New Study: The Court a Firm Names Matters More Than Its Number

What we counted

A case results page is a firm's own account of its own work. There is nothing wrong with that, and a firm with a long record should say so.

The question is what a reader can actually do with the page. The only way to answer it is to read every one of them and count what is on it.

So we fetched the results page of each of the 217 firms in the directory, honoring robots.txt, and recorded five things: how many results the page lists, what kinds of case they are attributed to, whether any court or county is named, whether the page carries a prior-results disclaimer, and whether we could read the page at all.

That last one matters more than it sounds.

We treat the 10 pages we could not read as a limit on our crawler rather than as a finding about the firms. Boatman Ricci's page, for instance, answers us with a refusal: our note on that record reads "the results page served 2851 characters of text and no figure or mention of a settlement, verdict or recovery, so there was nothing on it to count. Whether the results are held back from the served HTML or the page carries none is not something this check can tell.". Its profile leaves pillar B unscored rather than scoring it zero.

A page we cannot read leaves the sub-factor out of the denominator, because a firm must not lose points to the reach of our own crawl. That is the rule everywhere on this site. 13 firms have pillar B unscored for that reason.

The 99 firms with a readable page and nothing countable on it are a different case, and a real finding. They have built the page, linked it in the navigation, and put no result on it that a reader or a search engine receives.

What the pages carry

Across the 107 firms that do publish, the pages hold 2,705 results between them. The median firm publishes 14. Wingate, Russotti, Shapiro, Moses & Halperin, LLP publishes 240, the most of anyone here.

Case types are close to universal, and they are the most useful thing on the page. A firm whose results are all rear-end collisions has told you very little about a birth injury case. Across these pages, the types named by the most firms are motor vehicle, truck and wrongful death.

Venue is where it thins out. 28 of the 107 name a court or a county anywhere on the page, and even then the reference usually sits beside the page as a whole rather than beside a particular figure.

A settlement with no court, no year and no case number cannot be looked up by anyone, including us.

Disclaimers split the set nearly in half: 44 pages carry one, 63 do not. New York's attorney advertising rules effectively require the line on a page that advertises past results, and most of these firms are in New York.

The states are not evenly covered, so read that table as a description of this directory rather than of any state's bar. Most of what we have measured so far is New York injury firms, and the other markets are early.

Why this page prints no figures

We hold 2,160 dollar figures taken off these pages. We publish none of them. Not on this page, not on the profiles, not in the rankings.

The reason is that we have verified none of them.

Verifying a published result means matching it to a docket or to documentation: the court, the case number, the date, the disposition. Our records have fields for exactly that, and not one firm in the directory has a result in them yet.

Until a result is in that state, repeating the number would make us a louder version of the firm's own marketing, with the directory's name on it.

There is a practical reason too. Reading figures off a page mechanically is unreliable. 3 of the figures we extracted are in the billions, which is far likelier to be two numbers welded together, or a lifetime total sitting next to a single case, than a recovery of that size.

A count of what is on a page survives that kind of error. A dollar amount does not.

A firm that wants its results on this site with the amounts attached can claim its profile and submit them with the documentation. That is the only route, it is free, and it is the same route for everyone.

What our own pillar B measures

Pillar B of the LFL Certification Score is worth 15 of 100, and it is named for verified outcomes. It does not measure them.

It measures what a firm publishes about its outcomes, in three sub-factors: how many results are on the page, how specific they are, and whether the page carries a disclosure. We say so on the methodology page, and it is worth saying plainly here too, because the pillar's name promises more than the pillar delivers.

The sharpest edge of that is the first sub-factor. It pays its top band at 25 published results, so the 33 firms at or above that number earn the same points, whether they publish 25 results or 240.

Publishing more unverified figures past that point earns nothing, which is deliberate. But it does mean a firm that posts many unverifiable figures scores above a firm that posts a handful it could document, and that is the wrong way around.

It shows up in the totals. The 107 firms that publish results average 54.9 out of 100. The 110 that do not average 37.5.

Some of that gap is pillar B directly, by construction: 101 firms score zero in it. The rest is probably the same thing that produces a results page in the first place, which is a firm large enough to maintain its own website. Read the gap as an association, not as a measurement of lawyering.

The honest summary is that pillar B currently rewards disclosure and penalizes silence, and that is a defensible thing for 15 points out of a hundred to do. It is not the same as knowing who wins. Verified results would need docket-by-docket reading, and where we have done none, the profile says so rather than implying otherwise.

The averages above carry one more caveat. A firm whose results page we could not read keeps pillar B pending instead of scoring zero, which takes those points out of its denominator entirely, so the totals being averaged are not all shares of the same scale. What that does to a comparison is its own study.

Reading a results page yourself

The page is still worth reading. It is just worth reading for different things than the biggest number on it.

• Look for your own case type, not the headline figure. The largest figure on a page is one case, and you are not that case. The kind of case is the part that transfers to yours. 106 of the 107 pages name their case types, so this is usually answerable in a minute.
• Check whether anything is attached to a court. A result with a county, a year and a case number can be looked up at the courthouse or through the state's case search. Most cannot, and a firm that publishes them anyway is publishing an advertisement, which is lawful and is not evidence.
• Ask how many of these the firm tried. Settlements and verdicts sit on the same page and are not the same achievement. Which of the two a result was is usually not stated, and the firm will tell you on the call.
• Ask for the denominator. A results page is a selection the firm made from its own record. How many cases sit behind the ones shown, and how long they took, is not the sort of thing these pages carry. It is a fair question on the free call, and no firm should mind it.
• Weigh it against things you can check. What the firm publishes about fees and what its clients say are both verifiable to some degree. We measured what these firms publish about fees and what their Google reviews are worth, and both hold up better than a settlement figure.

Every profile in the personal injury directory shows the count and the case types we read off that firm's page, with a link to the page itself and the date we read it. No amounts, for the reason above. If we have got a count wrong, the firm can tell us and we will read the page again.

Results pages are read by crawling each firm's own public site and honoring robots.txt, and each record stores the page URL and the date it was read. Every page is re-read monthly and this study recomputes itself from whatever the new reading says, so the figures above move without anyone editing the sentences around them. The date in the byline is the most recent read. A page we cannot read leaves pillar B unscored rather than scoring the firm zero. Amounts are stored but never published: only a result verified against a court record or documentation is shown with its figure, and there are none at present.