Practitioners