Six diagnostic products that operate above your data team and your ERP — in the six structural gaps neither can reach.
20 years of cross-border experience, 5 languages. I know these situations from the inside — growth, MBO/MBI, buy-and-build and the pre-transaction window: where hidden value, risk and waste sit, and how to make them visible with source-checked figures before anyone uses them against you. I work on your file myself, no intermediary layer.
From growth to a grip on margin, cash and enterprise value.
The diagnostics work at every phase — ongoing operations, growth, MBO, MBI, buy-and-build and exit — and serve owners as well as acquirers.
The diagnostics apply at every phase — this is simply where the stakes run highest.
Pre-transaction is the window where what you find can still be repriced. Once due diligence opens, valuations crystallise — and the same findings become deductions applied by the buyer, not by you.
| What your data team does | Where GLO25 operates |
|---|---|
| Answers the stated question | Detects whether the stated question is the right one |
| Finds correlations across hundreds of variables | Isolates the single causal binding constraint |
| Extrapolates historical patterns forward | Reverse-engineers from the desired end-state back |
| Works with tables and time series | Works with entity-relation semantics (knowledge graph) |
| Analyses internal data | Detects external signals (BORME, KVK, M&A, regulation) |
| Builds models | Audits models independently — AI Act / DORA / bias |
| Simulates historical scenarios (Monte Carlo) | Simulates multi-agent behaviour with individual agency per stakeholder |
Products and instruments are one team: the products are what you receive, the instruments are the engine that finds it.
Boeing selected Baan in 1994 for 757 production. Before implementation, all processes were documented in flowcharts and best-practice templates were applied.
Result: stock-outs fell from ~350 per day to ~10. 25,000 employees consolidated into one system. Implementation time: half that of comparable competitors.
Lidl started project eLWIS in 2011 to replace its legacy system with SAP for Retail. A fundamental conflict went undetected: Lidl valued inventory at purchase price; SAP standard used selling price.
Final cost: ~€500M. Project cancelled 2018. Seven years without process modernisation.
The structural mismatch between physical flows and financial flows is not industry-specific. It appears wherever a physical unit — a pallet, a tonne, an hour, a rented asset — moves through multiple parties and generates a billing event at each node. The three conditions below determine whether the diagnostic applies. When all three are present, the leakage is measurable from day one.
| NACE | Sector | Structural leakage trigger |
|---|---|---|
| E38 | Waste management & recycling Afvalbeheer / Gestión de residuos / Abfallwirtschaft | Weight at intake ≠ weight at processing ≠ invoiced tonnage. Three parties, three measurement moments, structural divergence at €/tonne level. |
| N77 | Rental of moveable assets Verhuur / Alquiler de bienes muebles / Vermietung beweglicher Güter | Circular pooling model: asset leaves, is used, returns — return does not automatically trigger correct credit. Every unmatched return is an open balance. Scaffolding, machinery, industrial tooling, medical equipment, pallet pooling. |
| M69–71 | Professional services Accountancy / Legal / Technical advisory | The hour is the physical unit. At mergers and integrations, structural hour loss occurs: hours double-booked, unbilled, or misattributed. GLO25’s Hidden Task Detector quantifies simultaneous handling per client file. On average 12–18% of billable hours disappears during integration phases. |
| G46.46 | Pharmaceutical distribution Pharma wholesale / Hospital & pharmacy supply | Multi-party chain manufacturer → wholesaler → pharmacy / hospital with serialised unit tracking (FMD / EU 2016/161). Cold-chain breaks, expired-stock returns, batch recalls and controlled-substance ledgers each generate billing or credit events that systematically diverge from the physical movement. Reimbursement deductions from payors compound the gap. |
| C25 + C28 | Industrial manufacturing & machinery Mittelstand / Maakindustrie / Manufactura industrial | Multi-tier supply chain (sub-supplier → tier-1 → OEM → end-user → service & parts) with a billing event at each transfer. Warranty and RMA reconciliation: a returned defective unit rarely matches automatically against the warranty credit, the supplier recovery and the field-service work order — the four ledgers diverge by 4–9% of net revenue. Supplier-rebate accruals are double-handled across procurement, finance and BU controllers. Project-margin vs ERP-margin gaps on long-cycle orders are visible from day one of the diagnostic. |
| H49.4 + H50 | Road & maritime freight Land & sea transport / Logística terrestre y marítima | Loaded weight, weighbridge weight and invoiced weight diverge structurally on land. At sea, demurrage, container return charges and discharge-vs-loading weight mismatches compound the same gap. On multi-leg routes the mismatch accumulates per link in the chain. |
| H52.1 + H53 | Warehousing, 3PL & express logistics Opslag, distributie & koerier / Almacenamiento y mensajería | WMS stock count ≠ ERP invoice line. Same billing unit as asset pooling (unit × day). Acute at third-party logistics operators with multiple clients per site, and at courier networks where the physical scan stream ≠ invoice line on volume discounts, returns and B2B settlement — with duplicate claim handling on damage and loss. |
| G46 | Wholesale trade (non-pharma) Groothandel / Comercio mayorista / Großhandel | Return credit notes for rejected or damaged goods systematically unreconciled at multi-item partial returns. Credit note stays open; nobody owns it. |
| NACE | Sector | Adaptation required |
|---|---|---|
| F41–43 | Construction & infrastructure | Material flow per project vs progress billing. Variations and rejection are the leakage source. Project-level rather than transaction-level analysis. |
| C10–12 | Food & beverage processing | Weight loss in processing, shelf-life returns, by-product settlement. Leontief matrix applicable to raw material balance. |
| C20 | Chemical manufacturing | Batch-level billing, yield reconciliation, lot tracking. High per-flow complexity but methodologically direct. |
| C21 | Pharmaceutical manufacturing | Batch yield reconciliation, QP release, lot-level tracking and serialisation chain. High GMP and regulatory complexity, but the input-output matrix maps cleanly to bill-of-materials and yield variance — methodologically direct once the dossier structure is mirrored. |
| K65 + K66.2 | Insurance carriers & brokerage | Claims handling is the textbook reconciliation loop — physical loss event ≠ paid claim ≠ salvage recovery ≠ reinsurance recoverable, four parties with a billing event at each node. The same file is touched simultaneously by claims, underwriting, A/R, compliance and recoveries. Adaptation needed because the operating spine is Guidewire / Sapiens / Tia, not Baan — the Leontief matrix maps onto the claim-flow ledger rather than the BoM. |
| K64.91 + K64.92 | Specialty finance & leasing | Multi-party servicing chain: originator → servicer → SPV → investor → borrower. Reconciliation gaps between servicer reports and SPV ledgers cluster around fees, advances, recoveries and modifications — each generates a billing or accrual event that does not flow through cleanly. Hidden duplicate handling in collections, restructuring and waterfall accounting. Adaptation: SPV-level Leontief structure with separate residual / interest / fee strands. |
| G47 | Retail | Reverse logistics: physical return ≠ automatic credit. Scalable at chains with centralised returns processing. |
| D35 | Energy production & distribution | Meter reading vs consumption vs billing. Settlement lag in smart metering and grid balancing. High regulatory complexity. |
The five operational pain points that show up most in food & process manufacturing (NACE C10) — and where EBITDA leaks or can be protected. This is a proof slice: one sector now, the rest of the series to follow. Every figure below is labelled for how we know it — a measured source, or a modelled estimate.
In food & process manufacturing (NACE C10) the median EBITDA margin among SEC filers is 4.5%. GLO25 models a +153 bps median (p50) uplift toward roughly 6.0% — with a +202 bps ceiling — by closing five operational gaps: supply-chain blindness, batch traceability, retirement knowledge loss, compliance drag and M&A integration. The baseline is measured; the uplift is modelled, not guaranteed.
EBITDA margin = EBITDA ÷ RevenueEBITDA margin — operating profitability as a % of salesEBITDA — operating earnings before interest, tax, depreciation, amortisationRevenue — net sales over the same periodModelled target margin = baseline margin + modelled uplift (in basis points). 100 bps = 1 percentage point. Driver rates and input distributions stay under the IP line.
| When this applies | When it doesn’t |
|---|---|
|
|
Take a food processor at the sector baseline of 4.5% EBITDA. Closing the ranked gaps models a median (p50) uplift of +153 bps toward roughly 6.0%, with a +202 bps ceiling. Assumptions: SEC-filer sector median, gaps actually closed, one plant’s own baseline will differ. Baseline is a direct source; the uplift is a GLO25 model output, not a guarantee.
Five steps, shown as a swimlane: what you see, and what stays under the IP line. We put the shape and the discipline of the method in daylight — the engine itself stays under the hood. There is one step where we hold nothing back: step three, how every figure is labelled.
GLO25’s method runs five steps: source from authority data (SEC and CNAE filings), quantify down to EBITDA, simulate the uncertainty with Monte Carlo, label every figure as a direct source or a derivation, and name the single binding constraint to fix first. The proprietary engine stays under an IP line; the shape of the method and the source-labelling discipline are fully visible.
We start at authority sources — SEC filings, CNAE sector data, public financial records. Nothing begins with a guess; every input has a documented origin before we touch it.
We work the sector economics all the way down to EBITDA level — from broad sector figures to what actually lands on the margin line.
Every uncertain driver is a range, not a single guess. We run thousands of scenarios across those ranges and report the percentiles: p50, the median — the honest expected outcome — and a ceiling, a high percentile. That is why our figures read "+153 bps (p50), ceiling +202 bps" — never "you will win X%".
This is where GLO25 is different. Every number we hand you is labelled by how we know it — a direct source, or a derivation. Nothing hides behind a confident tone. At this step the transparency is the point: you can see exactly which figures are measured and which are modelled, and weigh them accordingly.
We translate the sector picture into what it means for your margin — your baseline, your gap, your exposure — not a generic benchmark.
You get one thing to fix first — the single binding constraint — and the route toward it. Not a forty-point list; the one move that unlocks the rest.
Reported uplift = { p50, ceiling } from N simulated scenariosp50 — median outcome — half the scenarios land above, half belowceiling — a high percentile — favourable but plausibleN — thousands of scenarios across each driver’s rangeWe report percentiles, never a single point estimate. The input distributions and driver rates stay under the IP line.
| When the method fits | When it doesn’t |
|---|---|
|
|
For food & process manufacturing we source a 4.5% baseline (direct source), quantify to EBITDA, simulate, and report +153 bps (p50) with a +202 bps ceiling (derivation). Every figure carries its stamp. Assumption: sector-level median — your own numbers move the result.
Every sector below carries a baseline EBITDA drawn from SEC section-level data, and the one pain point where the margin quietly hides. Food is worked out in full — as the example of how deep the diagnostic goes.
GLO25 quantifies twelve mid-market sectors, each with a SEC section-level EBITDA baseline and the one operational pain point where margin hides. Food & process manufacturing (NACE C10) is worked out in full: a measured 4.5% baseline and a modelled +153 bps (p50) uplift toward roughly 6.0%, ceiling +202 bps. Baselines are direct sources; uplifts are modelled, not guaranteed.
Baseline 4.5% EBITDA. Our model lifts that by +153 bps (p50) toward ~6.0%, with a ceiling of +202 bps.
Sector target margin = SEC baseline + modelled uplift (bps)SEC baseline — measured sector EBITDA margin (direct source)modelled uplift — p50 estimate of margin recoverable by closing the sector’s pain pointOnly food is modelled end-to-end today; the other eleven show a measured baseline and the pain point. Driver internals stay under the IP line.
| Worked out in full | Baseline-only today |
|---|---|
|
|
Food & process manufacturing: baseline 4.5% (SEC, direct source) + modelled +153 bps (p50) → roughly 6.0%, ceiling +202 bps (derivation). Assumption: section-level median — a single company’s baseline and uplift will differ.