A regional sales director resigns on a Friday. Three weeks later, a competitor announces his hire, and by the end of the quarter, four of his accounts have moved with him. You file under the Defend Trade Secrets Act, and the preliminary injunction hearing is on the calendar before your team has finished collecting.
By the time collection closes, you are holding six different exports.
The forensic image gives you USB insertion events with timestamps. The badge system gives you after-hours building access. The email metadata gives you a spike in outbound attachments to a personal address in the last month of employment. The CRM audit log gives you report exports and record views. Phone records give you outbound calls. And a second custodian, the junior rep who resigned two weeks later, gives you another set of all of the above.
Your litigation support analyst does what every analyst does. One tab per source. Sort by date. Start building the chronology.
That workbook will get you to a defensible timeline. It will not get you to the case.
What the tabs can show you and what they cannot
A spreadsheet is very good at sequence. Sort the USB events by timestamp, and you can see that the exfiltration happened over four nights in the final two weeks. Sort the badge records, and you can line up the after-hours entries against those same nights. That correlation is real, and it matters.
What the spreadsheet cannot do is tell you that the personal Gmail address in the sales director’s email metadata also appears in the junior rep’s export. Or that the phone number he called eleven times in March belongs to a former colleague who left for the same competitor a year earlier. Or that a fifth person, never in your custodian list, sits between two of your custodians and connects them.
Those are not sequence questions. They are relationship questions, and rows and columns cannot answer them. The information is genuinely present in your data. It is just distributed across six tabs in a form where no amount of sorting will surface it, because the connection exists between files rather than inside any one of them.
This is where the VLOOKUP era of trade secret work quietly fails. Not because the analyst missed something, but because the tool cannot represent the thing you are actually trying to prove.
The three questions that decide the case
Who else is involved? Almost no trade secret misappropriation is a solo act. There is usually a recruiter, a receiving manager, or a colleague who left first and opened the door. It looks like your custodian list was built from the people you already suspected. The second-degree connections are the ones that expand the case, and they only appear when every identifier from every source is treated as a single entity.
When did the relationship actually start? The exfiltration date is the easy part. The valuable date is the first contact with the competitor, because that is what converts a document-taking claim into a coordinated raid and reframes the timeline for the court. That date is rarely in the forensic image. It is in call records and calendar metadata, months before anyone touched a USB port.
What does the fact finder see? A forty-tab workbook is not an exhibit. A pivot table has ever persuaded nobody. At some point the pattern in your data has to become something a judge can look at for thirty seconds and understand.
What changes when you model relationships instead of rows
Link analysis inverts the structure. Instead of one tab per data source, every person, phone number, email address, device, and account becomes a single entity, appearing once no matter how many exports it showed up in. Every call, transfer, badge swipe, and message becomes a relationship between those entities.
The practical consequence is that duplication stops hiding your case and starts building it. The personal email address that appeared in three separate exports does not become three rows in three tabs. It becomes one entity with three connections, and the moment it has three connections, it is visible.
Relativity reads documents. SentVi maps relationships from the data underneath them, which is why the two tend to sit side by side on a matter rather than compete.
A few things this makes routine that were previously manual:
- Import call detail records, email metadata, badge logs, and financial records into one dataset rather than six workbooks
- Filter by relationship strength to strip out the noise of one-off contacts and leave only the sustained communication patterns
- Run a timeline view across every source at once, so the phone contact in March and the USB event in June appear on the same axis.
- Identify the shortest path between two custodians, including the intermediary nobody had named yet.
Because it runs on SQL Server, the record volume is not the constraint. Millions of entities and relationships are a normal working set, not an edge case.
The exhibit is the point.
The output that matters is not the analysis. It is the chart.
Charts built in SentVi can be saved and shared with people who do not hold a license, which means the trial team, co-counsel, and the client can all look at the same visual without anyone buying a seat. The relationship map that took your analyst three days to develop becomes the demonstrative that takes the court thirty seconds to absorb.
That is the actual gap between a spreadsheet and a case. The spreadsheet holds the evidence. The map explains it.
See it on your own data. Request a free demonstration at https://sentvi.com/link-analysis-book-a-demo/ and then receive a free evaluation copy of SentVi to test on your matter.
