In public perception, the funding application is the central artefact of the funding business. Over the life of a project, though, it is the smallest block of work. The empirical data — from the Federal Demonstration Partnership's Faculty Burden Survey[1], the European Commission's Horizon Europe interim evaluation[3] and the annual reports of the European Court of Auditors[5] — point in a direction that contradicts the picture of classic consultancy: around 20 per cent of the total effort on a funded project arises before the grant, around 80 per cent after it. This piece places the figures in context and describes why the structural imbalance between pre-award and post-award marks the dividing line between a consultancy fee and running a platform.
Pre-award vs post-award
US funding administration divides a project's life cycle consistently into two phases: pre-award covers everything before the grant notice — scouting programmes, delimiting the project, writing the application, costing the budget, consortium agreements, submission; post-award begins with the notice and covers the project start, drawing down funds, ongoing reporting, amendment requests, audit and final use of funds. European practice knows the same split under other names: the "application phase" and the "grant management phase" in the European Commission's Annotated Grant Agreement[6].
Operationally, the two phases differ not only in duration but in risk profile. Pre-award is a one-off effort with a clearly delimited outcome: the application is submitted and either approved or rejected. Post-award stretches across the whole project term — typically 36 to 48 months in Horizon Europe — plus a five-year review period after the last payment run, during which the European Commission and the European Court of Auditors may exercise their audit rights under Article 25 of the Annotated Grant Agreement[6]. For a grant notice issued in 2026, the documentation obligation therefore reaches into the early 2030s.
This ratio is no peculiarity of the European Union. The OECD's Frascati Manual names systematic data capture, attribution to people and projects, and the delimitation of research and development from product-related activity as standard requirements for reporting in R&D statistics[7]. Anyone whose research is publicly financed has to show exactly what was financed — across the whole project term.
Where the 20 per cent figure comes from
The empirical basis for the 20/80 thesis comes from the Faculty Burden Survey of the Federal Demonstration Partnership, a cooperation founded in 1988 between ten US federal agencies and over 150 research institutions. The first survey, in 2005, covered 73 universities and 23,325 principal investigators; 6,081 valid responses were analysed[1]. The central finding: 42 per cent of the personal time available in a federally funded research project went into administrative work, 58 per cent into active research.
"Faculty devoted an average of 42 % of their research time to pre- and post-award administrative tasks associated with the federally funded portion of their workload."— FDP Faculty Burden Survey, Research Management Review 17 (1)[1]
The follow-up surveys of 2012 and 2018 confirm the order of magnitude: 11,167 principal investigators from 111 institutions took part in the 2018 wave, and the administrative share stayed in the region of 44 per cent of working time[2]. That administrative time is not evenly distributed: pre-award activity — writing the application, the budget, approvals — accounts for around a fifth of it, and post-award work such as procurement, personnel administration, reporting, audit preparation and compliance for the remaining four fifths.
Add the figures up and the 20/80 profile emerges: of 100 person-hours a scientific lead invests in a funded project, around 58 go into research and around 42 into administration. Of those 42 administrative hours, roughly 8 to 9 fall on pre-award — the application writing phase in the narrow sense. The remaining 33 to 34 are spread across post-award. As a share of the total administrative effort, that is a ratio of about one fifth to four fifths.
This US finding carries over to the European framework, because the underlying processes are structurally identical. The European Commission's Horizon Europe interim evaluation of 30 April 2025 (SWD(2025) 110 final) puts recipients' administrative costs, based on a survey of participants, at 6 to 10 per cent of the project budget on average; 48 per cent of respondents report figures above 10 per cent, and 10 per cent report figures above 20 per cent[3]. That order of magnitude applies not to writing the application alone but to the whole administrative effort, reporting and audit included.
The 80 per cent in detail
What actually fills the 80 per cent? The European Commission's Annotated Grant Agreement, in version V2.0 as updated on 1 April 2025, lists the post-award obligations systematically[6]. The main categories:
- Time recording per person, per project, per working day — in Horizon Europe, since the move to daily rates, with minimum time records in auditable form[6].
- Technical reports at fixed intervals (typically half-yearly to yearly), deliverables at agreed milestones, ad hoc evidence on request from the project officer.
- Financial statements per reporting period — broken down by cost category (personnel, travel, equipment, subcontracting, other goods and services), evidenced by invoices, payment records, personnel files, time records.
- A Certificate on the Financial Statements (CFS) by an independent auditor, from a cumulative own contribution of €430,000 per recipient.
- Amendment requests for substantive or financial changes — reallocation between cost categories, extension of the term, changes to the consortium — each with reasons, the Commission's agreement and an adjustment to the grant agreement.
- Retention of documents for five years after the final payment; the European Commission and the European Court of Auditors may initiate audits at any time within that period[6].
How error-prone this post-award apparatus is in practice is shown by the European Court of Auditors' 2024 annual report. Of 99 research projects audited — 65 under Horizon 2020, 34 under Horizon Europe — 24 showed errors in staff cost accounting. In 5 of 34 Horizon Europe projects the daily rate calculation was specifically objected to, and in three cases it led to incorrect cost claims[5]. Simplifying the staff cost rules between Horizon 2020 and Horizon Europe did not lower the error rate — the change produced a new class of errors.
For German recipients, the national level comes on top. The General Ancillary Provisions for Grants on a Cost Basis (ANBest-P-Kosten), the BMBF's ancillary provisions and points 6 to 8 of the administrative provisions to section 44 of the Federal Budget Code define interim reports, the proof of use, use in accordance with the purpose, duties to notify project changes, and the retention of auditable vouchers for at least five years. In its 2024 annual report, the DFG reports that it managed around 30,940 live projects with a funding volume of around €3.9bn[10]; each of these ties up scientific and administrative capacity continuously across its term.
The macroeconomic dimension of this effort is mapped in ifo Schnelldienst 11/2024: the direct bureaucratic costs of the German economy are put at around €65bn a year, and in the Mittelstand the working time tied to bureaucracy amounts to around seven per cent of total working time[8]. In publicly funded R&D the share is regularly higher, because of the additional grant-specific obligations.
What classic consultancy covers
Classic funding consultancy has built its business model around the pre-award section. In the form visible in the market, it offers essentially four services: programme matching (which instrument fits the project), writing the application (the narrative, the technical description, impact), costing the budget, and submission. The contractual arrangements are predominantly success-fee-based — a share of the grant goes to the consultant if the project is approved — and generally end with the grant notice or a few weeks after it.
In the 20/80 breakdown, that range of services covers exactly the pre-award area. Measured against the project's total effort, that is — on the FDP methodology — around a fifth of the administrative person-time across the whole term, and on a narrower definition (writing the application only, not building the consortium or starting the project) closer to ten per cent. Focusing on this service is rational for the consultant: it is clearly delimited, it delivers a verifiable outcome (approved or rejected), and it scales through standardising the architecture of applications.
The concentration on pre-award has a second cause in the economic incentives of the success fee. A consultant taking a share of the approved grant realises their revenue with the grant notice. Post-award work — reporting, compliance, audit preparation — extends the commitment, increases liability exposure and is unattractive to monetise from a consultant's point of view, because it neither triggers a further success event nor scales in a standardised way. A number of consultants therefore offer post-award support as a separate, time-based engagement — in practice only selectively, alongside the project, never as continuous responsibility for the process.
In its position work on reducing bureaucracy, the Stifterverband reaches the same conclusion: the effort of applying, reviewing and reporting along the funding chain is designed such that individual units — from a university chair to a mid-sized company — can "barely cope" with it any more, and it permanently diverts capable people from their actual work[11]. The consultancy sector does not solve this problem; it addresses only the moment of application.
What it does not cover
The structural gap opens immediately after the notice. Four areas are systematically outside classic pre-award consultancy.
First, ongoing reporting. Half-yearly or yearly technical and financial reports require current, auditable actuals on time recording, use of materials, progress against work package milestones, and deviations. This data has to be produced and validated continuously across the whole reporting period. A consultancy engagement that ends with the grant notice has access neither to the recipient's operational source systems nor to the day-to-day work of the people involved.
Second, amendment requests and consortium management. Almost every multi-year R&D project sees reallocations between cost categories, staff changes, extensions or changes to the consortium along the way. Each of these changes requires a formal request to the funding body or the Commission, a fresh assessment of the eligibility of the items, and — in EU consortia — the agreement of all partners. The Annotated Grant Agreement describes this process in its own articles; operational responsibility lies with the coordinator[6].
Third, audit preparation. The ECA error statistics show that staff cost accounting is the single largest source of objections[5]. An audit taking place five years after the final payment meets staff who have left the company, technical documentation sitting in superseded repositories, project management artefacts archived in systems that have since changed. Reconstructing auditable evidence is a block of work in its own right, one that by the time it falls due can typically no longer be handled by whoever wrote the application.
Fourth, checking cumulation across several programmes. Public funding is regulated by EU state aid law, by rules on non-cumulation and by programme-specific own-contribution requirements. In a company receiving the research allowance, ZIM, Horizon Europe and state funds at the same time, every single person-hour and every single item of expenditure has to be assigned to one funding instrument — double funding is not permitted and leads to recovery. This control runs permanently, not at the moment of application.
The empirical consequences are visible in the NIH data. Success rates for R01-equivalent grants have for years been in the range of 15 to 22 per cent for first submissions; the majority of applications later approved are resubmissions (A1), whose success rate is around 20 to 30 per cent[9]. That means: even the pre-award service is usually delivered more than once. The real value creation, though, lies not in the individual application but in the ability to manage approved funding in line with the rules across its term.
The Horizon Europe interim evaluation quantifies the effect on the recipients' side. 32 per cent of respondents rate applying under Horizon Europe as more burdensome than under Horizon 2020; almost half see no difference[3][4]. The simplification dividend from lump sums is estimated at a reduction in administrative costs of between 14 and 30 per cent — measured across the whole project life cycle, not the application[3]. The relief therefore arises primarily in the post-award section, not in the pre-award phase.
What this implies for how a platform works
The 20/80 breakdown has a clear consequence for process architecture. If 80 per cent of the effort arises after the notice, the responsible infrastructure has to keep working after the notice. That means: not consultancy alongside the project, but a permanently operated process environment that runs the whole life cycle of a project — from delimiting it to the end of the retention period — as one coherent system.
For work on a funding case, that yields four structural features. First, every relevant piece of data — time records, vouchers, technical progress, correspondence with the agency or the Commission — has to arise in the system and be attributed there. Reconstructing it afterwards from email archives and spreadsheets produces exactly the errors the European Court of Auditors documents in its annual reports[5]. Second, the state of every single project — including outstanding obligations, reports due and amendments pending — has to be retrievable at any time without further research; that is the core of the retention obligation under Article 25 of the Annotated Grant Agreement[6].
Third, the platform has to work across all funding instruments running at the same time. Checking cumulation, the delimitation under section 7 FZulG, state aid consistency between the GBER, the de minimis regulation and sector-specific frameworks can only be assured at constant quality if all of a company's projects are represented in the same data model. Fourth, the human in the loop has to be structurally anchored: funding law is not rule-based in the sense that every decision could be delegated to a machine; a BSFZ review guide, a Commission AGA annotation or the ECA's thresholds are texts to be interpreted. A platform automates the mechanical parts — capture, attribution, deadline management, assembly — and returns the interpretive parts to the human workflow for qualified review.
This architectural approach differs from the classic consultancy engagement in one fundamental respect. The engagement is a contractual relationship exchanging work for a fee, and it ends with the grant notice. A platform is an operating model that keeps the recipient, across the whole term — and across the retention period that follows — in a process that produces the evidence the Court of Auditors, the tax office or the funding body will ask for.
The data supports the logic. The FDP Faculty Burden Survey documents that the administrative burden stayed stable across three waves — no statistically significant change between 2005 and 2018[2]. The Horizon Europe interim evaluation shows that the European Commission's simplification instruments have so far not decisively relieved the cost structure on the recipient side[3]. The ifo Institute puts the bureaucratic costs of the German economy at a magnitude that means substantial opportunity costs for the Mittelstand[8]. Research institutions report the same thing in other words through the Stifterverband[11].
upsmart is built for this asymmetry. The platform runs pre-award (programme scouting, delimiting the project, application data, submission) and post-award (time recording, managing vouchers, reporting, amendments, audit preparation, retention) in the same data model. An application is not produced as a final text file but set up as a structure from which both the application documents and, later, the interim reports and proofs of use can be derived in line with the rules. The funding application is not the product but the starting point of a process running over years, which is decided in the 80 per cent — not in the 20.
- [1]Decker, R. S.; Wimsatt, L.; Trice, A. G.; Konstan, J. A. — The FDP Faculty Burden Survey (Research Management Review, vol. 17, no. 1, 2010)Federal Demonstration Partnership / Research Management Review (PMC) · 2010Open source
- [2]2018 Faculty Workload Survey — Primary Report (Federal Demonstration Partnership, Phase VI, released 2020)Federal Demonstration Partnership (thefdp.org) · 2020Open source
- [3]Schneider, S. L. et al. — Results of the 2018 FDP Faculty Workload Survey: Input for Optimizing Time on Active Research. FDP plenary presentation, January 2019 (11,167 principal investigators, 56,869 invited; administrative time share 44.3%).Federal Demonstration Partnership (thefdp.org) · 2019Open source
- [4]Commission Staff Working Document — Interim Evaluation of the Horizon Europe Framework Programme for Research and Innovation (2021–2024), SWD(2025) 110 final, 30 April 2025European Commission, EUR-Lex · 2025Open source
- [5]Communication COM(2025) 189 final — Horizon Europe: Research and Innovation at the heart of competitiveness (30 April 2025)European Commission, Directorate-General for Research and Innovation · 2025Open source
- [6]Annual Reports concerning the 2024 Financial Year — chapter 5, "Single Market, Innovation and Digital" (staff costs Horizon 2020 / Horizon Europe; findings on daily rate calculations and objections in 24 of 99 audited projects).European Court of Auditors (ECA) · 2025Open source
- [7]EU audit in brief 2024 — Introducing the 2024 annual reports (short version with chapter overview and key findings on "Single Market, Innovation and Digital").European Court of Auditors (ECA) · 2025Open source
- [8]EU Grants AGA — Annotated Grant Agreement, V2.0, 01.04.2025 (record-keeping Articles 20–22, audit Art. 25)European Commission, Funding & Tenders Portal · 2025Open source
- [9]Frascati Manual 2015 — Guidelines for Collecting and Reporting Data on Research and Experimental DevelopmentOrganisation for Economic Co-operation and Development (OECD) · 2015Open source
- [10]ifo Schnelldienst 11/2024 — The cost of bureaucracy (economic output lost to high bureaucratic effort)ifo Institute — Leibniz Institute for Economic Research · 2024Open source
- [11]NIH Data Book — Success Rates: R01-Equivalent and Research Project Grants by Institute/Center and Submission NumberNational Institutes of Health, Office of Extramural Research (RePORT) · 2024Open source
- [12]Annual report 2024 — tasks and results (chapter on funding activity, review and reporting processes)German Research Foundation (DFG) · 2025Open source
- [13]Ways out of the bureaucracy trap — innovation system, position paperStifterverband für die Deutsche Wissenschaft · 2023Open source
HITL in the application process: what the Dell'Acqua study at BCG measured
758 consultants, a pre-registered field experiment, +25.1% on speed and +40% on quality — and where the lever finds its limits.
The reuse lever: how an application platform pays for itself from cycle 2–3
NIH R01 resubmission data show the empirical advantage of the second submission. What that means for the roadmap of German funding applications.
Horizon Europe more burdensome than H2020 — the interim evaluation in detail
32% of participants find Horizon Europe more burdensome than H2020, 48% report admin costs above 10%. What the EU documents itself.
From the analysis into the application.
We show the platform on a real case.