See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/357835745

Designing a modular approach towards innovation Example: Developing and

scaling Climate-Information Services

Technical Report · December 2021

CITATION READS

1 46

6 authors, including:

Jana Patricia Koerner Imbach Pablo

German National Metrology Institute CATIE - Tropical Agricultural Research and Higher Education Center
28 PUBLICATIONS 358 CITATIONS 125 PUBLICATIONS 3,278 CITATIONS

SEE PROFILE SEE PROFILE

Angelica Barlis
Consultative Group on International Agricultural Research
5 PUBLICATIONS 13 CITATIONS

SEE PROFILE

All content following this page was uploaded by Imbach Pablo on 14 January 2022.

The user has requested enhancement of the downloaded file.

Designing a modular approach towards innovation
Example: Developing and scaling Climate-Information Services
Jana Koerner, with contributions from Pablo Imbach, Elisabeth Simelton,
Yen Thi Nguyen, Angelica Barlis, and Kees Swaans
DECEMBER 2021

livelihood strategies, increasing the complexity and efforts

Key messages
required in designing and delivering “actionable” services.
 Food system transformations will require a new
approach towards innovation to be able to deliver In the following, the Info Note will introduce the concept of
impacts at scale. ‘modular design’ as a methodology that can help on one
hand to increase the variety of products or services offered
 Innovations need to be context-specific and respond
to myriads of different famers’ needs. to users, while on the other hand reducing the complexity
for the product or service designers and providers. We will
 Modular design is a ‘new’ way to design, bundle and illustrate the concept with an example of designing and
scale innovations.
scaling CIS by outlining the five main phases to be
 This Info Note outlines the main principles and undertaken for modular products or services. Finally, we
phases of modular design, with the example of will shortly discuss possible limitations or pitfalls of using
Climate Information Services (CIS). such a modular approach.
 Special attention would need to be given to aspects
of inclusion, data ownership and a potential misuse
of the approach to scale unsustainable practices. ‘Modularity’ as a new way to design,
package, and scale innovations
The concept of ‘modular design’ is a general set of design
Ingredients of innovation systems that principles widely used by private sector actors with the aim
deliver impact at scale to increase the offer made to users (‘external variety’),
There is increasing consensus among the development while reducing the complexity for the product- or service
and climate action community that the transformation of providers (‘internal variety’).
our food systems will both require and entail new In a nutshell, modular design is about decomposing an
technologies, organizations, and ways of working together. innovation into its different parts (‘components’), analyzing
Recent global stakeholder dialogues in the run up to the the relevance and interdependencies of these
UN Food System Summit 2021 and the COP26 have components, and then recomposing these components
developed a shared vision, in which national and regional into modules (Figure 1). Components are parts that work
hubs shall develop, test, and scale innovations to promote
better integration of data and digital systems that facilitate
a transformation in food systems (Koerner et al. n.d.).
However, challenges remain for innovations to be adopted
by farmers (as ‘end users’), as they need to be context-
specific and respond to diverse end-users needs (Barrett
et al. 2020). For example, Climate Information Services
(CIS) that enable farmers to adapt to unfavorable climate
and weather conditions, need to respond to different types
of farmers, farming systems, agroecological contexts, and
together to form a functioning item. From there, different  therewith also creating new business opportunities
stakeholders can use, re-use, reconfigure, or repurpose through outsourcing modules across the value chain
these components or entire modules (Ernst 2005), thus (Ernst 2005).
multiplying possible uses, and/or accelerating the process
However, modular approaches also have limitations and
from innovation development to scale.
trade-offs, that need to be considered carefully:
For example, a personal computer as modular system
 Creating new business opportunities for some, may
consists of different modules (e.g., monitor, keypad, disc
also reduce the unique selling position, or relevance,
player) which can be quickly changed. These modules
of others. In the context of science, risks of
again consist of different components (e.g., motherboard,
‘oversharing’ could be mediated by establishing which
central, or graphics processing units). These components
components must not be changed, and which ones
can also be changed, but this would require more technical
can be modified by the users (Landor 2017).
expertise, and to reopen the modules. In the context of
agricultural research for development, modular  Stable standards and design rules may also have the
approaches could be useful for downstream science opposite effect of slowing down of innovation when
applications, like climate services or digital applications. they ‘predict too much’ and thereby lock-in institutional
strategies.
Reducing time and costs of innovation design and
Therefore, ‘design rules need to be firm enough to
development
encourage modular innovation and recombination – but
The concept of modularity can be applied widely to any loose enough not to be constraining to the evolution of the
system, be it an ecological system, cellular metabolism, system (Sabel and Zeitlin 2004)’, and to maintain the
traffic flows, a power grid, or an economy (Marta 2017). independent modules’ resiliency.
Modularization saves time and efforts by:
 multiplying design options (including re-purposing) Example: Climate Information Services
through mix and match of modules; CIS can support farmers, agricultural value chains and
 allowing localized adaptation by non-expert users governments to take strategic, seasonal or immediate
through bottom-up modularity; decisions on planning and managing agricultural
production (Born et al. 2021). Many public and private
 allowing decentralized design processes and
initiatives experiment with innovative ways to develop and
parallelism in testing (Naik et al. 2020);

CCAFS INFO NOTE 2

scale CIS, but often remain at project or pilot levels lacking
Cut to fit modularity: Spreading the use of the same type
mechanisms for scaling and long-term sustainability.
of data and CIS to more locations and beneficiaries with
Challenges on the information supply side include e.g., the components and modules remaining largely
data gaps, limited institutional capacities for climate unchanged. However, this is only applicable within the
modeling and for cross-disciplinary research and same agricultural, climate and socio-economic conditions.
collaboration. On the services demand side, challenges
Component swapping / sectional modularity: Targeting
touch upon end-users’ adaptive capacity enabling changes
CIS to different purposes and contexts will entail to adapt
in behavior and agricultural practices, but also on
or modify the provided products and services, and their
addressing the differences in livelihoods context and
transfer channels, either by swapping components of an
strategies, cultural aspects, and gender and languages
otherwise unchanged service, or putting all components
(Hansen et al. 2019).
newly together.
Generally, there is a mismatch between the current supply
Bus – or slot modularity: Governmental or institutional
and potential demand for CIS. Albeit the climate and
policies and regulations that level the playing fields, could
weather information supply are mostly generated at
be seen as bus- or slot modules. Equally, this could be all
national levels by statal national-level meteorological
kinds of vehicles that could link and/or accommodate
institutions, the full process of translation into advisories,
different components or modules, e.g., business models
dissemination and access of the information is not fully
that offer different components or modules, that can form
accounted for in the service design. Implementation
innovation bundles.
requires a myriad of actors across different levels working
together for the service finally being used by farmers. Component-sharing modularity: Central to most, if not
all, modules of CIS will be a certain component of climate
Along the CIS ‘value chain’ a diversity of climate-informed
or weather data. These might also be needed in different
products and services should address different needs of
resolution levels, for which again a cut-to-fit modularity
multiple next- and end users in terms of content
could be considered.
(advisories) and frequency of delivery, formats, and
communication channels (Figure 2). Furthermore, to
sustain such services on the national level, beyond the How to design for modularity
lifetime of projects, CIS portfolios (of products and
The key goal of realizing modular designs is thus to group
services), need to be underpinned with business or
strongly interacting components or parts together and
financing models and scaling mechanisms.
separate weakly interacting ones. For example, regarding
CIS, crops and agro-ecosystems might be strongly
Explaining types of modularity with the example of CIS
interdependent factors, while the access to digital
scaling pathways
infrastructure might have a weaker relation to these
There are different ways to structure modules and factors. The most crucial step is to identify the boundaries
components. Brax and Toivonen (2007) suggest six of each system that shall be modularized, and the logic in
different forms (Figure 3), which might be useful to which to decompose the system into different modules and
illustrate different scaling pathways: their components, which then can be used, re-used, re-
packaged, or re-purposed (Ethiraj and Levinthal 2005).

CCAFS INFO NOTE 3

Developing a modular architecture for the to use the same information provided via TV also for SMS
CIS value chain services, or for TV but different crops (as different ‘product
family’).
The following five-phase-guide is adapted from the MB
Collaborations Modular Management Series (https://mb-
III. Defining the best fit module cut
collaborations.com/en/modular-system (retrieved 26
October 2021), which is to the knowledge of the authors is This phase is the ‘heart of the process’, where main crucial
currently the most detailed, openly accessible guide components are defined, and composed into modules.
available. These phases will be outlined, using the This module cut will then form the reference frame for all
example of CIS value chains. The phases I – IV are further product varieties. Differently than the previous phases, this
illustrated in Figure 4. phase concentrates on the different end-user needs, and
the main factors that influence their decisions.
Important to note, though, that the whole ‘toolkit’ of
modular design consists of approximately 70 tools and 20 Identifying main components (external variety):
methods to choose from and to adapt in each phase. Identifying the factors that influence current and future
users’ decisions. These can come partly from the portfolio-
I. Defining common objectives analysis, focus on users’ needs and experience cycles,
rather than on (or: adding to) static indicators. This
The objectives and a common vision of the modularization
exercise will result in different user groups/segments with
strategy are decisive for developing a modular product or
different preference profiles. With regard to CIS, factors
value chain architecture. For example, a national portfolio
that would influence users’ grouping could be e.g., the
of CIS might need to accommodate the different goals of
different crops and production systems, eco-systems,
the key actors: economic departments might focus on
climate extremes and variability, but also the farmers’
value chains of prioritized cash crops, while welfare
livelihoods, language, infrastructure and markets. User
departments might focus on the most vulnerable small
profiles are developed with tools and methodologies that
holder farmers. Disaster and risk reduction departments
lend from design thinking, e.g., users’ value propositions,
might prioritize early warnings. While generally,
customer journeys and -experience cycles.
governments would aim for social and economic impacts,
private sector actors would be more profit-oriented and The correct decomposition is decisive and will depend on
need viable business models. Design teams should the (mix of) module cut method(s). The right granularity
therefore represent the main stakeholders’ perspectives. must be found, and all variants of the product family must
be captured, e.g., in a variant tree. This can also be an
II. Strategic portfolio analysis iterative process.
With these different objectives in mind, the next step will Grouping components into modules (internal variety):
be an analysis of the current product or service portfolio (in Components are then grouped based on their relevance,
this Info Note: CIS), as well as the wished future portfolio. and their interdependencies. There are different ways to
Aim is to prevent complexity already at the beginning of the arrive at the best suited grouping of components into
value chain, while also providing for sufficient flexibility, modules, depending on the designers’ respective goals,
since also end-users needs will change over time. The products or value chains, and capacities. Thus, modules
main steps are: could be based on components interactions (‘Design
Structure Matrix’), on social or technical driving forces
Analysis of current situation of product portfolio:
Analyzing existing product program structure, trends, and (‘Modular Function Deployment’), or on their functional
structure (‘METUS method’).
economic key figures. With regard to CIS, key figures could
be e.g., costs of investment and implementation, reach and Regarding CIS, main functions that could structure
impact indicators, and approaches, mechanisms and modules could be the different purposes (‘uses’) of the CIS
methodologies to reach farmers (the CIS last-mile). products, different modes of dissemination (‘transfer’),
entailing different ways of formatting and packaging
Future analysis of product portfolio: Identifying future
trends and success factors, analyze and evaluate future (‘translate’), and consequently, feed from different data
sources and -formats (‘generation’).
scenarios. Regarding CIS, trends might be e.g., increasing
digitalization, or planned future investments in extension
services.
High-level analysis of internal variety: Determine
existing product components and variants. and evaluate
the re-use potential of components between different
product families. For example, a component of CIS could
be the mode of delivery, with variants then being SMS,
radio, TV dissemination. Potential for re-use could then be

CCAFS INFO NOTE 4

 Figure 4: Four steps to optimize Complexity, adapted from MB Collaborations

IV. Variety optimization of the module set

that facilitate the evaluation and implementation of the new
Finally, to find the most feasible balance between offering portfolio. For example, a national portfolio of CIS could be
a variety of different products and services, while reducing composed of different but complementing portfolios and
complexity as much as possible, the variety of the module road maps of the different key stakeholders, from the
sets need to be optimized. The main steps are: delivery of CIS all the way to how (and by whom) the
respective data is generated, packaged, and translated.
Mapping of interdependencies between internal and
external variety: How will a change of a user-related Discussion
variable (e.g., change of crop), influence the technical
characteristic (e.g., content of advisory) of the component Modularization processes are established in series of
(e.g., SMS service)? This can be made transparent, e.g., workshops with relevant stakeholders and users. Time and
by using a visual tool such as a variance-analysis diagram resources must be invested in ensuring that vulnerable and
(VAD). marginalized people can be properly represented.
However, a critical role will be to build in feedback loops,
Variety reduction: Finally, the number of emerging e.g., adding an interactive component to digital modules,
varieties is reduced by prioritizing the strongest and most where farmers could rate the services, and/or suggest new
relevant dependencies, thus creating modular product ones. This could help to ensure that the product- or service
families. For example, farmers with access to internet set-up remains relevant to changing needs and uses.
could technically be served by the digital product family of
CIS. If, however, the translation of digital services into the For example, with increasing digitalization, more users
local language of a small group of farmers would require could switch to online applications, which might entail
more effort and resources than training the respective extending these applications to other regions or crops.
extension technicians, it would be more efficient to serve Feedback loops could also provide wins for user
this farmer group by the ‘offline’ (personal extension) engagement and -learning components, and increase data
product family. collection points, e.g., through crowdsourcing data on the
impacts of CIS. At the same time, safeguards must be built
V. Organizational processes in right from the start on how to responsibly generate,
disseminate and use farmers’ data and ensuring their data
This process of creating a shared modular product- (or ownership.
service-) architecture will result in visualizations, road
maps and distributed stakeholders’ roles and action plans,

CCAFS INFO NOTE 5

A modular design approach could be used in the context Marta SP. 2017. The importance of being modular.
of developing a national portfolio (menu) for CIS in Vietnam Science (80-. ). 357, 128–129.
and/or Southeast Asia region. Main critical question would https://doi.org/10.1126/science.aan8075
be if such approach, even if developed and applied to Naik HS, Fritzsche A, Moeslein KM. 2020. Modularity in
deliver impacts at scale, would be able to cater for the making: simplifying solution space for user innovation.
context-specific needs of smallholder farmers, and not lend R&D Manag. 51, 57–72.
itself to be mis-used as a scaling vehicle for large-scale, https://doi.org/https://doi.org/10.1111/radm.12427
unsustainable practices and farming schemes. Sabel CF, Zeitlin J. 2004. Neither modularity nor relational
The project team plans to pilot an adapted approach for contracting: Inter-firm collaboration in the new economy.
Vietnam in 2022, which will require to generate the Enterp. Soc. 5, 388–403.
https://doi.org/10.1093/es/khh057
respective instruments and data, engagement with the
relevant stakeholders, and last not least, the demand for a
modularized national CIS portfolio.
This Info Note contributes to the CCAFS Climate
Services Menu for Southeast Asia (CLISM) project.
References
On the authors:
Barrett CB, Benton TG, Cooper KA, Fanzo J, Gandhi R,
Herrero M, James S, Kahn M, Mason-D’Croz D, Mathys Jana Koerner (j.korner@cgiar.org) is the CCAFS
A, Nelson RJ, Shen J, Thornton P, Bageant E, Fan S, Global Innovation Manager and member of the
Mude AG, Sibanda LM, Wood S. 2020. Bundling CGIAR/GIZ Task Force on Scaling.
innovations to transform agri-food systems. Nat. Sustain.
3, 974–976. https://doi.org/10.1038/s41893-020-00661- With contributions of:
8 Pablo Imbach (pablo.imbach@catie.ac.cr) is Climate
Born L, Prager S, Ramirez-Villegas J, Imbach P. 2021. A Action Lead at CATIE.
global meta-analysis of climate services and decision-
Elisabeth Simelton (e.simelton@cgiar.org) is
making in agriculture. Clim. Serv. 22.
Climate Change Scientist at World Agroforestry
https://doi.org/10.1016/j.cliser.2021.100231
(ICRAF).
Brax SA, Toivonen M. 2007. Modularization in business
service innovations. 2007 ISPIM Conf. ‘Innovation Yen Thi Nguyen (nguyenthi.yen@care.org.vn) is
Growth Challenges East West’ 12. Climate Change Advisor at CARE Vietnam.

Ernst D. 2005. Limits to Modularity : Reflections on Recent Angelica Barlis (a.barlis@cgiar.org) is Associate
Developments in Chip Design. Ind. Innov. Fellow at the Alliance of Bioversity International and
https://doi.org/10.1080/13662710500195918 CIAT.
Ethiraj SK, Levinthal DA. 2005. Modularity and Innovation Kees Swaans (c.swaans@cgiar.org) is Climate
in Complex Systems. SSRN Electron. J. Action Lead for Asia at the Alliance of Bioversity
https://doi.org/10.2139/ssrn.459920 International and CIAT.
Koerner J, Tasse A, Zeppenfeldt L, Healy-Thow S, Girvetz
E, Baethgen W, Dinesh D, Vermeulen S. n.d.
Transforming innovation systems to deliver impact at
scale. Cambridge Univerity Press.
Landor. 2017. Revolutionizing the way we manage brands:
The Brand Community Model [WWW Document]. URL
https://landor.com/revolutionizing-way-manage-brands-
brand-community-model

CCAFS INFO NOTE 6

View publication stats