Geoff Mulgan – Social and Public Labs, 2014 – Version 1
The radical’s dilemma: an overview of the
practice and prospects of Social and Public Labs
– Version 1
Geoff Mulgan, February 2014
Nesta has a great interest in the work of labs – teams using experimental methods
to address social and public challenges. We host the 70-strong Innovation Lab,
which includes a joint team with the UK Cabinet Office, and programmes with local
government and the health service as well as civil society. In the early 2000s Nesta
set up Futurelab working in education, and subsequently spun out, and we recently
launched a joint venture to spin one of the most successful innovation labs out of
government: the Behavioural Insights Team.
We’ve also done research on innovation methods and labs in the public sector
worldwide (including the Open Book of Social Innovation which documented
many of these, and a forthcoming study with Bloomberg Philanthropies on
innovation teams (i-teams) in national, regional and city government). And we
collaborate with other Labs and innovation teams around the world, through SIX
and other networks, sharing experiences and methods.
This note summarises a personal view of the field of innovation labs - and what
might lie ahead – largely based on Nesta experience. It looks at the ways in which
labs need to be both insiders and outsiders at the same time – and the practical
challenges of the classic ‘radical’s dilemma’. If they stand too much inside the
system they risk losing their radical edge; if they stand too far outside they risk
having little impact. It follows that the most crucial skill they need to learn is how
to navigate the inherently unstable role of being both insiders and outsiders;
campaigners and deliverers; visionaries and pragmatists.
Background – what is a lab?
Laboratories developed in the 18th-19
centuries in science and technology,
ideas. They offered a safe space for trying out ideas - before the successes were
then taken out into the world. Since then labs have become common in
chemistry, physics, electronics, and biology. Versions of labs are present in most
Some labs are deliberately very removed from real life. But from an early date
agricultural labs showed how labs could be more integrated with the outside
world, with research centres like Rothamsted (set up in the mid-19
providing an environment in which new crops and fertilisers, and combinations
of the two, could be experimented with.
century, thanks to various strands of positivism, utopian thinking and
reform. The proponents believed that small scale experiments could
demonstrate the potential direction of social change, part of a broader movement
of utopian ideas (many of which included practical expressions). Robert Owen,
for example, saw his cooperatives, schools and healthcare in 19
Later on, psychology led the way in extending
Paris in the 1890s.
experimentalism as an alternative to blueprints – from John Stuart Mill’s
advocacy of living experiments (and of the role of the state in providing space for
people to experiment); to John Dewey’s arguments for practical experimentation
in education; to Karl Popper’s account of the virtues of incremental
experimentalism. Popper argued that experiment was preferable to top down
design of new institutions, economies and laws, because it allowed for evolution,
adaptation and improvement on a small scale to improve ideas before they were
generalised. He also argued that one virtue of piecemeal social engineering was
that it treated each new issue as sui generis, and not as the basis for
generalisations. Contemporary theorists, such as Roberto Mangabeira Unger,
have drawn on the legacy of Dewey and others to show how experimentalism can
form part of a much more ambitious approach to politics in the 21
Words - There is no shared definition of what constitutes a social or public lab,
though it might be expected to include experimentation in a safe space at one
remove from everyday reality, with the goal of generating useful ideas that address
social needs and demonstrating their effectiveness.
However, sometimes the word ‘laboratories’ is used metaphorically. States and
cities are often described as laboratories of reform (in the 1930s the US Supreme
Court Justice Brandeis wrote that a ‘state may, if its citizens choose, serve as a
laboratory; and try novel social and economic experiments without risk to the
rest of the country’). The word is also currently being used by many
organisations which look more like consultancies or events organisers. A quick
google search also finds dozens of other organisations using the language of labs
or social labs to describe everything from brand development and marketing to
experimental methods to design or discover new ways of working that address
social and public needs.
The landscape of public and social labs
There are now many different kinds of lab applying method to social problems or
public sector. Some can be found in universities in relation to social action,
research, experimentation, and a new generation of social science parks looks
likely to take this into new fields, such as computational social science.
There are several hundred ‘living labs’, mainly technology based, and enabling
some user input to shaping technologies. More recently, labs of various kinds
have spread into governments, some using design methods, some focused on
data, or using challenges to elicit ideas. Many of these were documented in the
Open Book of Social Innovation; a more thorough study of public sector ones is
coming out shortly in the forthcoming Nesta/Bloomberg Philanthopies research
study on i-teams. Other useful overviews include one prepared by SIG/MaRS,
spiral summarised below – better understanding needs and opportunities;
generating ideas; and testing them in practice. Labs don’t usually include
capacities to take ideas on into implementation and scale. Some are closely tied
into big institutions with power and money, others have very few means to
spread their best ideas. But most aspire to influence whole systems and not just
By the field in which they work (education and healthcare to
from understanding issues, through generating ideas to implementation
experiments), advise, use open innovation methods, or primarily work
through funding others.
labs within governments, to ones at arms-length and others wholly
These variables can be mixed in almost any combination – though a more
thorough research of labs would show particular clusters.
Methods for Labs
The simplest way to differentiate Labs is through the first of these variables – the
distinctive methods they use. These are some of the main ones:
DESIGN: these labs try to introduce design thinking into government or civil
Adelaide, the relatively short-lived Helsinki Design Lab and DesignGov in
Australia, Futuregov in the UK, and others such as Region 27 in France. Others
using design methods but less oriented to government include the Institute
without Boundaries and Stanford’s Design for Change Lab. There has been a
steady growth in the number of labs using design methods despite some
setbacks. Design approaches provide a very useful complement to traditional
bureaucratic, top down policy methods. As I’ve shown elsewhere some elements
of design thinking are not unique to design – the use of ethnography and citizen
input; rapid prototyping etc - but can still be powerful.
Others – notably
to fields predominantly based on text and numbers.
CITIZEN-LED IDEAS INCUBATORS: another group of labs share many similar
methods, and see themselves as incubators of ideas derived from citizens rather
than experts. These often use a mix of tools – from engagement methods to rapid
implementation, some drawing on the traditions of intermediate technology.
Many of the organisations or programmes called ‘Social Innovation Labs’ are of
this kind – such as the SILs set up by OASIS in India, BRAC’s Social Innovation
Lab in Bangladesh, MaRS Solution Lab in Toronto, the Lien Centre’s Social
Collaboratory in Singapore, the Sociallab in Chile, or the Goodlab in Hong Kong.
Often there is a strong ethos of empowerment.
DATA AND DIGITAL TECHNOLOGY: Another group of labs emphasise data, with
many brought together in the Open Government Partnership. These include
Code for America and the teams around its various fellows, groups like the Office
for Urban Mechanics in Boston, ODI in the UK, and the team around the new CIO
in Mexico. These various datalabs have yet to take a stable form but generally
involve small teams of programmers working with public servants, or civil
society, to design new ways of combining public data or developing web-based
MySociety in the UK. The Living Labs are a related group which primarily focus
on developing new technologies with some involvement of users. Across Europe
there are also many Living Labs receiving public funding, usually from R&D
programmes and linked in the European Network of Living Labs. The Global
Living Labs organisation, now established as private company, has a more
commercial approach, working mainly with city administrations helping them
procure technology-based solutions.
EXPERIMENT-BASED/PSYCHOLOGY: these labs emphasise the use of formal
development. A prime example of a lab based on psychology is the Behavioural
Insights Team, recently spun out of the UK government into a partnership with
Nesta, and emulated in the US and Singapore. These are labs in quite a strict
sense in that they run experimental trials, and pay attention to data.
ORGANISATION-BASED: these are Labs working within a single organisation to
generate new ideas and options. UNICEF’s Labs in Kosovo, Uganda, Zimbabwe,
and Copenhagen are good current examples.
PROCESS-ORIENTED LABS: process-led labs use multi-stakeholder processes
Various consultancies promote this approach including Forum for the Future and
range of projects, and generally use a range of different methods. Nesta’s
Innovation Lab is a good example – providing intensive support both to small
scale experiments and subsequent scaling up; supporting systemic innovation in
localities; ‘rapid results’ methods; venture investment and acceleration; and tools
to promote adoption of innovations.
Other more hybrid labs include Change
and social accelerators and incubators around the world. They aim to create new
ventures, or offshoots of existing firms, that address social needs, for example in
health. Some focus on intensive support for cohorts of start-ups, with the
primary aim of getting them ready for follow on funding, and in some cases
contracts. Nesta will soon publish a ‘Field Guide to Accelerators’ drawing on
global best practice.
There are many labs defined more by the field of operation. For example:
In education, Futurelab in the UK was a spinout from Nesta focused on
examples include the Innovation Unit (a spinout from the Department for
Education), New York’s education I-zone, and the lab created by the Office
of Personnel Management in the US Federal Government.
Improvement in the US and MIT Agelab, and in the UK, the NHS Innovation
Institute, Health Launchpad and the NHS Regional Innovation Funds.
A systematic survey would undoubtedly find hundreds of these more specialised
labs working in different fields.
The state of knowledge and craft
The social and public labs have much less history to draw on than labs in science
and technology. So far there has been little serious assessment of the
effectiveness of different methods. Many labs are run by enthusiasts for
particular methods, and see their role more as advocacy than testing. Some can
point to the impact achieved by particular projects and programmes, but few
have yet had any independent validation of their claims.
A series of recent meetings have brought together many labs to share
experiences and insights (eg hosted by Kennisland, MaRS, SITRA, SIX and others),
and the craft knowledge of the field is advancing fast, helped by a strong ethos of
learning and honesty and some helpful recent books.
Efficacy of method – there is a great deal of experiment underway with
methods, and some convergence: linking work to big systemic problems;
close engagement of the people most affected by those issues; co-design;
fast prototyping; and bringing together coalitions of supporters. This mix
of methods shows the scale of ambition of many labs. But the tricky
questions mainly flow from the scale of ambition: i) the timescales
necessary for achieving significant change on this scale; ii) the very wide
range of skills needed to influence the conditions for systems to change;
iii) the issues of power and politics raised by more radical ideas. More
modest lab models may have a higher chance of success: for example
generating new service models within existing NGOs or professions.
organisations which have mature systems for adoption and scale (eg the
classic examples like Bell Labs, or university based labs developing new
biotechnology solutions). Most social innovation Labs by contrast are not
sufficiently linked into systems of investment, scale and adoption, and
implementation. A lot of attention has been paid recently to this issue –
for example, improving the incentives for existing public services to adopt
promising or proven new innovations.
demonstration – showing a new method in the hope that this will lead to
take up by others. All implicitly aim to catalyse demonstrable change. Yet
it is inherently hard to prove the overall impact of a lab’s work, let alone
value for money (a problem shared by parallel organisations like the MIT
Media Lab). Case studies can show the successful growth and spread of
new ideas – but most really transformative ideas are likely to take 10-20
years to spread (digital platforms are the exception rather than the rule).
So the most obvious – if imperfect – short-term metric of success is being
Labs and the ‘radical’s dilemma’
Perhaps the fundamental challenge facing labs is the classic ‘radical’s dilemma’ –
do you work from the outside to create a coherent alternative to the status quo,
but risk being ignored and marginalised; or do you work within the system and
directly influence the levers of power, but risk being co-opted and shifted from
radical to incremental change?
Some of the Labs seek to combine top down and bottom up, inside and outside –
and this must be the right route to attempt. But it requires a great deal of subtlety
– mobilising champions and advocates within power structures while also
experimenting outside; orchestrating small scale evidence and showing its
relevance to the larger scale issues. There is, inevitably, no simple formula.
Indeed, this bridging of inside and outside is inherently unstable because of the
range of variables involved.
I prepared a simple chart on these dilemmas in relation to systemic change –
more to map the options rather than to prescribe answers, which are bound to
vary depending on the state of the field.
around complex problems will continue to grow, and that in response labs will
The ones based on method are likely to become more sophisticated in their use of
methods and demonstration of results; augmenting their core methods (eg data
or design) with other methods to improve impact (eg knowledge of policy,
economics, organisational design). Some may become brand leaders globally
(eg for design, data, behavioural insights etc).
We should expect more sophistication in addressing the radical’s dilemma and
managing roles which straddle inside and outside: how to bring in supporters
and champions; how to organise innovation in ways that improve the chances of
adoption of ideas; how to advocate systemic change, and so on.
Some labs will continue to focus primarily on problem solving and use a range
of methods according to the nature of the problem, based on more generic
innovation skills (the approach taken by Nesta).
We should expect more labs to take an explicitly experimental approach – ie
works with control groups (while hopefully avoiding the more simplistic faith in
RCTs as a universal panacea).
Some may develop place-based testbeds, with towns or cities serving as more
overt laboratories for change.
And we should expect many labs to be set up within existing organisations
(such as global NGOs) or networks of organisations (eg in fields such as childcare
or drugs treatment), potentially sacrificing radicalism for better prospects of
seeing ideas taken up.
The field as a whole will also hopefully gather more insights into practical
questions, such as what scale is optimal; what scope of work is ideal (eg how
many different projects or types of project at any one time?); how to organise
handle failure and get the right balance between a healthy openness to learn
from failure, and the risk of making failure so acceptable that people don’t
struggle through to success?
If Labs spread we should also expect more attention to the ethics of
experimentation, since there are very important issues of handling risk and
consent involved, far more than with consumer products.
Finally, there is the question of the relationships to politics – how much can or
future programmes, and how much should they seek to be insulated? When they
generate radical ideas, how much should Labs move into campaigning and
advocacy? How much could Labs become part of the mainstream toolkit of
Mayors and Ministers?
There will be many different, and valid, answers to these questions. But this feels
like a good time for labs to share experiences; to interrogate each others’
methods; and to move beyond advocacy to deepening effectiveness and impact.
SIX is the Social Innovation Exchange; www.socialinnovationexchange.org and
Social Innovation Europe.
My book ‘The Locust and the Bee’ includes a chapter on utopian experimentation
centuries, and shows how many bold utopians also put their
ideas into practice.
the most ambitious account of the maximalist approach to social innovation
WISIR paper “What is a Design Lab?,” the SiG@MaRS report “Labs: Designing
At Nesta we’ve published a series of blogs on how the methods used by
demand and use, and how ideas can progress along the innovation spiral.
creating new labs
See Zaid Hassan, The Social Labs Revolution, 2014, which eloquently describes
Christian Bason, Leading Public Sector Innovation: Co-Creating for a Better
Design Lab/SITRA, 2011
See Nesta’s overview of performance management which discusses the
risk management around social experiments, including reversibility, choice, how