Screening Studio
operating-performance screen

Synergy signal engine

Directional screening signal, not a forecast, valuation, or guarantee of realized synergy.

examples
derived from current inputs
indicator contribution
Why the hero index reads this way
An explanation of the deterministic rule, not causal evidence. Weights are a documented rule-of-thumb, not fitted coefficients.
sensitivity
Which levers move the index
One-at-a-time changes to the current deal, where the signal is sensitive, not an attainable optimum.
evidence

What the model actually shows.

Out-of-sample metrics from the extension-suite handoff CSVs. R² values sit on different target scales and are never read as cross-target improvements.

target family
Benchmark context
asset-turnover validation
Mean reversion & stability
Dropping reversion-prone level features collapses R² toward the naive baseline.
industry concentration
Target-side SIC2 diagnostic
Four large bins cover 219 of 597 test deals; only chemicals shows positive within-bin R².
Read correctly: the asset-turnover model leans on reversion-prone utilisation features (≈ 30% of |SHAP|). That is why the headline result is framed as learnable asset-productivity drift, not proven synergy creation.
sme transfer

The owner-company overlay.

A theory-backed decision-support overlay that re-weights the listed-firm channel emphasis for private-company deal features. It is not trained on SME outcomes.

This is not an empirically validated SME prediction model. ORBIS evidence supports feature-computability only.
examples
acquiror sme
Private-company inputs €m
target sme
Private-company inputs €m
context
Deal profile
ownership
acquisition experience
illustrative realisation tilt · not validated on SME outcomes
Channel-realisation overlay
1.0x
illustrative · not validated
This multiplier is a theory-based re-weighting, not trained on or validated against any SME outcomes (ORBIS supports feature-computability only). Read the direction to structure the integration conversation; do not treat the number as an SME synergy estimate.
adjusted channel emphasis
Where realisation shifts
active reasoning
Why it tilts
about

Built at the intersection of M&A, machine learning, and decision support.

A public portfolio artifact showing how thesis research can become an interactive, bounded analytical product.

about the builder

Yanick Annema

I hold an MSc in Financial Engineering & Management from the University of Twente and work in M&A and Corporate Finance at Moore MKW. My work sits where applied machine learning meets finance: interpretable models, honest validation, and tools that make quantitative evidence usable in real advisory conversations.

This dashboard is based on my MSc thesis on whether pre-deal information can predict post-merger operating performance. It turns the research into a compact screening studio for exploring candidate deals, model evidence, attribution, and SME transfer limits.

what this is

A screening workbench that ranks M&A candidates with a transparent operating-performance rule, shown alongside separate out-of-sample evidence from a model learned on listed-firm deals, with a feature-computability gate for SME transfer.

what you enter

Acquiror and target fundamentals you already hold early-stage (revenue, EBIT, assets, balance-sheet items), plus deal context.

what you get

Target-family indicators with transparent contribution, out-of-sample model evidence per target, and an SME transfer overlay.

what it is not

Not a valuation, not a synergy guarantee, not a forecast value, and not an empirically validated SME prediction model.

orbis computability
SME feature-computability boundary
ORBIS supports constructing selected accounting features for revenue-reporting SMEs. It supplies no labelled SME post-merger outcomes; it gates feasibility, not validation.
10
directly computable
5
proxy-only
9
unavailable
2
comparability risk

Headline: asset turnover and EBIT/assets are computable for revenue-reporting SMEs; CFROA is proxy-only; retained earnings, cash, CAPEX and all deal-level controls are unavailable in ordinary ORBIS accounts.

supports
What you can claim
  • Ranking deal candidates by a transparent operating-performance screening rule.
  • Asset-turnover change is the most learnable target in the family.
  • Reporting each target's out-of-sample model evidence (R², ρ, n).
  • Explaining model attribution via SHAP on listed-firm test data.
  • Mapping which SME features are computable from ORBIS.
does not support
What you cannot claim
  • A forecast value, valuation, or guaranteed realized synergy.
  • Causal attribution: SHAP and contribution bars are explanations, not mechanisms.
  • Cross-target R² improvement claims (target scales differ).
  • Empirically validated SME prediction.
  • Synergy creation from asset-turnover drift (it partly reflects asset-productivity reversion).
Index vs. model: the live screening index is a deterministic rule over your inputs. The R²/ρ figures are separate out-of-sample results from the trained model. The app keeps them visually and conceptually distinct on purpose.