Six systems are losing vendor support. For each, one export is enough to count what is actually in the estate — dead code, unsupported components, orphaned records — and price the exit as a floor. Counted, not estimated.
One document, same six parts, whichever system it is about. Page counts are from the sample report.
| Object type | Objects | Never executed | Share |
|---|---|---|---|
| REPORT | 900 | 270 | 30% |
| CLASS | 450 | 135 | 30% |
| FUNCTION | 300 | 129 | 43% |
| ENHANCEMENT | 180 | 63 | 35% |
| TABLE | 220 | 88 | 40% |
| INTERFACE | 150 | 45 | 30% |
| FORM | 200 | 85 | 43% |
| Total | 2,400 | 815 | 34% |
This is the whole argument of the product in one table: a third of the custom objects are never executed, and every dead object is scope somebody would otherwise pay a systems integrator to migrate.
Each finding in the report carries the same three things — the count, the filter that produces it, and the consequence. That is why a client's own consultant can check the work instead of trusting it.
Sample data, real method. Open the SAP sample
| System | Clock | The export | Price |
|---|---|---|---|
| Dynamics GP | 31 Dec 2029 | Table CSVs or a company backup | $2,500 |
| SAP ECC | 2027 maintenance | Object list + 12-month usage | $4,500 |
| BizTalk | 2028 mainstream | Bindings.xml | $2,500 |
| Gentran/Sterling | No vendor priority | 20–50 sample X12 documents | $3,500 |
| VMware | Per-core renewal | RVTools vInfo / vHost | $1,500 |
| Delphi / VB6 | Compiler dead 2008 | Source tree (.dpr, .pas, .dfm) | $2,500 |
Every number is produced by code reading the files you send. Nothing is estimated, sampled or inferred from a benchmark, so if a count looks wrong you can open the export and run the same filter.
The effort figure is a floor, not a quote. The coefficients are printed on its page, and the exclusions are named in the same block: functional design, data cleansing, testing cycles, licence change and training.
Each report ends with what the export cannot show — screen usage, hand-installed DLLs, stored procedures, contracts, or what people actually do at month-end — and asks for the specific artefact that would close the gap.
Naming the blind spots is the difference between an assessment and a sales document.
Two business days from the file arriving. If it is late for a reason on our side, the report is free.
The document, one revision included, and the machine-readable findings file. Yours to use and forward.
Files stay on a private volume, are never shared, and are deleted on request. An NDA on your paper is fine.
The counts are produced by code; the report is written and checked by a person. A tool cannot decide which finding matters or what the exclusions are, and those two sections are the ones a CFO reads.
No. Every report is built from an export you already know how to produce: table CSVs, an RVTools file, a bindings export, a folder of X12 documents, an object list, or a source tree.
Ask. The method is the same for any system with a published end-of-support date and an export a client can hand over; the scanner is the only part that has to be written.
Yes — every sample report on /samples is the real output of the same code, generated against a synthetic estate. Read it end to end.