Seventeen entries sit in this category and, read as causes of action, they have almost nothing in common. A retailer whose product pages stopped appearing after a replatform — a move to a different e-commerce system, which changes every URL at once. A person establishing what a search result displayed on one Tuesday two years ago. Breach of contract, negligence, defamation, the Lanham Act, and a class of technical failure with no legal name of its own.
Read the same seventeen as evidence problems and there are four.
The four questions
Every entry here is labeled with the single question it turns on. The label sits above the prose because it is the first thing that has to be settled and, in practice, the last thing anyone settles.
- Causation — what caused the change. Visibility moved, and the dispute is about what moved it.
- Standard of care — whether the work met the standard that applied when it was performed. Not the standard now. The standard then.
- Reconstruction — what existed, when, and under whose control. What that page said in March, what the results page looked like, who held the account.
- Valuation — what the visibility was worth. Not whether something was lost, but what the thing that was lost consisted of.
The label is not taxonomy for its own sake. It settles three things before anybody spends money: what evidence you have to collect, and how fast; which kind of expert the matter needs, because reconstruction and valuation are not the same skill; and whether the case is provable at all. A reconstruction matter usually turns on records that still exist somewhere. A causation matter turns on records that are deleting themselves while you read this. Google Search Console, the reporting property a site owner verifies, holds sixteen months of performance data on a rolling window, and raw server access logs are commonly rotated every thirty to ninety days.
Why these are grouped by relationship, not by cause of action
The four disciplines here describe the relationship the dispute came out of. Somebody was hired. Something was published. A mark was used. A system was changed, or a platform changed something on its own. That is a deliberate choice over the two more obvious schemes.
Grouping by symptom is the first alternative and the worse one. Organized that way, half this category collapses into one entry called "traffic went down," and a collapse caused by a botched migration sits beside one caused by a competitor's link campaign. Technically they are nearly the same exercise: establish the baseline, date the break, enumerate the candidate causes, eliminate them. As lawsuits they have nothing in common — different defendant, different duty, different discovery.
The second alternative, grouping by cause of action, splits matters that share every methodological problem they have. A false advertising claim between competitors and a defamation claim by an individual are different bodies of law and the same evidentiary exercise: prove what was displayed, for which queries, on which dates, to how many people. Sorting by relationship keeps the technical work together and leaves the legal theory to you. You already know what you are pleading. What you may not know is which records exist and how long they last.
So: commercial and contract disputes (four), where retainers are written in categories rather than artifacts and performance is rarely measurable against the contract. Reputation and content disputes (four), where prominence is a property of a layout on a date, not a number. Trademark, advertising and competition (four), where the technical record proves display and stops short of confusion. Technical and platform failures (five), the most testable matters here and, for that reason, the ones where an overreaching opinion is easiest to take apart.
Causation is the hard one
Of the four, causation is answered badly most often, and always the same way. Two lines go on a chart. One is a date — a launch, a demand letter, an agency's last day. The other is organic traffic. They move at about the same time, and the opinion treats the coincidence as the finding.
Organic search has no stable counterfactual. The engine changes continuously and only some changes are announced. Query demand is seasonal. Competitors do their own work. A break in a chart is consistent with all of it. The question can still be answered, and the entries here describe what answering it takes, including confirmed update dates from a neutral record, comparison against pages nobody touched, and evidence separating what changed on the site from what changed at the engine.
Here is the uncomfortable part, and the most useful thing on this page. In a meaningful share of matters that reach me, the honest answer after looking at the data is that the data does not support the theory. Not that the theory is wrong — that the surviving records cannot carry it. The logs rotated. The break predates the conduct. The comparison sites moved the same way on the same day. That conclusion is worth a great deal reached before a complaint is drafted, and it is ruinous when it surfaces for the first time in a deposition, out of an opposing expert's mouth, two years in.
The overlap is deliberate
These four questions are not mutually exclusive. One fact pattern routinely raises three at once.
An agency is hired, rebuilds a site, moves it to new URLs, and traffic falls seventy percent on launch day. That is a standard of care question, because there is published, dated guidance on how a site move is supposed to be executed and the redirect map — the list pairing every old URL with its new destination — either matched that guidance or did not. It is a causation question, because a core algorithm update, one of the broad ranking changes the engine announces and dates, ran the same month and the defense will say so. It is a valuation question, because the figure in the complaint has to be built from traffic that never arrived. And if the site was rebuilt in place with no archival copy, it is a reconstruction question too.
An expert whose report addresses one of those and ignores the other three has not written a narrow report. He has handed the remaining three to the opposing expert, whose framing will then be the only framing in front of the court. In my experience that is the most common structural weakness in reports here — not error, but silence exactly where the other side is loudest.
How to read this category
Start from the discipline that matches the relationship, then read the label. Causation means preservation is urgent. Reconstruction means the work is retrieval and dating, and can usually still be done late. Standard of care means the question is what the guidance said on the date of the work, not today. Valuation means a division of labor, because the search opinion should stop at the traffic counterfactual.
Nothing here is legal advice. The companion category covers the rules that decide whether any of this is heard. Establishing what happened is half the problem; getting a court to receive the proof of it is the other half.
Published