Lessons in scaling for Delhi and beyond

Emma Gibbs

Emma Gibbs, International Education Consultant, was one of the team from Education Development Trust working as ‘learning partner’ to STiR Education as they embarked upon their scaling journey in Delhi. She reflects on the role of learning partner as well as the impact of the study’s findings.

The learning partner role is a fascinating one with the potential to be a hugely powerful tool for organisations and projects starting the – often daunting – journey to scale. When it works well, the learning partner can provide ‘scalers’ with constructive learning and insight and help them to overcome challenges. However, the risk with such roles is that having an outside organisation acting as a ‘critical friend’ could quickly change from being constructive to confrontational; a critical friend could very easily become just a critic. Happily, our experience with STiR was wholly positive, so why was this? To my mind, it was due in no small part to how both parties approached the opportunity: we were both committed to being transparent and constructive with a genuine interest in learning. Importantly, we also had, as an additional factor, a focus on and respect for sharing our findings with the wider sector. 

It was a privilege to be able to observe STiR’s work. STiR is an international NGO that supports education systems to re-ignite the intrinsic motivation of their workforce. They do so by working with teachers to form teacher networks, and with education system officials to support the networks. STiR’s growth has been rapid: they started their work in Delhi in 2012, with an initial pilot programme involving 25 teachers. Just a few years later, in 2017, STiR began working across the whole Delhi education system – covering all 1,050 schools – and becoming one of three major partners to the Delhi Government in their ambitious programme of system reform. 

Scaling: an iterative, social process

They recognised that successful scaling of innovation projects is seldom straightforward and that projects such as theirs do not always translate into impact at scale. They also recognised scaling within an education system is an iterative, social process with a complex network of ever-changing variables. STiR therefore sought out a partner to support rapid learning, reflection and adaptation throughout their first years of scale-up – us.

Our learning partner role was funded by the Department for International Development from 2017 to early 2019. In this role, we have provided support and challenge, as an expert outsider, as STiR scaled their model to a system-wide programme. We based our approach on current scaling frameworks and lean evaluation thinking to provide a continued ‘drip-feed’ of findings to STiR throughout.

Findings for Delhi and beyond

The result has been a report with findings that straddle two different levels: those which are relevant for STiR and the Delhi Government as they continue to learn and adapt their model to ‘stick’ in an ever changing system; and those which are wider ‘scaling Insights’, contributing learning for use by the wider community of organisations trying to scale-up. 

On the former, we identified clear findings about workforce development. This includes the need to understand which system roles are crucial to helping a programme genuinely embed at scale, but also the need to understand what skills, knowledge and attitudes will help people in these roles to be successful. 

For example, early on in our research, we identified the need for STiR to secure the buy-in of headteachers if the programme was to succeed at scale. As part of our iterative approach we used this initial insight to refine our later fieldwork cycles, focusing on exploring and codifying the beliefs and attitudes of highly engaged headteachers in Delhi. STiR can now use this information to develop new ways of engaging with headteachers going forward. 

As for the latter, our wider scaling insights are based on lessons learned from the scale-up process in Delhi. In order to make these lessons valuable beyond STiR’s model and the Delhi system, we have also constructed a number of key questions for scalers to accompany each insight. These aim to guide organisations in their thinking as they work on iterating and adapting the expansion of their own models. 

Each of the eight insights, summarised below, are further explored in the report, right.

1.  Successful scaling partnerships do not always feel easy
Successful collaboration between the Delhi Government and STiR has required skilled management from both parties, alignment of values and purposeful re-negotiation throughout the scaling journey.  Our analysis suggests that a number of complementary qualities and ways of working which have been critical to the successful partnership. 

2. ‘Scaling an attitude’ is not the same as ‘roll out’
Scaling new practices within an existing workforce requires different management from rolling out or disseminating an intervention. We argue that deep scaling requires increasing willingness to implement adaptive management, and a willingness to encourage and learn from ‘positive deviance’. 

3. Build a broad coalition for change
All scaling partners will have to respond to changes in the external political environment that they are in. Success for scalers lies in putting wide-reaching relationships with commissioners and government bodies at the centre of their approach.   

4. It’s not always obvious where the power lies
Political Economy Analyses are often highly theoretical, bearing little resemblance to the reality of the day-to-day functioning of the system. No amount of initial pre-programme analysis can reveal the reality of hidden powers and accountabilities and so scalers should be prepared to consider the complex network of relationships and power structures on an ongoing basis. 

5. System alignment is a marathon not a sprint
With system alignment, the challenge for scalers is in understanding where their organisation or interventions fits into the system. The key consideration here for scalers is learning when to be influenced to align with the existing system, and when to ‘hold your nerve’ on an existing programme model.  

6. Quick data is not bad data
Rapid, SMART, user-generated data collection, even if imperfect, can still drive important changes in behaviours, and increase demand for ever-smarter data. Scalers may find that this process can also foster increased willingness to understand, generate and use data within the system. 

7. Volume should not overshadow quality 
It can be beneficial to focus first on reaching scale and then allowing quality to catch up over time. This allows scalers to see how the model works in practice. After some time, implementers will be better placed to home in on improving quality and understanding how to do this in a personalised and bespoke way. 

8. Blurred responsibilities are not necessarily a problem 
Future programmes should work with donors and governments to create system-level theories of change that are comfortable with a gradual blurring of role responsibilities, and their implications:  increasingly expansive – even overlapping – roles of different actors; and a decreasing likelihood that an evaluation can ever untangle the impact of different interventions.

Ultimately, the experience of acting as a learning partner to STiR has highlighted to me the benefits that come when two organisations are willing to work in sincere partnership with each other, and particularly when both are open to collecting insights for genuine learning and adaptation. Given the breadth of findings uncovered as part of this work both for STiR and for the wider community of potential scalers, I think many more of us could stand to benefit from having a critical friend. 

Our full report is available to download, right, and we look forward to continuing the debate on how to take successful initiatives to scale.