Samples of past work

  • Policy & Strategy Design

    Ideation workshops and tailored advice and support to innovation agencies and research funders, including Innovate UK, NASA SBIR, Innovation Norway and FFG (Austria).

    Evidence syntheses and reviews of policy instruments for BEIS UK and the British Council.

    Climate toolkits for infrastructure PPPs for the World Bank.

    (Incl. projects delivered in previous roles)

  • Evaluation & Impact Assessment

    Quality assurance and grantee support in a European experimentation fund for testing novel innovation support programs (total budget €4M).

    Process evaluation and methodological review of the AfriPlastics Challenge Prize by Global Affairs Canada (prize pot: $5M).

    Assessment of grant allocations from Iceland’s Technology Development Fund.

    (Incl. projects delivered in previous roles)

  • Ecosystem Development

    Sectoral digital readiness assessments in partnership with UNDP and UNEP.

    A review of ecosystem needs and opportunities for the Icelandic Health Tech Cluster.

    Assessment of the Icelandic innovation environment (Nýsköpunarlandið 2.0), in partnership with Northstack and Rannís.

    (Incl. projects delivered in previous roles)

Testimonials

Eszter is a long-term collaborator of ours at Challenge Works, both from her time at Nesta and at StateUp. Eszter has raised our game on evidence and evaluation: she helped design and deliver the first experiments we carried out within the delivery of our challenge prizes, helping us to identify and implement improvements in the fairness of our judging processes. At StateUp, Eszter played a leading role in delivering the evaluation report of our Afri-Plastics Challenge. This was hugely useful to us in helping us conceptualise and evidence a previously under-evaluated aspect of our work, the use of challenge prizes to drive scale. Eszter's work engaged deeply with the motivations and design of our initiative, and the report has helped shape our approach to scaling innovation. She has also been an excellent informal collaborator - sharing feedback, making connections and brainstorming together with us. I am very happy to endorse Eszter [/EC Insights] as a first-class collaborator and thought partner.

Olivier Usher, Director, Challenge Works

Case study

The challenge

Challenge Works (formerly Nesta Challenges) manages challenge prize competitions that allocate millions in funding to innovative solutions. They wanted to assess whether their evaluation process was consistently identifying the most promising innovations and operating fairly across different applicant groups.

Our approach

Working with Challenge Works through the Innovation Growth Lab, we conducted an organisational experiment to test their evaluation system:

  • Designed a testing framework that integrated with their existing operations

  • Analysed evaluation data from two major prize competitions

  • Examined consistency in evaluator decision-making and potential bias patterns

Results & impact

Our analysis found that whilst the process showed no gender or regional bias, there were significant inconsistencies in how evaluators scored identical proposals. This "evaluation noise" affected which innovations received funding.

Challenge Works used our findings to:

  • Implement a revised scoring system that improved evaluation consistency

  • Introduce feedback mechanisms for evaluators

  • Establish quality control measures for ongoing process improvement

These changes enhanced their ability to identify and fund the most promising innovations whilst maintaining their reputation for fair evaluation. You can read more about our findings here.

When allocating resources to innovation, research, or development programmes, even incremental improvements to your selection process can significantly increase impact. Our evidence-based approach identifies optimisation opportunities and delivers practical solutions that work in real-world implementation.

Ready to explore how we can help? Get in touch to discuss your next challenge.