Skip to content

Methodology & data index

decivo MVP-Kosten-Index2026.06

We don't estimate prices, we estimate complexity — and convert it transparently into three procurement paths. 1 complexity point (CP) = 4.0 productive developer hours. Market ranges are aggregated from published agency prices and calibrated against real fixed-price projects. Dataset free to use under CC BY 4.0.

1 · Complexity points instead of prices

Money is a derived quantity. The source quantity is the complexity point: 1 CP = 4.0 productive developer hours (implementation + test, excluding overhead). This decouples the model from hourly rates — market prices change, the feature catalogue stays stable. There is deliberately no synergy discount: traceability beats false precision.

Complexity bands

BandCPLabel
S0–30Schlankes Validierungs-MVP
M31–65Vollwertiges Web-App-MVP
L66–110Komplexes Produkt-MVP
XL≥ 111Plattform-Scope (kein MVP mehr)

2 · Calibration

Reference scope “web-app MVP” (foundation, auth, social login, CRUD, file upload, admin, transactional emails, onboarding, GDPR) = 48 CP.

  • Classic path: 48 × 4.0 h × 1.30 overhead = 249.6 h × €80–140/h → €19,968–34,944 (median €23,712). Matches three independent market sources → externally validated.
  • AI-native path: the same scope reaches ~16 effective CP via the leverage matrix → below the fixed-price floor, output €12,500 → matches the real decivo entry price → internally validated (Endelevu: ~10–14 h of agent time instead of ~422 h team-equivalent).
  • The spread (classic median €23,712 vs. AI-native €12,500, −47%) reproduces the “€50–80k → €15–30k” thesis for larger scopes. This is the index headline figure.

3 · AI leverage matrix

Core differentiator of this calculator: AI doesn't accelerate everything equally. With parallel coding agents, standardisable work (UI, CRUD, scaffolding) drops to 25–30% of the classic effort, security- and compliance-critical work only to ~82%, because human review dominates. Recalibrated Q3/2026 against real decivo projects, updated quarterly.

CategoryEffective effort vs. classic
Foundation25 %
UX & design25 %
Data & content30 %
Admin & analytics30 %
Accounts & access38 %
Communication40 %
AI features48 %
Payments50 %
Real-time50 %
Integrations55 %
Architecture60 %
Mobile65 %
Compliance82 %

4 · Formulas

  • Classic: Stunden = CP × 4 × 1.3; Cost = Stunden × 80/95/140 €/h (minimum 8000 €).
  • AI-nativ: effCP = Σ (CP × Leverage); hours = effCP × 4 × 1.25; price = max(12500 €, round50(effStunden × 115 €)).
  • DIY: founder hours = CP × 9,0/6,0/3,5 (by skill level); rework risk = P(Band) × 0,8 × AI-Festpreis.

The AI-native fixed price is delivered in decivo packages (net): Code Prototype €12,500 as the base, then iterative prototype loops at €4,500 each to the MVP/final product. Optionally upstream: Innovation Workshop €7,500 (incl. clickable prototype), Clickable Prototype €3,500, UX validation loop €1,350 per sprint.

5 · Market index by project type

Market ranges (p25/median/p75) aggregated from published prices of German development agencies, as of June 2026. Sources linked per row.

Project typep25Medianp75Sources
Landing page MVP / smoke test5.000 €8.000 €15.000 €tj-labs.de
Web app MVP20.000 €27.500 €40.000 €codeguides.degroenewold-it.solutionseasy.bi
B2B SaaS MVP35.000 €55.000 €90.000 €codeguides.de
Two-sided marketplace MVP40.000 €60.000 €100.000 €easy.bi
AI product MVP30.000 €50.000 €120.000 €easy.bi
Booking platform20.000 €30.000 €50.000 €groenewold-it.solutions
Internal tool12.000 €20.000 €35.000 €groenewold-it.solutions
Native mobile app MVP25.000 €40.000 €80.000 €codeguides.detj-labs.de

First-party data points from real decivo fixed-price projects will be added with the next quarterly update.

6 · Download the dataset

CC BY 4.0 — Nutzung mit Quellenangabe 'decivo MVP-Kosten-Index' und Link ausdrücklich erwünscht