Scope
One target · public filings only
This is not a story about a hidden error. It is a demonstration of the harder question: given figures reported at different times, can the difference be traced, explained from the source, and separated from what the record genuinely cannot resolve?
Groupon's FY2016 results appear two ways in two annual reports. The difference is disclosed — the filing attributes it to a retrospective adjustment for discontinued operations. A competent analyst would expect that, and this page does not present it as a discovery.
What follows is the part that has to be right before any finding is worth anything: every figure traced to its filing, the stated cause quoted rather than inferred, and the record's own boundary marked where the documents stop answering.
| Source | FY2016 revenue | Filed |
|---|---|---|
| Form 10-K, FY2016 acc. 0001490281-17-000024 |
$3,143,354,000 | 2017-02-15 |
| Form 10-K, FY2017 · comparative acc. 0001490281-18-000018 |
$3,013,615,000 | 2018-02-14 |
| Difference | −$129,739,000 | −4.13% |
The output is a reviewable trail, not a verdict. A human decides what it means; the method's job is to make that decision traceable to the documents behind it.
Leave a work email and, if you want, one line about the process. The worked example comes back by email. No call, no sequence, no follow-up unless you reply.
Not a platform, not a workflow sprint. One defined piece of analytical work: take a named company's public filings, reconstruct what changed between disclosures, quote the stated cause, and mark what the record cannot resolve.
Pricing is a hypothesis, not a rate card. Public sources only for now: no access to confidential data rooms, no reliance opinion, and no investment recommendation. Those boundaries exist because the safeguards for handling client documents are not built yet — not because the analysis wouldn't benefit from them.
This page demonstrates document analysis, evidence tracing and the reconciliation of reported figures. It does not demonstrate proven client ROI, measured time savings, or validated automation performance — VaultIQ has no production results to point to.
It is not a discovered discrepancy, and nothing here should be read as one: the FY2016 difference is explained by the company's own disclosure. The demonstration is the reconstruction and the classification, not the surprise.
It is also not investment advice, not a diligence opinion, and not something a firm could rely on in an investment decision. It is one worked example of the method, prepared from public filings, showing what a reviewable finding looks like and where the evidence stops.