Overview
This site is written for attorneys litigating internet defamation matters, and it is built around the two questions counsel actually asks: what does this element turn on, and how far does this method carry? Nearly every competing site is organized by platform instead, and that choice costs the reader more than it appears to.
The scheme is six groups. On the proof side, publication, attribution and quantum. On the method side, authentication, production and analysis. Two pages in each, twelve in total, and the placement of a page in a group is an argument about what kind of evidence question it is.
What organizing by platform produces
Platform organization is the obvious scheme, and I understand why it gets chosen. A page for Facebook, a page for Google, a page for Yelp, a page for Reddit, a page for X, a page for review sites in general. It is quick to build, it maps onto the way a client describes the problem in the first phone call, and it produces six pages that all say the same thing about Meta.
The reason it collapses is that the platform is a variable inside an answer rather than the answer itself. Take one question: can the author of an anonymous post be named? The structure of the answer does not change with the service. There is a chain running from content to account to connection to person; four of its links are records; the connection link is held by an access provider and the earlier ones by the operator; the final link is not a record at all and never has been. Only two details vary by platform, which are whether that operator publishes a preservation window and what tier of process reaches which category of record. A per-platform site therefore writes the chain six times and changes two sentences, and a reader who visits a second page learns almost nothing on it.
Platform pages also age badly in exactly the place they claim authority. Policy pages change without notice, and the honest ones on this site carry the date they were read for that reason. A page whose entire structure is here is what this operator does is a page whose entire structure is a dated claim. What an element turns on does not move on that schedule, and neither does whether a method reproduces.
The proof half: what the element turns on
Six things have to be established in a case of this kind, and they behave differently enough that treating them as one subject is the root of most wasted expert spend. They divide into three groups by what settles them.
Publication holds the two questions about whether the content went out and how far it traveled. Both live close to the records, and both are routinely asked to do more than records can. One reader satisfies publication as an element, so everything about audience goes to the size of the injury rather than to the element. A diff dates a change; whether the change amounts to republication is an argument made on top of it.
Attribution holds who wrote it and what the contemporaneous record shows. Both are exercises in narrowing rather than closing. Records reach an account, a connection, and what a page said on a date. The person at the keyboard and the truth of the statement come from somewhere else entirely.
Quantum holds the two that are not technical questions at all. Linkage a machine can observe between content and a person can be measured; whether any reader made the connection cannot. Exposure and timing are in the record; the quantum is modeled, and every model in this area turns a proxy into a number through assumptions belonging to whoever chose them.
Grouping this way makes an uncomfortable fact visible on arrival, which is the point of it. Two of the six elements do not need an expert, and on those two the useful thing for counsel to be told is not to spend the money.
The method half: how far it carries
The other six pages are methods, grouped by the same logic applied to a different axis: not what the question turns on, but how far the answer travels before it needs help.
Authentication is the reproducible tier. Process, documented at the time it happened, producing a claim narrow enough that another examiner can check every part of it against the same records. It is the least glamorous group on the site and the one that decides the most exhibits.
Production is reproducible once a scope and a date are stated. What comes back depends on what an operator holds and what it published on the day process was served, and three of the operators that come up most in this subject publish nothing about retention at all. That finding is why this group exists as its own tier rather than being folded into analysis: the uncertainty is not in the analyst's method, it is in the source, and it has to be attributed there.
Analysis is where the assumptions live, and it is where the one contested method on this site sits. An analysis group that contained no contested method would be a marketing document rather than an inventory.
What this scheme will not do
Three things, stated plainly so nobody spends time looking for them here.
- It does not tell you whether anything is actionable. I am not an attorney. The grouping sorts evidence, not claims. Every page names the point at which its subject stops being technical, and past that line the question belongs to counsel.
- It is not a platform playbook. No page here walks through what to do about content on a named service, because that is a different question asked by a different reader, and answering it here would produce precisely the duplication this scheme exists to avoid.
- It does not rank. A page carrying not a technical question or contested is not a weaker page than one carrying reproducible. Those two verdicts took more work to reach than any of the others, and they are the two this market does not print.
There is a fourth, which is that no grouping scheme survives contact with a matter that does not fit it. A case built entirely on a private message thread, or on content that was never indexed, touches half of these pages and none of the other half. The scheme is a way of reading a case quickly, not a template for one.
The practical effect is that a matter gets read in two passes. The first sorts the elements: what does each turn on, and which of them are record questions at all. The second sorts the methods against whatever the first pass demanded, and prices it. Counsel who has made both passes knows, before any conversation with any expert, which parts of the matter a technical record can move and which parts no expert touches. Knowing the second list is what makes the first list worth paying for.
Published