Guide · For boards and sponsors
What a board verifies before it meets a finalist. The license. The disciplinary record. The work the candidate actually ran. And the gap a friendly reference will not show.
The candidate's referees describe a career. The record says whether that description is the one the board can hire.
General counsel reference checks and background are a record check. They run before the board meets a finalist, inside a retained general counsel search, not as a courtesy round after someone has already decided. Of 52 GC and CLO files since 2017, inside 400+ placements (230+ in-house), the shortlist is a set of assessment dossiers at 24 days. Median brief to accepted offer is 11 weeks, range 8 to 16. The firm that holds those files was founded in early 2017 by Lorenzo Sartori.
A NASDAQ-listed medtech company in Boston is the file a committee can read without a composite wearing a city name. The search was blind. The sitting GC was not shopped. Four dossiers reached the board on day 24. The accepted offer closed in week 11. The 12-month retention checkpoint held. The dossiers named examinations, product lines, and public-company work. They did not say “life sciences experience” and stop.
| Check | Record | Who holds it | When |
|---|---|---|---|
| License status | Public attorney search in every jurisdiction claimed | Search, before the dossier | Inside the 24 days to shortlist |
| Public discipline | Same bar record; New York also requires a disciplinary-body letter for in-house registration | Search | Before the dossier |
| In-house registration | State directory where the seat requires it and the lawyer is not generally admitted | Search | Before offer |
| Named work | Deals run, examinations sat, board work done | People who saw that work | Written into the day-24 dossier |
| Back-channel | Not a quiet ask-around | Search, after mutual interest | Not in week 1 of a blind succession |
Median shortlist 24 days; median brief to accepted offer 11 weeks (range 8 to 16). Of 52 GC and CLO files, 2017 to 2026. Boston file: four dossiers on day 24, accepted offer in week 11.
Source: Sartori & Partners corporate track record as published on the case studies, 2017-2026. License steps: state bars and 22 NYCRR Part 522.
A dossier that cannot answer the record is a CV with compliments attached.
First, every bar the candidate claims, and any bar the public record adds that the CV omitted. Status is the fact: active, inactive, not eligible, suspended, disbarred, administrative. A title on a proxy circular is not a license. ABA Model Rule 5.5(d) lets a lawyer admitted in another United States or foreign jurisdiction serve an employer from an office where they are not generally admitted, and it withholds that permission from anyone disbarred or suspended in any jurisdiction. Comment 17 to the same rule says the host state may still require registration, a client-protection assessment, and continuing legal education. Registration is not general admission. A board that stops at “admitted in New York” has not asked the question the rule asks.
Second, public discipline, read rather than summarized by the candidate. California’s State Bar tells the public to look up license status and disciplinary and administrative history. That lookup is the check. A candidate’s sentence “no issues” is the claim. The two are not interchangeable.
Third, in-house registration where the company will sit the lawyer and the lawyer is not generally admitted there. New York’s current text of 22 NYCRR § 522.7(a) requires an in-house counsel who is not admitted in New York to file within 90 days of the later of commencement of employment or 15 April 2020. Section 522.2 asks for a certificate of good standing from each jurisdiction of licensure. The Court of Appeals registration instructions also require a letter from the disciplinary body in each of those jurisdictions on whether charges were filed and how they were disposed. Section 522.7(c) makes failure to comply professional misconduct. The court system publishes a searchable directory of registered in-house counsel. From 1 January 2026 the Appellate Division, Second Department, also requires an affidavit that the applicant has read the New York Rules of Professional Conduct and affidavits from at least two people who have known the applicant for five years. California runs a separate Registered In-House Counsel registration. It is not the same thing as passing the California bar. The CV should not be allowed to blur them.
Fourth, the work the scorecard named, at the grain the case studies already use. On a first general counsel search for a venture-backed SaaS company in California, the disqualifiers were specific: last three years spent in management only, no SaaS revenue contract actually closed, a plan to arrive with a four-person team. The dossiers that survived recorded deals closed, privacy programmes stood up, and board work done. On a chief compliance officer search for a NYSE-listed payments company in New York, the dossiers named examinations. They did not say “regulatory experience.” On a deputy general counsel search for a PE-backed industrial group in Houston, the dossiers named the carve-out, the transition-services agreement, and the last time the person had written for a securities calendar. A sponsor who accepts “industrials experience” from a referee has waived the only test that file was run to pass.
Fifth, the dates. Employment that overlaps a license suspension, a registration that lapsed between employers, a “GC” title that was a deputy seat: those are document facts. They do not require a judgment about character. They require the CV and the public record to be put on the same page before anyone is described as board-ready. The research programme has said since 2019 that a candidate’s claim and the verified record are allowed to diverge, and that the published figure is the gap. Reference checks are that rule applied to a hire.
The right person saw the matter. The wrong person heard about it, or still reports to the incumbent.
Call the person who can confirm the line in the dossier. If the scorecard says the candidate owned an audit-committee pack, the referee is the audit chair or the chief executive who received it, not a peer who liked working with them. If the scorecard says they ran a carve-out, the referee is the operator who lived the transition agreement, not a law-school classmate. If the scorecard says they sat an examination, the referee is the person in the company who was in the room, and the public order if one exists. The Boston dossiers named examinations, product lines, and public-company work because those were the things a director could test. A reference that cannot be tied to one of those lines is a character sketch.
Do not call the sitting legal department. On the Boston file the incumbent was still in the chair, outside counsel still reported to that GC, and outside counsel would have been the first leak. Listed medtech departments on that case typically run eight to twenty lawyers. That is a small building. A “confidential” call to the deputy, the employment counsel, or the panel partner is how the building finds out the board has started. The same constraint is why talent acquisition could not post the role. Posting and phoning the team are the same leak in different stationery.
Lawyer-referees are under a second limit that has nothing to do with the search. ABA Model Rule 1.6 prohibits a lawyer from revealing information relating to a representation unless the client gives informed consent, the disclosure is impliedly authorized to carry out the representation, or a listed exception applies. Former outside counsel, a former GC still bound by the company’s privilege, and a current panel partner cannot give a rich account of the matters just because a board asks. What they can confirm is often narrow: that the person held the role, that a matter existed, that they would work with them again. A board that treats a guarded answer as a negative, or a warm answer as a waiver of privilege, has misunderstood the rule. Ask questions a lawyer can answer. Do not push them past it.
Candidate-supplied names are necessary and insufficient. People who agree to be listed have already decided to be helpful. Take the call. Write down what they actually saw. Then hold it next to the bar record and the named work, which is the comparison the firm owes a committee. The search holds the calls. The board does not run a parallel set. A director who “knows someone” in the candidate’s city routes the name to the person running the file. Two uncoordinated conversations are how a succession stops being blind. The product the board is paying for is the general counsel search, and the reference is a part of that product, not a side hobby of the committee.
A general counsel practices on a license, a registration, or both. The CV is not the register.
Run the public search in every jurisdiction the candidate names, and in any jurisdiction where the company will put them to work. Read the status line and the disciplinary history, not a screenshot the candidate forwards. California publishes both license status and disciplinary and administrative history and tells the public to use that lookup. New York’s in-house registration file is more specific than a status bit: good standing from each licensing jurisdiction, plus the disciplinary body’s own letter on charges. That letter is the document. A referee’s recollection of it is not.
Multi-state careers are the ordinary GC career. Rule 5.5(d) is built for them, and so is the mistake. A lawyer in good standing in Illinois, sitting in New York, advising a Delaware corporation, with a California subsidiary that wants them on the product calls, is not one question. Are they disbarred or suspended anywhere? That single condition defeats the in-house exception. Are they registered where the host state requires it? Comment 17 says registration may be mandatory even though the model rule allows the work. Did the registration die when they changed employers? New York § 522.5 terminates registration when the employment ends, with a 30-day window to re-register if a new in-house employer picks them up immediately. A gap longer than that window is a new application, not a clerical detail. Foreign admission is a further question: Rule 5.5(e) requires a member in good standing of a recognized legal profession, subject to real discipline by a real authority, and advice on United States law has to rest on a lawyer licensed to give it.
Administrative status matters as much as a published censure. Inactive, not eligible to practice, suspended for fees or education, or registered in-house when the job description assumes a full local license: each of those is a different conversation with the board, and each of them is visible before anyone’s reputation is debated. The Second Department’s 1 January 2026 affidavit rule is a useful picture of what a court itself asks when it registers in-house counsel: two people who have known the applicant for five years, on character, plus a sworn description of the duties. That is a registration requirement in one judicial department. It is not our questionnaire. It is evidence that “trust me” is not the standard the regulator uses, so it should not be the standard a sponsor uses either.
Do the license check before the calls, not after a glowing reference has made the calls feel unnecessary. A disciplinary hit does not automatically end a candidacy. An undisclosed one does. The dossier should say what the record shows, what the candidate said when asked, and whether the seat’s registration is already in place or is a condition of the start date. Leaving registration to “the lawyers will sort it in the first month” is how a company discovers, after announcement, that its general counsel is not entitled to do the job it just described to the exchange.
The information a director wants from a friend is usually real. The way they go and get it is how the search stops being confidential.
A back-channel is a call to someone the candidate did not list. Boards like it because supplied references feel staged, and they are right that supplied references are staged. They are wrong about the remedy. On a blind general counsel succession the remedy is not a quieter leak. The Boston medtech case states the failure mode without decoration: everyone relaxes, someone quietly asks around, and by week three the market is trading the name. That is not a rate. It is the published account of how this kind of search gets bungled. The file itself was run the other way. Blind both ways. Company unnamed until mutual interest. Sitting GC not shopped. No posting. Four dossiers on day 24.
The density of the market is why the failure mode is fast. The Boston case describes a life-sciences field in which everyone lunches, and a legal department of eight to twenty lawyers. A clumsy call to a friend who might know someone is a leak because the friend can do the arithmetic: medtech, Boston, a GC-shaped question, this month. Private equity sponsors repeat the error one layer up. An operating partner phones a former colleague at the candidate’s current fund, or at the law firm on the carve-out, “just to get a sense.” The colleague does not experience that call as informal. They experience it as the seat being open. The Houston deputy search was explicit about the document that had to be confirmed — the carve-out, the transition agreement, the securities calendar — and it was confirmed in a dossier, not in a sponsor’s side conversation.
The handling rule is short. No back-channel before mutual interest. After mutual interest, a named list of people who saw the named work, told to the candidate as categories rather than as a surprise, placed by the search and not by individual directors, written up as findings with the question that was asked. Disagreement between a referee and the CV goes in the dossier as a gap, not as a verdict delivered in the corridor. If a director has already made the call, the search needs to know the same day: who, what was asked, what was said, whether the company was named. Pretending the call did not happen is how the second call happens.
There is a legal version of the same mess, and it is not optional once a vendor is in the chain. The Consumer Financial Protection Bureau’s Circular 2024-06, issued 24 October 2024, says an employer may not use background dossiers, scores, or similar third-party reports for hiring and then skip the Fair Credit Reporting Act. Those dossiers are often consumer reports. Permission comes first. Notice comes before adverse action and with it: a copy of the report, and the statement of rights. The FTC’s guide for employers, Using Consumer Reports: What Employers Need to Know, is the same sequence, starting with a standalone written disclosure and written authorization. A conversation the board holds itself is not, by itself, a consumer report. A consumer report is a communication from a consumer reporting agency. The moment a vendor compiles public records, employment history, or interview notes into a product the board relies on, the circular applies to that product. Adding a director’s own calls on top does not turn the vendor’s half off. Whether a search firm’s written memorandum is itself that product is a question for the company’s counsel. The safe operating rule is the FTC sequence whenever someone other than the company assembles the file.
Jurisdictions and matters are written down. Nobody is phoned. A blind succession stays blind.
License status, public discipline, registration. This does not require the market to know the company is looking.
People who saw the named work. Candidate knows the categories. Directors do not run a parallel set.
Findings and gaps, in writing. Boston: four names on this day. Programme GC and CLO shortlist: 24 days.
Boston accepted offer. Programme median 11 weeks, range 8 to 16, of 38 files. The reference is already in the file.
Where a candidate’s claim and the verified record diverge, publish the gap rather than the flattering number.
This is the unflattering part. The references a candidate is willing to supply are the account that flatters them. Hiring on that account is how boards miss the gap.
The Sartori research programme has run since 2019. It says, in terms, that where a candidate’s claim and the verified data diverge, the programme publishes the gap rather than the flattering number. The clearest case it names is portable originations: what a partner says will travel, against what the record supports. That case is a partner fact. The rule is not confined to it. A general counsel’s CV, and the people that general counsel offers as references, are claims. The bar record, the registration file, and the matters the person actually ran are the verification. When they diverge, the divergence is the finding. Rounding it into “strong references” is the flattering number the programme refuses to publish.
No rate at which general counsel references contradict a CV is published on the research page, the about page, or the case studies. What those pages do show is the shape of the gap when a file is done properly. California screened out lawyers whose last three years were management only and lawyers who had never closed a SaaS revenue contract. New York’s payments mandate refused “regulatory experience” as a substitute for a named examination. Houston refused “industrials experience” as a substitute for the carve-out and the securities calendar. Boston refused a life-sciences slogan in place of examinations, product lines, and public-company work. Each of those is a candidate claim that would have passed a friendly reference and failed a dossier.
The map is a different instrument and should not be asked to do this job. The research programme describes a global map approaching 1.5 million lawyer profiles, used to see who sits in a market. It is not a disciplinary database. License status still comes from the bar. Interview cohorts and the quarterly survey, running since 2019, are aggregate. They are not a substitute for the check on one general counsel. Mandate telemetry is the firm’s own closed files. On this subject that telemetry, as published, is the 52 GC and CLO searches: dossiers at 24 days, accepted offer at 11 weeks median. It is not a hidden percentage of bad references.
Four names the directors can sit with. Each one already tested against the record. Not a stack of compliments stapled to a CV.
On the Boston file the board interviewed four and hired one. The shortlist arrived on day 24, which is the GC and CLO figure across the 52, and the accepted offer closed in week 11, which is the median of that book. The 12-month checkpoint held. Programme retention of placed in-house leaders is 97 percent at 12 months and 91 percent at 24 months. Those percentages are the book. They are not a property of this company, and they are not what a reference “predicts.” Offer acceptance on the offers that go out is 96 percent. Retained corporate mandates complete at 92 percent. A reference process that starts after those offers are emotionally done is not what produced the clocks.
The client’s own sentence on that shortlist is below, verbatim. It runs further than the file fact we lock. The case study is explicit: the words run to two years in post; the checkpoint we report for the file is 12 months, which held. We do not convert the sentence into a second metric.
The shortlist arrived in just over three weeks and every name could already sit with our board. The person we hired is still in post two years later.
Timeline: Day 24 shortlist · week 11 offer · one of 52 GC/CLO files
Timeline: Day 24 · week 10 · in post at 24 months · one of 52
Read the two files as a pair, not as a promise. Boston is a succession with an incumbent in the building and a 12-month checkpoint that held. California is a first hire, still in post at 24 months, which is a file fact on that case and sits next to the programme rate of 91 percent at 24 months rather than replacing it. Both shortlists landed on day 24. Both refused a slogan. A sponsor comparing a portfolio GC with either of them should ask for the same shape: the matter, the person who saw it, the license, and the gap. The Houston deputy case is the portfolio version of that shape. The New York payments case is the examination version. Neither is a general counsel reference statistic, and neither is quoted here as one.
A reference process that will do anything a director suggests is not a process a general counsel can defend.
We will not treat supplied references as clearance. We will not call the sitting legal department, the incumbent’s deputy, or the panel partner who still bills that incumbent. We will not let a director run a parallel back-channel, and we will not start one before mutual interest. We will not describe a vendor background score as a bar check. The license is the state’s record. The score, if a board insists on buying one, is a consumer report with a statute attached, and it does not replace the lookup.
We will not publish the flattering number where the record diverges. That sentence is the research programme’s, in force since 2019, and it binds this desk. We will not state how often general counsel references come back worse than the CV. That rate is not on the research page, not on the about page, and not in the case studies. Inventing it would be the flattering move in reverse: a hard number with no base. We will not use the market map, approaching 1.5 million profiles, as if it were a disciplinary register. We will not attach the 97 percent and 91 percent retention rates to a named company. They are programme rates, of the placed in-house leaders inside the 230+.
We will not run the check on contingency, and we will not post the seat so that referees arrive on their own. The search those rules belong to is a retained general counsel search. If the honest reading of the record is that the deputy already meets it, the answer is to say so, not to collect references until a search feels justified. The decision between those two is promote the deputy or hire a new GC.
The record, not a mood. Every jurisdiction the candidate claims, and the public disciplinary history on that license. In-house registration where the seat sits and the lawyer is not generally admitted there. The matters the scorecard named: deals the person ran, examinations they sat, board work they did. And the gap between that record and the account the candidate's own referees give. Of 52 GC and CLO files since 2017, that package is the dossier, not a courtesy call after the offer. The method around it is how to run a general counsel search.
People who saw the work the dossier will name: a former chief executive or audit-committee chair the candidate reported to, a predecessor or successor in the legal seat, and outside counsel on a named matter only to the extent a lawyer can speak without breaching client confidence. Do not call the sitting legal department while the incumbent is still in the chair. On the NASDAQ-listed medtech succession in Boston, outside counsel still reported to the sitting GC and would have been the first leak. Candidate-supplied names are taken. They are not the record.
Look up every jurisdiction on the CV in that bar's public attorney search, and read status plus public discipline. California publishes license status and disciplinary and administrative history. New York's in-house counsel rule, 22 NYCRR Part 522, requires a certificate of good standing from each jurisdiction, and the Court of Appeals registration instructions require a letter from each disciplinary body on whether charges were filed. The application is due within 90 days of the later of the start of employment or 15 April 2020, and noncompliance is professional misconduct. ABA Model Rule 5.5(d) withholds the in-house exception from a lawyer disbarred or suspended in any jurisdiction. A clean record in one state does not clear the others.
It should not run one. The Boston succession case states the failure mode: everyone relaxes, someone quietly asks around, and by week three the market is trading the name. That file was not run that way. It was blind, the sitting GC was not shopped, and four dossiers reached the board on day 24. After mutual interest is confirmed, calls about the named work go through the search, the candidate knows the categories, and the finding is written into the dossier. A director who knows someone routes the name. The director does not dial.
It applies when a third party compiles the dossier and the company uses it for a hiring decision. The Consumer Financial Protection Bureau's Circular 2024-06, issued 24 October 2024, says background dossiers obtained from third parties are often consumer reports. The employer needs the candidate's permission before procuring the report, and notice before and when it takes adverse action. A conversation the board holds itself is not, by itself, a consumer report: a consumer report is a communication from a consumer reporting agency. The moment a vendor compiles the calls, the public records, or a score, the statute is in. Whether a search firm's written memorandum is that product is a question for the company's own counsel. The FTC's employer guidance is the operating sequence whenever it is: a standalone written disclosure, written authorization, then a copy of the report and a summary of rights before an adverse step.
Inside the dossier, which on GC and CLO files reaches the company at 24 days, not after a verbal offer. Median brief to accepted offer is 11 weeks, range 8 to 16, of 52 GC and CLO files since 2017. The NASDAQ-listed medtech in Boston had four dossiers on day 24 and an accepted offer in week 11. The 12-month retention checkpoint on that file held. Offer acceptance across corporate mandates is 96 percent, and 92 percent of retained corporate mandates are completed. Those are programme rates. They are not a promise that a friendly reference predicts retention. To brief the search those checks sit inside, use general counsel search.
The clocks and the named files are the case studies. The gap is the research programme. The license steps are the rules, cited to the rule.
Sartori clocks are the published GC and CLO book: 52 files since 2017, dossiers at 24 days, accepted offer at 11 weeks median (range 8 to 16). The Boston file is four dossiers on day 24, accepted offer in week 11, 12-month checkpoint held. The research programme has run since 2019. Part 522 and the California State Bar lookups are the regulators' own pages. CFPB Circular 2024-06 is 24 October 2024. No reference-failure rate is stated, because none is published on the research page, the about page, or the case studies.
This page is the check. The neighbouring pages are the search it sits inside and the succession it was written from.
The retained method these checks sit inside: Mandate Blueprint, a mapped field, dossiers at 24 days, accepted offer at 11 weeks median of 52 GC and CLO files.
Read the search methodBlind succession. Four dossiers on day 24. Accepted offer in week 11. The 12-month checkpoint held. The case that states what a quiet ask-around does.
Read the Boston fileThe scorecard the references are scored against: board access, the compliance book, the secretary book, and the regulator that actually reaches this company.
Read the scorecardAlso on the company path: hiring your first general counsel, hiring in-house counsel for a PE-backed company, lateral partner due diligence when the reference file is a panel firm’s incoming partner rather than a lawyer on the company’s payroll, and the research programme that states the gap. To brief: contact the corporate desk or book a call.
For boards and sponsors
License, discipline, and the named work, written into the dossier. Shortlist at 24 days. Median accepted offer at 11 weeks, of 52 GC and CLO files since 2017.
