How we know what we publish.

Who answers

Every Sidebar member is a legal management professional at a law firm with 10 to 200 attorneys. That is the whole panel. There are no vendors, no consultants who serve firms rather than work at one, and no attendees who joined a list.

Membership is by application and manual review. Before anyone is admitted we verify their title, their employer, and their firm size through LinkedIn and their work email domain. A founder reads every application. Applications from outside the 10 to 200 attorney band are declined, as are applications from people who sell to firms rather than run them.

What they are asked

One question a week, by email, every Tuesday. The question is written by a founder and is specific enough to answer from experience in about five minutes.

Replies are private. A member writes back to the question and no other member ever sees what they wrote. There is no forum, no comment thread, and no peer-to-peer surface anywhere in the product, so nobody is performing for an audience when they answer.

Answers are self-reported. We are recording what a firm says it does, which is the honest description of any practitioner research, and it is worth reading the figures on this site with that in mind.

How a finding is produced

When the week closes, every reply to that question goes to Claude at once, labeled only by firm size band, practice area, and technology stack. Names, email addresses, and firm names are stripped before the model sees anything.

The model identifies the themes, the tensions, and the points of agreement across the replies. A founder reviews the draft before anything is published or sent.

Percentages are computed from the replies to that question, not estimated by the model. Where a figure appears on this site, it is the share of responding members whose answer fell into that category.

When we publish a figure, and when we do not

We publish a figure only when at least five members replied to that question. Below that floor the week still produces a synthesis for members, but no number from it reaches a public page.

We publish the floor rather than the count behind each individual figure, and you should weigh what that means. Five is a small number, and a share drawn from a narrow panel is a different instrument from a large open survey. We would rather state the threshold plainly and let you judge it than imply a scale we do not have. What the panel gives up in size it gains in specificity: every reply comes from a verified legal management professional at a firm with 10 to 200 attorneys, which is a harder group to reach than a broad industry sample.

Figures on this site describe mid-size firms. They are not a claim about the profession generally, and where a public benchmark covers the same ground we cite it alongside our own finding so you can see both.

What is public and what is not

Public pages carry the aggregate pattern: what most firms report, what the minority does differently, and what that means for running a firm.

The breakdown by firm size, practice area, and technology stack stays with members, as does any individual firm's story. That is not a marketing gate. Segment-level data and single-firm detail are the two things that could identify a member or a member's employer, and they are the reason people answer candidly in the first place.

Corrections

Articles here are living pages. They are updated as new cycles touch the same question, and the date on each page is the date it was last revised rather than the date it first went up.

If a figure or a characterization on this site looks wrong to you, tell us and we will check it against the underlying replies. Corrections are made on the page itself.