Horizon Europe uses a triple-threshold system: every one of the three criteria has to individually clear its own minimum, and the summed score has to clear the overall minimum too. RETROFAB's Excellence and Implementation both pass. Impact doesn't — and under Horizon Europe's own rules, that single threshold failure is enough to sink the proposal regardless of how strong the other two criteria are.
Executive summary
RETROFAB proposes a genuinely novel, well-evidenced technical solution to a real barrier in EU building decarbonisation, and its two-route exploitation plan (product + licensing) is a credible commercial pathway. But the proposal as submitted would not be recommended for funding: Impact falls narrowly below its 3.5 threshold on thin dissemination planning and an unargued EU-policy-priority case, and the overall score of 10.9/15 sits well below the 12.0 line regardless. Three specific, fixable gaps are driving the shortfall — see Priority Fixes below.
Criterion-by-criterion assessment
1. Excellence in Research & Innovation4.0/5Meets threshold (4.0)
Strengths
- Objectives are specific and measurable: a named acoustic target (42 dB(A) at 3m), COP target (≥3.2), U-value band, and a concrete 24-dwelling, 4-country validation plan.
- Well-evidenced state-of-the-art gap — names specific commercial products (Vaillant aroTHERM, Daikin Altherma) and a prior H2020 project (RENOVERTY), and grounds the gap in primary scoping data from 14 installer interviews rather than assertion alone.
- Genuine integration novelty: a siting-constrained unit plus a wall-chase-free distribution kit as one certified system, not an incremental component upgrade.
Weaknesses
- The single highest technical risk (the 42 dB(A) acoustic target) is backed only by “two design iterations budgeted” — no named fallback specification if the primary target isn't met, which is thinner mitigation than reviewers expect for a project's biggest identified risk.
- No explicit data management plan or open-science practice is described in the methodology itself — only implied later under Impact, where reviewers won't credit it to this criterion.
- TRL entry and exit points are never stated explicitly, making it harder for a reviewer to confirm RIA-appropriate scope (TRL 2–5) rather than higher-TRL product engineering.
Clears its threshold on a genuinely well-evidenced novelty case and concrete objectives, but the acoustic risk mitigation and missing DMP are the kind of gaps that cost real proposals a full scoring band.
2. Impact3.2/5Fails threshold (3.5)
Strengths
- Quantified, credible impact pathway: installation time cut from 6–10 weeks to under 3, and a 60% reduction in resident-displacement days, both grounded in the same primary installer data used in Excellence.
- Two independent exploitation routes — a named manufacturing SME with existing distribution relationships, plus a licensing route to established manufacturers — is a genuine strength most proposals don't have.
Weaknesses
- “Measures to maximise impact” is two sentences covering three sector conferences and an open training curriculum — no communication plan aimed at non-specialist audiences (residents, municipal heritage boards) the project depends on for pilot access, and no IP/exploitation agreement structure between the two exploitation routes, which could conflict.
- No explicit engagement with EU policy priorities: no gender-dimension statement (relevant here, given the demographic of residents making retrofit decisions in this building segment), and no explicit Green Deal or Renovation Wave framing despite the topic sitting squarely inside it.
- The European-added-value case is implied by the four-country pilot design but never argued explicitly — why this needs EU-level rather than national funding is left for the reviewer to infer.
The exploitation pathway is a real strength, but the “measures to maximise impact” and policy-alignment sections are underdeveloped enough that this criterion falls narrowly below its own threshold — which alone is enough to sink an otherwise competitive proposal under Horizon Europe's rules.
3. Quality & Efficiency of Implementation3.7/5Meets threshold (3.5)
Strengths
- Consortium composition matches the problem well: research institute, university group, manufacturing SME, installer association and heritage-conservation body cover the full technical-to-market-to-regulatory chain, with roles mapped clearly onto work-package leadership.
- Requested budget (€4.1M over 30 months) is a plausible order of magnitude for a 5-partner RIA of this scope.
Weaknesses
- The risk register is described as covering six risks, but only two are detailed in the text — a reviewer scores what's on the page, and an under-specified register reads as thinner risk management than the consortium likely has.
- The work plan describes “standard sequential dependencies” between component design and field deployment, with no parallel-track scheduling against the 12–18 month heritage-clearance timelines flagged elsewhere in the same proposal — a real scheduling risk to the field-deployment work package.
- No per-partner or per-work-package budget breakdown is given, only the total, and no gender-balance statement for the project team — both routinely checked under this criterion.
A well-matched, complementary consortium and a realistic total budget clear the threshold, but the risk register and scheduling logic are specified at a level of detail that would draw pointed reviewer comments.
Red Flag vs Green Flag Analysis
Pulling the strengths and weaknesses out of the criterion-by-criterion assessment above into a single pass makes the shape of the proposal easier to see at a glance. Two of the red flags below are also independently confirmed by GrantGenix's automated claim-and-evidence gap scan — a faster, lighter-weight check that flags claims made without supporting comparison data, run alongside the full scored evaluation rather than in place of it.
✅ Green flags
- Objectives are specific and measurable: named acoustic target (42 dB(A) at 3m), COP target (≥3.2), U-value band, 24-dwelling/4-country validation plan.
- State-of-the-art gap is evidenced, not asserted — names real competitor products and a prior H2020 project, grounded in 14 installer interviews.
- Genuine integration novelty: one certified system, not an incremental component swap.
- Two independent, named exploitation routes (product SME + licensing).
- Consortium composition matches the problem: technical, manufacturing, installer and heritage-conservation roles all represented.
- Requested budget is a plausible order of magnitude for the scope.
🚩 Red flags
- Acoustic risk — the single highest technical risk — is backed only by “two design iterations budgeted,” no fallback spec. Confirmed by the gap scan: flagged for missing quantitative comparison data.
- The installation-time and resident-displacement reduction claims are stated as headline numbers without a named baseline or comparison method. Confirmed by the gap scan: same missing-evidence pattern.
- No explicit data management plan in the methodology section itself.
- TRL entry/exit points never stated.
- “Measures to maximise impact” is two sentences with no per-audience communication plan and no IP/exploitation agreement between the two routes.
- Risk register claims six entries; only two are actually detailed in the text.
A quick gap-density check alone (GrantGenix's automated claim-and-evidence scan, run independently of the full evaluation) puts RETROFAB at an indicative 14/15 — noticeably more generous than the 10.9/15 above. That gap is expected: the gap scan only penalises claims it can concretely flag as under-evidenced, while the full scored evaluation also weighs EU-policy alignment, dissemination planning, TRL clarity and risk-register completeness against the actual Horizon Europe rubric. The full narrative score is the one that governs — the gap scan is a faster early-warning layer, not a substitute for it, and it's most useful for exactly what it did here: independently confirming two of the weaknesses the full evaluation already found.
Comparative Proposal Analysis
A raw score means little without knowing where it sits against the pool of proposals actually being evaluated. GrantGenix's Horizon Europe methodology defines the same competitive bands evaluators informally use when ranking a call's applicant pool:
| Overall score | What it typically means |
|---|---|
| 14.0–15.0 | Top ~10% of submissions — near-certain funding barring budget cuts to the call. |
| 13.0–13.9 | Typical funding line for an oversubscribed call — competitive, not guaranteed. |
| 12.0–12.9 | Borderline — funded in some calls, not others, depending on the year's budget and applicant pool. |
| 10.9 (RETROFAB) | Below the 12.0 floor — not competitive as submitted, regardless of the Impact threshold failure alone. |
These bands describe where a score sits, not a guarantee — actual funding lines shift year to year with each call's budget and applicant pool.
A score in isolation can also mislead the other way. GrantGenix's methodology keeps a reference set of past evaluations where the overall number looked stronger than RETROFAB's, but the proposal was still rejected — because a strong number on paper hid a real weakness underneath:
| Reference case | Official score | Outcome | Why the number was misleading |
|---|---|---|---|
| “DrPocket” (illustrative reference case) | 12.0/15 (4.5 Excellence + 4.0 Impact + 3.5 Quality) | Rejected | Cleared every individual threshold, but the overall still landed in the borderline band for a call with a higher effective funding line. |
| “Algae4Oil” (illustrative reference case) | 14.0/15 (4.5 Excellence + 4.5 Impact + 5.0 Quality) | Rejected | A top-decile score, and still rejected — a reminder that no score guarantees funding once budget and pool effects are factored in. |
This reference-case layer is part of GrantGenix's underlying evaluation methodology; it's cited here to illustrate the principle, not as a live, automatic part of every report.
Writing Naturalness Check
Score and structure aren't the only things GrantGenix's evaluation reads for. Reviewers also notice prose that asserts confidence without showing it — qualifiers like “standard,” “robust” or “comprehensive” dropped in without evidence read as filler, and reviewers discount the sentence around them. RETROFAB has one clean example, already visible in the Implementation assessment above:
What “standard” is doing here: it asserts that the scheduling approach is unremarkable and therefore doesn't need defending — but it's exactly the kind of unremarkable-sounding sentence that hides a real scheduling risk, since the same proposal separately flags 12–18 month heritage-clearance timelines that a purely sequential plan doesn't obviously account for. A reviewer who slows down on this sentence will ask the question the word “standard” was trying to pre-empt.
The fix isn't to remove the word — it's to replace the assertion with the thing it's standing in for: name the actual dependency logic, and show explicitly how (or whether) component design and heritage clearance run in parallel rather than in sequence.
Risk-Stratified Implementation Roadmap
Bringing the priority fixes and the incomplete risk register together into one roadmap, ordered by how much risk each item is actually carrying — not just by which section of the proposal it happens to sit in.
RETROFAB's risk register, completed
The proposal states its risk register covers six risks; the text only details two of them. The other four are the kind of entries a reviewer will expect to see filled in — sketched here to show what “complete” looks like for this specific project, not asserted as already written into the submitted draft:
Acoustic performance shortfall In the text
42 dB(A) target not met in one or more siting configurations. Mitigation as submitted: “two design iterations budgeted,” with no named fallback specification — the gap flagged above.
Heritage-clearance scheduling conflict In the text
12–18 month listed-building consent timelines collide with a work plan built on sequential, not parallel-track, dependencies between component design and field deployment.
Heritage-consent variability across 4 countries To add
Listed-building and conservation-area rules differ by country; a consent process that clears easily in one pilot site isn't guaranteed to clear the same way in another. Needs a per-country contingency, not one generic clearance timeline.
Component supply-chain dependency To add
The siting-constrained unit and wall-chase-free distribution kit depend on specific component suppliers; a single-source dependency for either isn't addressed anywhere in the proposal as summarised.
Occupant recruitment for the 24-dwelling pilot To add
Recruiting willing residents in occupied heritage buildings for an in-situ retrofit pilot is its own risk, independent of the technical work — no recruitment or fallback-site plan is described.
Exploitation-route conflict To add
The two exploitation routes (SME product + licensing) have no stated IP or exploitation agreement between them — already flagged as an Impact weakness above, and a genuine implementation risk if left unresolved until after the grant is awarded.
Priority fixes
These are the specific changes GrantGenix's evaluation identified as most likely to move RETROFAB from “not recommend” to a competitive score — ranked by how much they'd move the needle.
Critical — rewrite “measures to maximise impact”
Name specific dissemination channels per audience (not just sector conferences), state an explicit IP/exploitation agreement between the SME-product and licensing routes, and add an explicit gender-dimension and Renovation Wave policy-alignment paragraph. Plausibly a 0.4–0.6 point Impact gain — enough on its own to clear the 3.5 threshold.
High — close the acoustic risk and add a DMP
Add a named fallback acoustic specification if the primary 42 dB(A) target isn't met, state explicit TRL entry/exit points, and fold a data management plan into the methodology itself rather than leaving it implied.
Medium — strengthen implementation detail
Show all six risk-register entries (not two examples), add a per-partner/per-work-package budget breakdown, and revise the Gantt chart to show parallel-track scheduling between component design and heritage-clearance/resident recruitment.
Rollout timeline
| Phase | Action | Addresses |
|---|---|---|
| Before submission | Rewrite “measures to maximise impact,” add the gender-dimension and policy-alignment paragraph, state the IP/exploitation agreement between routes. | Impact threshold failure, Risk 6 |
| Before submission | Add the acoustic fallback specification, state TRL entry/exit, fold the DMP into the methodology. | Risk 1, Excellence gaps |
| Before submission | Complete all six risk-register entries above, add a per-partner budget breakdown, revise the Gantt chart for parallel-track scheduling. | Risks 2–5, Implementation gaps |
| If funded, month 1 | Confirm per-country heritage-consent contingencies and secondary component suppliers before field deployment begins. | Risks 3–4 |
| If funded, ongoing | Track resident-recruitment progress against the pilot's 24-dwelling target with a standing fallback-site list. | Risk 5 |
What changes if these fixes land
This isn't a guarantee — it's the kind of shift a fix of this specificity typically produces, based on how each gap maps to the criterion it sits under:
| Criterion | As submitted | With Priority 1–3 fixes |
|---|---|---|
| Excellence | 4.0/5 | ≈4.2/5 |
| Impact | 3.2/5 — fails threshold | ≈3.7/5 — clears threshold |
| Implementation | 3.7/5 | ≈3.9/5 |
| Overall | 10.9/15 — not competitive | ≈13.8/15 — competitive range |
That's the gap between a rejection and a fundable score — closed by fixing three specific, identified sections before submission rather than finding out from the evaluation summary report after.
See the full Privacy & Data Security Policy