Why We Built ResiliKit the Way We Did

We did not build ResiliKit because people need more emergency information. We built it because information alone does not create the capability to act.

That is the claim. This article explains the reasoning and evidence behind it, what we got right, what we got wrong, and what we still do not know.

The problem we observed

The United Kingdom is not short of emergency preparedness guidance. The government's Prepare campaign, the National Risk Register, local resilience forum advice, and a growing number of apps and checklists all tell people what to do. The UK Public Preparedness Survey (2025) reports that roughly half of households feel only slightly or not at all prepared for emergencies.

The question that interested us was not whether people had access to the right information. Most did. It was why access to information was not translating into household action.

Researchers have described this as the limitation of the "information deficit model": the assumption that if people receive the right information they will act on it. A substantial body of evidence suggests this assumption is unreliable. Awareness of risk does not reliably produce preparedness behaviour, even when people say they intend to prepare.

I had observed this gap professionally. Through years of working in the resilience sector, including working overseas after major disasters, I had seen well-designed guidance sit unread and households that understood the risks but had not taken basic preparedness steps. Information and knowledge are necessary. They are frequently insufficient. The harder problem lies in the chain from knowing to intending to preparing to being capable of acting under disruption.

From observation to inquiry

In 2024, I began doctoral research at a northern university. The original interest was organisational resilience, not household preparedness. That changed.

While reading for my first assignment in late 2025, I spent time with the disaster preparedness and capability literature in depth for the first time. Amartya Sen's capability approach, which distinguishes between the resources people have and what they are actually able to do with those resources, offered a way of thinking about the problem that felt closer to reality than the information-deficit framing. A household might know it should have an emergency plan. It might even want one. But whether it can create and maintain one depends on what Sen calls "conversion factors": income, time, housing, health, digital access, language.

Early ResiliKit experimentation and academic exploration developed alongside one another. Neither preceded the other. Reading sharpened the product problem. Building the product exposed questions the literature did not answer. That relationship is a strength, because it keeps the research grounded in practical problems. It is also a risk, because the researcher has a stake in the product succeeding. We try to manage that tension through transparency.

What we built, and why

ResiliKit is a digital tool that helps households create, maintain, and practise their own emergency preparedness plans. It is built around several design principles, each with a different evidential basis.

Household as the unit of analysis. Most preparedness guidance addresses individuals or speaks generically about "your family." We designed for the household, because the evidence on preparedness barriers is household-level: it is the household that shares a building, a budget, dependencies, and needs. We drew on ONS household composition data to identify eleven household types and personalise guidance accordingly.

Nine capability domains. We organised preparedness into nine domains rather than by hazard type. The structure was synthesised from multiple frameworks, including Karacaoglu's Disaster Literacy Capability Scale, the UK National Occupational Standards for Resilience and Emergencies, and European preparedness models. No single source dictated the architecture. It reflects practitioner judgement applied to a synthesis of evidence.

Practical artefacts, not advice. Each micro-tool guides the user through creating a specific, usable output: an emergency contact card with their actual contacts, a grab bag checklist calibrated to their household, a medication record with their actual medications. The design hypothesis is that producing tangible artefacts closes the gap between knowing what to do and having done it more effectively than providing information alone.

A 72-hour planning horizon. We adopted the 72-hour self-sufficiency convention from international preparedness practice, including the EU Preparedness Union Strategy, Finland's 72tuntia.fi, and Sweden's "If Crisis or War Comes" guidance. Seventy-two hours provides a concrete, communicable anchor for household planning. It is a pragmatic planning horizon informed by international practice, not a scientifically derived specification. We continue to review this framing against emerging UK and international guidance.

Accessibility and sustainability. ResiliGroup is a Community Interest Company with an asset lock protecting its community purpose. We charge modest amounts because hosting and operating the platform costs money, the CIC needs a sustainable operating model, and income helps fund continued development and research. We are testing whether a social-enterprise model can make meaningful household preparedness support as accessible as practicable while remaining financially sustainable. The question we find interesting is whether a socially useful preparedness service can sustain itself without creating unnecessary barriers to participation.

What external challenge taught us

As development progressed, we exposed ResiliKit to increasingly structured challenge from prospective users, academics and practitioners. The feedback exposed problems at every level. Our messaging used multiple unanchored timeframes that created a confusing picture of what the user was committing to. Tier names were linguistically misleading: "Individual" implied a single person when the product was household-based, and "Essentials" implied core requirements when it was actually the paid upgrade. The points-based scoring system lacked meaning for users who could not form a mental model of what the score represented.

We accepted every substantive challenge. Nothing was rejected outright. That fact matters: structured external critique found issues everywhere it looked.

What we changed our minds about

One of the most consequential changes was the retirement of ResiliScore.

We had created a single 0-100 number to make plan completion progress visible. The name implied it measured something about resilience. External reviewers challenged whether the number meant anything to users. But the deeper problem emerged from the evidence review we conducted in April 2026.

The resilience measurement field is, as Beccari (2016) documented, framework-rich and validation-poor: 106 composite resilience indices, with limited evidence of their validity. Bakkensen et al. (2017) tested five US resilience indices against observed disaster outcomes and found that some performed opposite to theory. Against this evidence, calling a plan completion percentage a "resilience score" was not defensible.

"Plan Progress" replaced it, with an explicit disclaimer: "Plan Progress shows how much of your plan you've built. It does not measure how prepared you actually are." The underlying formula was unchanged. The problem was what the name implied.

ResiliKit now uses three distinct signals. Plan Progress tracks plan completion. Assurance tracks whether artefacts have been created, briefly practised, and kept current. A short self-assessment captures how prepared the household feels. We separated these because the evidence showed that conflating different constructs into a single number produces misleading signals.

We describe this approach as evidence-responsive. Evidence, testing, and external challenge should be capable of changing the product, not merely cited to justify decisions already made. Evidence-informed design is not the same as evidence-proven effectiveness. Our design decisions are traceable to evidence. The product's effectiveness in improving household capability has not been demonstrated.

So, why is ResiliKit different?

We are not claiming to have invented household preparedness, capability theory, checklists, or behavioural interventions. The potential contribution lies in combining existing knowledge into a practical household system, then systematically testing and revising that system.

Our strongest contribution is not an invention but an integration: translating existing ideas into an operational system that produces usable artefacts, tracks meaningful signals, maintains itself over time, and is governed by a commitment to revise its own design when evidence or challenge warrants it. The distinction between novelty and usefulness matters. Whether this integration actually produces better outcomes for households is the question we have not yet answered.

So what does this mean for ResiliKit now?

We recently completed an internal evidence origin study, reconstructing the evidential basis for every major design decision. The exercise forced us to trace decisions back to their evidential origins, distinguish evidence from practitioner judgement and convention, correct claims that had outrun their evidence, identify assumptions requiring further research, and flag product features requiring testing.

Some decisions survived scrutiny. Others did not. ResiliScore was retired. Claims about what our signals measure were tightened. A proposed measurement instrument described as "in development" in our governance documentation was corrected, because no development had occurred. Conventions we had treated as settled were reclassified as untested and flagged for investigation.

The point is not that ResiliKit has an impressive internal audit process. The point is that we have deliberately built a system capable of changing its mind.

What we do not yet know

The most important thing we do not know is whether using ResiliKit improves real-world household outcomes during a disruption. No outcome study has been conducted. The limited empirical evidence in this field is not encouraging: a retrospective observational study of 1,114 households in New York City found that having a kit, plan, and medication did not reduce risk of adverse outcomes following Superstorm Sandy (Clay et al., 2020). That is a single study, in a single city, using a retrospective cross-sectional design. But it is a serious challenge to the assumption that planning artefacts translate into capability under real conditions.

We do have early formative signals. Anecdotal feedback from initial users, and more structured observations from academic partners who reviewed the product, suggest that interaction with ResiliKit caused some people to think differently about what preparedness resources they already had, what they lacked, what they might actually need, and what they would do during disruption. This is not outcome evidence. It does not demonstrate effectiveness or improved resilience. It does, however, provide an early indication that the product may be affecting the cognitive mechanism it was designed to address: moving people from passive receipt of preparedness information towards active consideration of their own circumstances.

The research question that follows is more useful than the admission that precedes it. Does that change in thinking subsequently produce meaningful preparedness behaviour, maintained capability, or improved outcomes? That question has not been answered. Answering it is the next priority.

A working hypothesis

ResiliKit is a working hypothesis. It is built on evidence-informed design, tested by external challenge, revised in response to evidence review, and governed by a legal structure that protects its social mission.

We believe the approach is sound: treat the household as the unit of analysis, produce practical artefacts rather than advice, track meaningful signals without overclaiming what they represent, and remain willing to change when the evidence warrants it.

Can we demonstrate that helping households turn preparedness information into practical action produces meaningful and sustained capability? That is the research and product-development challenge ahead. We intend to pursue it, and to report what we find, including findings that challenge our own design.

We did not build ResiliKit from a single evidence base. We are building it through an ongoing cycle of evidence, design, challenge, and revision. This article is part of that cycle.

References

1. Bakkensen, L.A., Fox-Lent, C., Read, L.K. and Linkov, I. (2017) "Validating resilience and vulnerability indices in the context of natural disasters," Risk Analysis, 37(5), pp. 982-1004.

2. Beccari, B. (2016) "A comparative analysis of disaster risk, vulnerability and resilience composite indicators," PLOS Currents Disasters, Edition 1.

3. Clay, L.A., Goetschius, J.B., Papas, M.A., Trainor, J., Martins, N. and Kendra, J.M. (2020) "Does preparedness matter? The influence of household preparedness on disaster outcomes during Superstorm Sandy," Disaster Medicine and Public Health Preparedness, 14(1), pp. 71-79. doi: 10.1017/dmp.2019.78.

4. Karacaoglu, D. (2024) Disaster Literacy Capability Scale (DLCS). Delphi-derived instrument measuring self-assessed confidence across disaster literacy skills.

5. Sen, A. (1999) Development as Freedom. Oxford: Oxford University Press.

6. UK Cabinet Office (2025) UK Public Preparedness Survey 2025. Official Statistics In Development.

Explore ResiliKit

See how ResiliKit translates household preparedness guidance into practical planning tools.

ResiliGroup CIC is a Community Interest Company registered in England and Wales. ResiliKit is its household preparedness product. This article reflects ResiliGroup's own analysis and does not represent the views of any academic institution, the UK Government, or any other organisation.