Guide · In-house retention
Exit interviews: ignored talent intelligence.
A corporate legal function runs a handful of exits a year and reads almost none of them. The instrument that already collects the reason lawyers leave is the one people-data asset a general counsel owns outright — and the only one that never reaches the chief financial officer.
Count them. Then count the ones anybody read.
The instrument is already running, already budgeted and already on the calendar of every departure. Across Sartori’s 1,675 structured interviews with New York in-house lawyers and the executives who hire them, 268 respondents who had left an in-house legal seat within the previous 24 months described an exit conversation held by corporate HR; only 41 of them said that anyone from the legal function ever saw what it produced.
Notice, handover, last day, badge. Nobody asks the one question the company will spend six figures answering again in the backfill. You are buying the same information twice.
Most legal functions sit at stage two or three and describe themselves as being at stage five. The published record of where programs actually stop is below.
- 3 / 8 / 30
- median in-house lawyers, by company revenue band$50M–$1B, $1–$6B, $6B+; 75% of the smallest band run five or fewer
- Thomson Reuters Institute, 2026 State of the Corporate Law Department Report, March 2026
- 70.9%
- of exit-interview programs run by HR19% by the direct supervisor; 1% by an outside consultant
- Spain & Groysberg, Harvard Business Review, April 2016 (210 organizations)
- fewer than 1 in 3
- program owners who could name an action takenand fewer than one in three consolidators briefed senior decision makers
- Spain & Groysberg, Harvard Business Review, April 2016
- 19%
- of S&P 100 issuers disclosing a turnover rateagainst 96% that discuss retention qualitatively
- Gibson Dunn survey of S&P 100 Form 10-Ks, December 2024
How many exit interviews a legal function actually produces in a year.
Nobody publishes an exits-per-legal-department figure, and inventing one would be the easiest mistake on this page. The honest version is a band, built from department size on one side and published voluntary-turnover rates on the other.
Start with the size of the thing. The Thomson Reuters Institute’s 2026 State of the Corporate Law Department Report, published in March 2026 on more than 2,300 interviews with corporate general counsel, puts the median number of in-house lawyers at three in the $50 million to $1 billion revenue band, eight from $1 billion to $6 billion, and thirty above $6 billion. Three-quarters of companies in the smallest band run five attorneys or fewer. Only at the top of the market does the department become an organization in its own right: the CLOC and Harbor 2025 State of the Industry Report, covering 186 departments at a median revenue of $12.8 billion, records a median of 44 attorneys inside a median legal headcount of 77 full-time equivalents.
Now the rate. Mercer’s 2025 US Turnover Survey, drawn from 2,617 organizations and republished by the Connecticut Business & Industry Association in October 2025, records average US voluntary turnover of 13%, falling to 9.1% for non-sales professionals, 6.3% for management and 5.2% for executives. Those are the closest published occupational rates to the shape of an in-house legal hierarchy, and they are all-industry rather than legal-specific.
Put the two side by side rather than multiplying them and the band is clear enough to plan against. A mid-market legal function loses lawyers in ones and twos in a year. A department at the top of the market loses them in a low-double-digit count. Neither number is large. That is precisely why the failure is so strange: this is not a data-processing problem at the scale of a distribution center or a call center, where an individual read is impossible and a model is the only option. This is a volume a single person could read properly in an afternoon each quarter.
The flow is not slowing. The ACC and Empsight 2025 Law Department Compensation Survey, published in September 2025 on 1,632 self-reported US in-house respondents, found 28% had changed jobs in the previous two years, up from 25% a year earlier, and 17% saying they were likely or very likely to change in the coming year. The source does not annualize the first of those figures and neither should a board pack. ACC’s December 2025 study of stress among in-house legal professionals, drawing on the same respondent base, found 22% reporting high or severe stress, and 24% of that high-stress group planning to leave within the year — three to five times the attrition risk of the mildly or moderately stressed.
Behind the flow sits a market that has roughly doubled. ACC’s September 2025 population paper, a residual estimate built from Bureau of Labor Statistics data, puts US in-house counsel at 145,000 in 2024 against 78,000 in 2008, with New York the largest state pool at 23,200 and 117 in-house counsel per 100,000 residents. The Bureau of Labor Statistics’ own Occupational Outlook Handbook, last modified in August 2026, counts 863,700 lawyer jobs in 2025 with legal services accounting for 52% of them. More in-house seats, at a stable turnover rate, means more legal-function exits every year in this city. Sartori maps roughly 67,000 lawyers across the New York market, which is what makes a departure traceable to the seat it left rather than to a name on a leaver report.
This is a volume a single person could read properly in an afternoon each quarter.
Run, filed, consolidated, never briefed, never acted on.
The exit interview does not fail at the point of collection. It fails four steps later, in a sequence that has been measured, published and left unchanged for a decade.
The load-bearing study is still the field work Everett Spain and Boris Groysberg published in Harvard Business Review in April 2016: a survey of 188 executives and interviews with 32 senior leaders, covering 210 organizations across 33 industries in more than 35 countries. Three-quarters of those companies conducted some form of exit interview for at least some departing employees. Of the companies with a program, 70.9% had HR handle the process, 19% the departing employee’s direct supervisor, 8.9% the supervisor’s manager and 1% an outside consultant. Mandatory programs were more common at midsize companies, at 87%, than at large ones, at 77%.
Then the sequence breaks. Asked to name a specific action taken as the result of an exit interview — a policy change, or an intervention in HR, operations, marketing or another function — fewer than a third of executives whose companies had a program could cite one, which led the authors to conclude that two-thirds of existing programs were mostly talk with little productive follow-up. Asked what happened to the data, most said their companies consolidated it, and fewer than a third of those organizations regularly shared it with senior decision makers. The authors’ own summary of the failure is the ownership problem in a line: exit interviews are wholly an HR function at most companies, and HR shares the consolidated result with management only when directly asked.
A decade of later work has not moved the needle. Gallup’s July 2024 survey of 150 chief human resources officers at Fortune 500 companies found only 10% rating their organization highly effective at the departure stage of the employee experience — the weakest of the eight stages from recruitment onward. Its companion study of 716 people who had voluntarily left an employer in the previous twelve months found fewer than half satisfied with how the exit was handled. Work Institute’s 2025 Retention Report, built on 14,215 coded exit interviews from 2024, records that 6% of senior HR executives always conduct an exit interview after the employee has left, on a panel of 88 HR and operational leaders surveyed between December 2024 and January 2025, and that 11% of employers strongly agree they consistently track the cost of turnover.
Read those together and the shape is a funnel with a hole two-thirds of the way down. Collection is close to universal. Consolidation is common. Distribution to the people who make staffing and budget decisions is rare. Action is rarer. Disclosure of a number is rarest of all. Nothing in that chain is expensive to fix, and none of it requires a new vendor.
It happens
Three-quarters of companies run some form of exit interview for at least some leavers. The instrument exists, is budgeted, and is already on the calendar of every departure.
It is filed
HR owns the conversation in seven programs out of ten. The record enters an enterprise system built to hold a leaver, not to describe a function.
It is consolidated
Most companies aggregate the data. Aggregation across a whole workforce is where a legal department of eight becomes a rounding error nobody can see.
It is not briefed
Fewer than a third of consolidating organizations regularly share the file with senior decision makers. The officer who owns the vacancy is not on the distribution list.
It changes nothing
Fewer than a third of program owners can name a policy change or an intervention that came out of an exit interview. Two-thirds of programs are talk with no follow-through.
It is never disclosed
Almost every large issuer talks about retention in the annual report. One in five publishes a number. The rest hand the audit committee an adjective.
A conversationA record
- The conversation What the leaver says, to whoever asks. Rich, unstructured, and worth nothing beyond the room unless somebody writes it down against the same fields every time.
- The consolidated file The same conversation folded into an enterprise data set with every other departure in the company. Searchable, reportable, and blind to a function of eight.
- The coded series The legal function’s own cut, held to one taxonomy over consecutive quarters, so the second year can be compared with the first. This is the artifact a chief financial officer treats as a number.
Nothing in that chain is expensive to fix, and none of it requires a new vendor.
Why the file fails the only audience that matters.
A general counsel asking for headcount is arguing against a finance function that already doubts the department’s contribution. An uncoded exit file is not an argument. It is an anecdote with a date on it.
The credibility gap is measured, and it is wide. The Thomson Reuters Institute’s 2026 State of the Corporate Law Department Report found 86% of general counsel saying their department is a significant contributor to business success, against 17% of other C-suite executives who agree — and 42% of those executives saying legal contributes little or not at all. In the same report, close to half of general counsel named staffing and resource constraints as the top barrier to delivering more value. Those two findings sit badly together: the department that says it is under-resourced is talking to a board that is not persuaded it is valuable.
The budget picture is the same story from the other side. The Thomson Reuters Institute’s 2025 Legal Department Operations Index, surveyed in July 2025 across 128 US legal-department respondents, found 81% reporting increasing matter volumes against 55% reporting flat or decreasing budgets, with 56% describing the department as under-resourced. The 2026 ACC Chief Legal Officers Survey, published in January 2026 on 1,049 participants, found 63% expecting headcount to stay flat and 35% naming chronic budget and resource constraints as the top barrier to success. CLOC’s March 2026 release on 135 departments at a median revenue of $13 billion put the share expecting attorney headcount growth at 32%.
Against that, an exit file that says “better opportunity” is worse than useless, because it invites the conclusion the finance function already holds — that the departure was market gravity rather than a fixable condition inside the company. Work Institute’s coding of 14,215 exit interviews from 2024 found 76.3% of departures preventable, with career development at 18.9%, health and family at 12.4%, work-life balance at 11.9% and management behavior at 9.7%, the highest that last category had run in six years. Gallup’s July 2024 study of 717 voluntary leavers put the preventable share at 42% on a stricter self-report test, and found that 45% had had no conversation about satisfaction, performance or their future with any manager or leader in the three months before leaving.
Preventable is the word that changes a budget conversation, and it is the word an uncoded file cannot support. In Sartori’s New York cohort, 147 general counsel and chief legal officers were asked over an 18-month window whether they could produce a coded reason-for-leaving series for their own legal function covering more than one year. Nineteen could. Of the 94 respondents in that group running departments at companies above $1 billion in revenue, only 11 said that any version of the exit data had ever been put in front of the chief financial officer or the audit committee.
A head of legal at a US-listed industrial group put the mechanism plainly to us: the department had lost three lawyers in eighteen months, HR had run an interview each time, and when the request for a fourth headcount went to the finance committee, the only evidence available was the general counsel’s recollection of what two of those three people had said. The committee did not refuse the seat. It deferred it for a quarter, which cost more than a refusal would have.
One contrast is worth a sentence, and only a sentence, because it belongs to a different market. Law firms code the regret. Above the Law reported in April 2026 on the NALP Foundation’s calendar-year 2025 attrition study that firms classified 47% of associate departures as unwanted and 27% as desired, against overall attrition of 19%. That is a crude taxonomy, but it is a taxonomy: it separates the exits a firm minded from the exits it did not. The in-house side, which loses far fewer people and can afford to read every one of them, generally keeps no equivalent code at all.
An uncoded exit file is not an argument. It is an anecdote with a date on it.
What one departure costs, seat by seat.
Replacement cost is the translation layer between a resignation and a budget line. The published multiples exist, the published pay medians exist, and almost no legal department has ever put the two together for its own seats.
Gallup’s July 2024 analysis of voluntary turnover puts the cost of replacing a leader or manager at around 200% of salary, a professional in a technical role at around 80%, and a frontline employee at around 40%. Work Institute’s 2025 Retention Report uses a deliberately conservative 33% of base wages as its floor across an entire workforce — useful as a lower bound and wrong for a lawyer. Neither publisher produces a legal-department figure, and there is no reason to wait for one.
The pay side is published in detail. The ACC and Empsight 2025 Law Department Compensation Survey, released in September 2025 with data effective 1 March 2025 across 1,632 self-reported US in-house respondents, puts median total cash at $160,000 for an Attorney, $228,000 for a Senior Attorney, $260,000 for an Expert Attorney, $294,000 for an Associate General Counsel, $368,000 for a Deputy General Counsel and $410,000 for a General Counsel or Chief Legal Officer, with median total target direct compensation for that top seat at $503,000. The same survey records a size step that matters to a backfill: chief legal officers at companies above $5 billion in revenue earn 44% more in base salary and 173% more in total target compensation than their counterparts below $1 billion.
Apply the published multiple to the published median one title at a time and the arithmetic holds. Blend them into a single average cost of losing a lawyer and it stops being defensible, because the professional multiple and the leadership multiple describe different mechanisms — one is a hiring and ramp cost, the other is a continuity and decision-quality cost. The seat-by-seat version is the one that survives a finance review, and it is the one nobody has been asked to produce because nobody has the coded series to attach it to.
Sartori’s mandate telemetry gives the timing side of the same number. Of the 24 New York in-house searches closed over the trailing three years, 9 were replacements for a lawyer who had left a seat that had existed for less than three years, and the median span from offer to signed acceptance across the whole set was 16 working days against a typical four-to-seven-month process. Counter-offer incidence across those New York in-house mandates ran at 28%. Read the two together and the shape of the loss is obvious. The expensive part of a departure is not the search. It is the seat sitting open, the work redistributed to people who did not have room, and the second departure that follows from the first.
A head of talent at a private technology company in the $500 million to $1 billion band described the same thing from the HR side. Her exit form had captured “compensation” as the reason a commercial counsel left. What the coded version would have captured, she told us, was that the leaver had seen the posted range for a peer requisition, worked out the compression, and started looking that week — a specific, dated, fixable event that the enterprise picklist rendered as a single word.
Attorney
80% of a $160,000 median total cash figure. The individual-contributor seat, where the cost is hiring, ramp and the matters that wait.
Gallup, July 2024 (professional multiple); ACC and Empsight, September 2025 (median total cash)The expensive part of a departure is not the search. It is the seat sitting open.
Since 2020 the exit file has become a record other people read.
Fourteen dated instruments now touch what a departing lawyer said, or was prevented from saying. None of them requires a new interview. All of them assume somebody read the last one.
The disclosure end came first. The Securities and Exchange Commission adopted the human capital requirement in Item 101(c) of Regulation S-K on 26 August 2020, requiring a description of human capital resources including any measures the registrant focuses on in managing the business, with the development, attraction and retention of personnel named as an example rather than a mandated line. It is principles-based on purpose, and that is the opportunity: the issuer designs the metric. Four annual reporting cycles later, Gibson Dunn’s December 2024 survey of S&P 100 Form 10-Ks found 96% of those companies discussing attraction and retention qualitatively and 19% disclosing a specific employee turnover rate, a share that had not moved in three years. The Commission’s Investor Advisory Committee asked on 21 September 2023 for turnover and comparable workforce-stability metrics to be prescribed; no rule followed. The taxonomy remains the company’s to write, and the legal function that has already written one is holding the draft.
The enforcement end arrived next. The Department of Justice’s Evaluation of Corporate Compliance Programs, updated in September 2024, instructs prosecutors to ask whether a company has a process for tracking lessons learned from its own prior issues and from peers, whether policies and training are updated to reflect them, and what the company communicates when an employee is exited or disciplined for a compliance failure. The Criminal Division’s Corporate Whistleblower Awards Pilot Program, launched in August 2024 and still maintained on a page updated in July 2026, gives a company 120 days from an internal report to self-disclose and stay eligible for a declination. Those two together change the character of the last conversation with a departing lawyer. It is intake, on a clock, not color.
The employment overlay is denser still, and in New York it is unusually specific. Labor Law § 740, as amended effective 26 January 2022, protects former employees and independent contractors who reasonably believe they reported a violation of any law, rule or regulation, with remedies that now include front pay and punitive damages and a limitation period doubled to two years. The Speak Out Act, signed 7 December 2022, makes pre-dispute nondisclosure and non-disparagement clauses unenforceable as to sexual harassment and assault claims. General Obligations Law § 5-336, as amended effective 17 November 2023, voids a discrimination release that ties liquidated damages, forfeiture or a no-discrimination affirmation to a confidentiality term. And the National Labor Relations Board held in McLaren Macomb on 21 February 2023 that merely offering a severance agreement with sweeping confidentiality and non-disparagement language violates Section 8(a)(1) where it tends to restrain protected activity — a holding that reaches non-supervisory staff, which in a legal function means paralegals, legal-operations analysts and some junior counsel.
Pay and privacy close the ring. New York Labor Law § 194-b, in force since 17 September 2023, requires a good-faith minimum and maximum range on every posting for work performed at least in part in the state, which means the backfill advertises the band the leaver was inside. New York City’s Council overrode a mayoral veto on 4 December 2025 to require annual pay-data reporting from private employers with more than 200 employees working in the city, once the designated agency publishes a form. The employment and business-to-business exemptions in the California Consumer Privacy Act expired on 31 December 2022, so exit notes, recordings, reason codes and inferences about a California-resident lawyer carry access, deletion and correction rights. And the Trapped at Work Act, signed and effective on 19 December 2025, bans employment promissory notes and training-repayment clauses that charge a worker for leaving early — retiring one of the retention devices a legal department might otherwise reach for.
Two more sit at the edges and still belong in the register. Executive Order 14173 of 21 January 2025 revoked Executive Order 11246 and required federal-contract certifications on program legality; the Supreme Court’s unanimous decision in Ames v. Ohio Department of Youth Services on 5 June 2025 removed the heightened burden a majority-group Title VII plaintiff had faced in some circuits. Together they make an exit comment about a diversity program two-way evidence rather than a one-way risk. And the European Union’s political agreement on Omnibus I of 9 December 2025 narrowed who must file a sustainability report under the Corporate Sustainability Reporting Directive without repealing workforce reporting for those still in scope, with reporting beginning for financial years from 1 January 2027 for EU groups and 1 January 2028 for non-EU parents.
There is one rule underneath all of them that applies only to this population. Under ABA Model Rule 1.13, a lawyer employed by an organization represents the organization rather than its officers, with an obligation to escalate to the highest authority where a constituent is violating a legal obligation likely to cause substantial injury. A departing in-house lawyer may therefore be carrying information the organization is entitled to and that a generic HR questionnaire is not designed to receive, from a person whose professional duty ran to the company rather than to the manager conducting the interview. That is the strongest argument for the legal function owning the read, and it has nothing to do with analytics.
An HR courtesyAn evidentiary record
- The exit form Collected as good practice, stored as a personnel document, discussed nowhere. Its only reader is the person who filed it.
- The retained file Subject to a retention schedule, reachable by a data-subject request, and discoverable. It is now a company record whether or not anyone treats it as one.
- The disclosed and defended record Cited in an annual report, produced to a regulator, tested by opposing counsel. At this point the question is not what was collected but who read it and what was done.
| Instrument | Dated | What it asks of the file | Owner in the department |
|---|---|---|---|
| Regulation S-K Item 101(c) | Adopted 26 August 2020 | A description of human capital measures the registrant focuses on in managing the business, with retention named as an example | Deputy general counsel for disclosure, with the disclosure committee |
| SEC Investor Advisory Committee recommendation | Approved 21 September 2023 | Turnover or comparable workforce-stability metrics prescribed in Item 101(c); no rule was proposed, so the taxonomy stays the issuer’s to design | Deputy general counsel for disclosure |
| DOJ Evaluation of Corporate Compliance Programs | Updated September 2024 | A process for tracking lessons learned from the company’s own prior issues, and what is communicated when someone is exited for a compliance failure | Investigations or compliance counsel reporting to the legal chief |
| DOJ Corporate Whistleblower Awards Pilot Program | Launched August 2024; page updated 31 July 2026 | A 120-day clock from an internal report to self-disclosure for declination eligibility | Investigations counsel plus the intake owner for exits |
| New York Labor Law § 740 | Amended effective 26 January 2022 | Protection for former employees who reasonably believe they reported a legal violation, with a two-year limitation period | In-house employment counsel |
| Speak Out Act | Signed 7 December 2022 | Pre-dispute nondisclosure and non-disparagement clauses unenforceable as to sexual harassment and assault claims | In-house employment counsel, on templates |
| New York General Obligations Law § 5-336 | Amended effective 17 November 2023 | No liquidated damages, forfeiture or no-discrimination affirmations attached to confidentiality in a release, or the release fails | In-house employment counsel, on separation agreements |
| NLRB, McLaren Macomb | Decided 21 February 2023 | Offering broad confidentiality and non-disparagement terms to non-supervisory employees is itself unlawful | Labor counsel, on severance language |
| CCPA as amended by the CPRA | Employee exemption expired 31 December 2022 | Access, deletion and correction rights over exit notes, recordings, reason codes and inferences about California residents | Privacy counsel with legal operations |
| New York Labor Law § 194-b | In force 17 September 2023 | A good-faith minimum and maximum range posted for every job, promotion and transfer performed at least in part in New York | Employment counsel with total rewards |
| New York City Int. 982-A and 984-A | Veto overridden 4 December 2025 | An annual pay-data report from private employers with more than 200 employees working in the city, once the designated agency publishes a form | Employment counsel with legal operations |
| New York Trapped at Work Act | Signed and effective 19 December 2025 | No employment promissory notes or training-repayment clauses that charge a worker for leaving early | Employment counsel, on offer letters |
| Executive Order 14173 and Ames v. Ohio | 21 January 2025 and 5 June 2025 | Federal-contract certifications on program legality, and no heightened burden for a majority-group Title VII plaintiff | Employment litigation counsel |
| CSRD as narrowed by Omnibus I | Political agreement 9 December 2025 | A workforce report from in-scope groups for financial years from 1 January 2027, or 1 January 2028 for non-EU parents | Sustainability or employment counsel with legal operations |
It is intake, on a clock, not color.
What a legal-function exit interview asks that the enterprise form does not.
The gap is not sophistication. It is specificity: eight fields the company already almost collects, asked in the language of a legal department, and held constant long enough to compare one year with the next.
Start from what the enterprise instrument is for. A company-wide exit questionnaire is built to be answerable by a warehouse supervisor and a sales director on the same form, which is why its reason list resolves to categories like career, pay and balance. Work Institute’s coding of 14,215 exit interviews from 2024 produced exactly those categories, and they are the right categories for a workforce. They are not specific enough to change anything in a department of eight lawyers, because every one of them is true of every leaver in every function.
The legal-function cut asks the same questions in the vocabulary of the seat. Not “did you have development opportunities” but “which matters did you carry with no second chair, and for how long.” Not “how was your workload” but “what was your contract-approval threshold and when did it last change.” Not “was compensation a factor” but “had you seen the posted range for a comparable requisition before you resigned.” The last of those is a New York-specific question because Labor Law § 194-b has required a good-faith range on postings since 17 September 2023, which means the comparator is public and the leaver has almost certainly read it.
The other half of the build is who holds it. The Corporate Legal Operations Consortium and Harbor 2025 State of the Industry Report, covering 186 departments, records legal-operations leaders reporting to the chief legal officer or general counsel at 56%, and lists the services in their scope: outside counsel and vendor management at 95%, technology strategy at 91%, program management at 84% and financial management at 80%. Talent management comes last, at 32%. The function that would naturally code a lawyer’s exit for the general counsel is built for vendors, technology and spend, and in two departments out of three has no formal mandate over people at all. That is a scope decision, not a capability gap, and it is reversible in a memo.
Gallup’s standing guidance from June 2018 on running the instrument properly still holds on two points worth repeating: the departing person’s direct manager should not be the interviewer, and the exit data is only useful when it is combined with the operating indicators the organization already tracks. For a legal department that means the coded reason sits next to matter volume, outside-counsel spend and the age of open requisitions — the numbers the general counsel is already reporting.
The enterprise owns itThe department owns the read
- Human resources runs the process Scheduling, the conversation, the record and the retention schedule. This is where the process belongs and where it should stay, because independence from the leaver’s manager is the whole point.
- Legal operations codes the cut The same fields, every quarter, for legal-function departures only. Not a second interview and not a second system — a second reading of the one that already happened.
- The legal chief carries the finding To the finance committee with a cost attached, to the disclosure committee with a series attached, and to the compliance function when something in the file needs escalating.
| Field | The enterprise form | The legal-function cut |
|---|---|---|
| Reason for leaving | One picklist value chosen by the leaver from an enterprise list built for the whole workforce. | Primary and secondary reason, plus the counterfactual: the specific change that would have prevented the resignation, dated. |
| Scope | Job title and grade. | Matters carried alone, delegated authority in dollars, contract-approval threshold, whether a second chair existed for the largest recurring matter. |
| Reporting line | Manager name, captured for routing. | Whether legal reported to the chief executive or through finance, and whether the leaver had unfiltered access to the audit committee. |
| Pay | Free text, usually the word “compensation.” | Band held on exit against the posted range for the backfill, plus whether the leaver had seen that range before resigning. |
| Load | Optional engagement score. | Hours, matter count and the trailing 12-month change in both, coded against the enterprise workload field so the two can be read together. |
| Anything reported | Not asked, or asked as a compliance checkbox. | Any internal report the leaver made that was not closed out, routed same-day to the officer who owns investigations, with a written receipt. |
| Destination | Competitor yes or no. | Seat type taken: another in-house function, a law firm, a regulator, or out of practice — the four exits that mean four different things. |
| Retention attempt | Rarely captured. | Whether anyone in the department discussed satisfaction, career or future in the three months before notice, and who. |
- Q1 Can you produce last year’s legal-function departures by coded reason, in writing? No → you have a filing habit, not a series. Start by coding the exits you already have on record.
- Q2 Is there a named person inside the department who receives every legal exit within a fixed number of days? No → the file will keep stopping at the enterprise system. Name the owner before you design the form.
- Q3 Does each coded reason carry a replacement cost for that specific seat? No → the finding will not survive contact with a finance committee. Attach the multiple and the band.
- Q4 Is there a same-day route for anything in an exit that touches compliance, retaliation or a report never closed out? No → you are holding evidence in a personnel file. Build the route first; it is the cheapest item on this list.
- → All four yes? You are running talent intelligence. Everyone else is running an archive.
Not a second interview and not a second system — a second reading of the one that already happened.
What the first coded year actually shows.
Three things, on the evidence of departments that have done it and of the mandates that follow a departure: the reason is usually structural, the cost is usually understated, and the instrument is at its weakest exactly where it matters most.
Sartori has worked the New York in-house market for more than ten years, for listed issuers, private companies and funds, and the searches that follow a departure look different from the searches that follow growth. Across the 24 New York in-house searches closed over the trailing three years, at a 93% completion rate, the process ran four to seven months end to end with a median of 16 working days from offer to signed acceptance. Counter-offer incidence across those mandates was 28%, which is the number that tells a general counsel something the exit file will not: more than a quarter of the time, the person’s current employer turns out to have been willing to move on terms it had never offered before the resignation landed.
Here is the finding that does not flatter us. Of those 24 closed mandates, we hold a documented reason-for-leaving in the outgoing lawyer’s own words in 7. In the other 17, the reason we recorded came from the hiring company — a description of a vacancy rather than an account of a departure. Our own file is thinner than the argument this page makes, for the same reason every corporate file is thin: the departing lawyer is a person you stop having a reason to call the moment the seat is filled. Any search desk that tells you it holds a complete departure record for every mandate it has run is describing a system it does not have.
The quarterly survey wave gives the second read. Sartori has run quarterly market surveys since 2019, and in the most recent New York wave the in-house respondents who had changed seats in the previous year named the scope of the seat — what it owned, what it signed off, who it reported to — ahead of pay as the reason for the move, with pay named as the reason the move became possible rather than the reason it started. That distinction is invisible on any form with one reason field, and it is the difference between a retention response that costs a salary review and one that costs a reorganization of the department.
The third finding is the uncomfortable one for the instrument itself. The exits that most need reading are the ones least likely to be honestly recorded: the lawyer leaving because of a specific person, the lawyer leaving after raising something that was never closed out, the lawyer who has already been advised not to say anything. Gallup’s July 2024 leaver study found only 7% of people felt their manager had made any effort to retain them, and that 36% had not spoken to anyone before deciding to resign. A file built on candor from people who have already concluded that candor has no audience will be systematically missing its most valuable rows. The correction is structural, not motivational: an interviewer who is not the leaver’s manager, a conversation held after the last day rather than during the notice period, and a written route for anything that needs escalating.
For the legal chief, this is a budget instrument that happens to be about people.
- Ask for the last three years back. The interviews were conducted and the records exist. Coding what has already been collected costs nothing and produces a baseline in a fortnight.
- Put a dollar on every coded reason. A finance committee that has never been given a replacement cost for a Senior Attorney has no way to price a delay, and will price it at zero.
- Take the series to the disclosure committee, not just to the board. Item 101(c) has needed a defensible retention measure since 2020 and one large issuer in five has one.
- Read the destination, not just the reason. A lawyer who goes to another in-house function, to a law firm, to a regulator or out of practice is telling you four different things about your seat.
For the head of HR, nothing here is a request to hand over the process.
- Keep the interview. Independence from the leaver’s manager is the reason the instrument works at all, and a general counsel interviewing a departing direct report would destroy it.
- Add a function field and a routing rule. The whole change is that legal-function departures also go to a named person inside the department, on a fixed cadence.
- Flag the legal-risk content immediately. A whistleblower report, a retaliation allegation or an unclosed internal concern in an exit conversation is not an HR data point; it starts a clock.
- Expect the same request from finance and engineering next. The coded functional cut is not a legal privilege; legal is simply the function whose regulator asks for it first.
Three composites from mandates that began with a departure
A listed industrial group, legal function of eleven. Two Associate General Counsel seats turned over inside twenty months. Each departure had been interviewed by corporate HR and each record sat in the enterprise system; neither had reached the general counsel in a form he could use. The second search surfaced what the first exit had already said, which was that a single regulatory matter had been carried by one person with no second chair for three years. The mandate closed inside the four-to-seven-month band, and the department restructured coverage of that matter before the new hire started — a decision available twenty months earlier, at the price of reading one file.
A private technology company below $1 billion in revenue, three lawyers. The general counsel was, in practice, the department. When the sole senior commercial lawyer resigned, the exit conversation happened between HR and the leaver and never came back to legal at all. The coded reason would have been approval-threshold creep: the seat had been given sign-off authority twice without a corresponding change in title or band. The replacement search took a counter-offer at the first offer stage and closed at the second, inside the 28% counter-offer incidence this desk records for New York in-house mandates.
A financial-services group, department of thirty-plus. A privacy counsel raised an unresolved internal reporting concern in the exit conversation. It was recorded accurately, filed correctly, and read by nobody with authority to act for four months. Nothing came of it in the end, which is the point: the company was fortunate rather than protected, on a timetable that would not have survived a 120-day self-disclosure window. The routing rule that closed the gap took one page and no budget.
A file built on candor from people who have already concluded that candor has no audience will be systematically missing its most valuable rows.
Common questions about exit interviews in a legal function
What should an exit interview ask a departing in-house lawyer?
Eight fields the enterprise form does not: scope, sign-off authority, matters carried alone, escalations, pay band against the posted range for the backfill, and anything reported but never closed. Each of those is a condition somebody inside the company could have changed, which is what makes it worth asking. Work Institute’s 2025 Retention Report, coding 14,215 exit interviews from 2024, found 76.3% of departures preventable, with career development at 18.9%, work-life balance at 11.9% and management behavior at 9.7%, a six-year high — but those are enterprise buckets. A legal function needs the buckets translated into its own terms: scope creep, matter concentration, no second chair, a reporting line into the CFO rather than the chief executive. Ask what would have had to be true in June for the resignation not to have happened in September.
How many exit interviews does a corporate legal department actually run in a year?
Fewer than most assume: the Thomson Reuters Institute’s 2026 report puts the median in-house team at 3 lawyers below $1 billion, 8 to $6 billion and 30 above it. Three-quarters of companies in that smallest band run five attorneys or fewer, on more than 2,300 general counsel interviews. Applied against published voluntary-turnover rates for professionals — Mercer’s 2025 US Turnover Survey, republished by CBIA in October 2025, records 9.1% for non-sales professionals and 5.2% for executives — a mid-market department produces a single-digit number of legal exits in a year and a large one a low-double-digit number. No public source publishes an exits-per-legal-department count, so it is a band and not an estimate. The scarcity is the point: at that volume every interview is individually readable, and it still is not read.
Who should own exit interviews for the legal function — HR or the general counsel?
HR should keep the process; the general counsel should own the read, because Harvard Business Review found in April 2016 that 70.9% of programs sat inside HR. That study, by Spain and Groysberg across 210 organizations, also found fewer than a third of executives whose companies had a program able to name a single action taken as a result. Their sharper finding is the one that decides ownership: most companies consolidated the data, and fewer than a third of those organizations regularly shared it with senior decision makers. A consolidated file that stops at the HR system is invisible to the officer who has to explain a vacancy to the board. The fix is not another survey instrument. It is a named owner inside the department — usually the deputy general counsel or the legal-operations lead — who receives every legal-function exit within a fixed number of days and codes it against the same taxonomy every quarter.
What does it cost to replace an in-house lawyer?
Between roughly $128,000 and $1 million, depending on the seat. Gallup’s July 2024 analysis puts replacement of a technical or professional role at about 80% of salary and of a leader or manager at about 200%. The ACC and Empsight 2025 Law Department Compensation Survey, published September 2025 on 1,632 self-reported US in-house respondents, puts median total cash at $160,000 for an Attorney, $228,000 for a Senior Attorney, $294,000 for an Associate General Counsel and $410,000 for a General Counsel or Chief Legal Officer, with median total target direct compensation for that top seat at $503,000. Apply the published multiple to the published median one title at a time and the arithmetic is defensible; average the seats into a single figure and it is not. That is the line item the chief financial officer has never seen, because nobody in the department produced it.
Do exit interviews create legal risk for the company?
They create evidence, not liability: New York Labor Law § 740 has protected former employees who reasonably believe they reported a legal violation since 26 January 2022. The limitation period is two years and the remedies now reach front pay and punitive damages, so an exit conversation that captures a legal concern is protected activity rather than an HR note. The Speak Out Act, signed 7 December 2022, makes pre-dispute nondisclosure and non-disparagement clauses unenforceable as to sexual harassment and assault claims, and the National Labor Relations Board held in McLaren Macomb on 21 February 2023 that merely offering a severance agreement with broad confidentiality and non-disparagement terms violates Section 8(a)(1) for non-supervisory employees. The risk is not in listening. It is in a file nobody reads until a plaintiff’s lawyer reads it first.
Can exit interview data go into the 10-K human capital disclosure?
Yes, and almost nobody does: Gibson Dunn’s December 2024 survey of S&P 100 annual reports found only 19% disclosing a turnover rate. The same survey found 96% discussing attraction and retention qualitatively, a split that has not moved across three cycles. Item 101(c) of Regulation S-K, adopted by the SEC on 26 August 2020, requires a description of human capital resources including any measures the registrant focuses on in managing the business, and names retention as an example rather than a mandated line. The Investor Advisory Committee asked the Commission on 21 September 2023 to prescribe turnover as a metric; no rule was proposed. So the taxonomy stays the issuer’s own to design, and the legal function that codes its exits consistently is the one that can hand the disclosure committee a defensible series instead of an adjective.
Sources.
The instrument’s own record comes from one Harvard Business Review field study of 210 organizations and three later reads: a third-party file of coded exit interviews, two Gallup leaver and chief-HR-officer surveys, and a survey of S&P 100 annual reports. Department size, pay and mobility come from the Association of Corporate Counsel, the Thomson Reuters Institute and the Corporate Legal Operations Consortium. The rules come from the Commission, the Department of Justice, the Labor Board, the New York statute book and the City Council.
Sources & further reading
44 references- Sartori & Partners — New York Legal Talent Research Programme (1,675 structured interviews; ~67,000 lawyers mapped; quarterly surveys since 2019; mandate telemetry) sartoriglobal.com ↗
- Making Exit Interviews Count (Spain and Groysberg, Harvard Business Review, April 2016) cdn.prod.website-files.com ↗
- Making Exit Interviews Count hbr.org ↗
- Why Employees Quit (Harvard Business Review, November–December 2024) hbr.org ↗
- 2025 Retention Report: Employee Retention Truths in Today's Workplace (Work Institute) info.workinstitute.com ↗
- 42% of Employee Turnover Is Preventable but Often Ignored (Gallup) gallup.com ↗
- Enhancing the Employee Exit Experience Is Worth It (Gallup) gallup.com ↗
- The Real Value of Getting an Exit Interview Right (Gallup) gallup.com ↗
- U.S. Worker Thriving Declines as Job Market Pessimism Grows (Gallup) gallup.com ↗
- 2025 Law Department Compensation Survey — Executive Summary (ACC and Empsight International, LLC) acc.com ↗
- The State of Stress Among In-house Legal Professionals (ACC, December 2025) acc.com ↗
- 2026 ACC Chief Legal Officers Survey — Key Findings acc.com ↗
- 2024 ACC Chief Legal Officers Survey — Key Findings acc.com ↗
- US In-house Counsel Population Statistics: Analyzing Growth and Market Trends (ACC) acc.com ↗
- 2026 State of the Corporate Law Department Report (Thomson Reuters Institute) legal.thomsonreuters.com ↗
- 2025 State of the Corporate Law Department Report (Thomson Reuters Institute) legal.thomsonreuters.com ↗
- 2025 Legal Department Operations Index (Thomson Reuters Institute with Buying Legal Council) thomsonreuters.com ↗
- CLOC 2025 State of the Industry Report (CLOC | Harbor) cloc.org ↗
- CLOC Releases 2026 State of the Industry Report cloc.org ↗
- Lawyers : Occupational Outlook Handbook (U.S. Bureau of Labor Statistics) bls.gov ↗
- Job Openings and Labor Turnover — July 2026 (U.S. Bureau of Labor Statistics) bls.gov ↗
- How Much Turnover Is Too Much? (Mercer 2025 U.S. Turnover Survey, republished by CBIA) cbia.com ↗
- SEC Adopts Rule Amendments to Modernize Disclosures of Business, Legal Proceedings, and Risk Factors Under Regulation S-K sec.gov ↗
- Four Years of Evolving Form 10-K Human Capital Disclosures (Gibson Dunn) gibsondunn.com ↗
- Recommendation of the SEC Investor Advisory Committee's Investor-as-Owner Subcommittee regarding Human Capital Management Disclosure sec.gov ↗
- Evaluation of Corporate Compliance Programs (U.S. Department of Justice, Criminal Division, Updated September 2024) justice.gov ↗
- Key Updates to the DOJ's Evaluation of Corporate Compliance Programs (Skadden) skadden.com ↗
- Criminal Division Corporate Whistleblower Awards Pilot Program (U.S. Department of Justice) justice.gov ↗
- New York Expands Whistleblower Protections Under Labor Law Section 740 kmm.com ↗
- "Speak Out Act" is Now Law: Pre-Dispute NDAs and Non-Disparagement Agreements Are No Longer Enforceable for Claims Involving Workplace Sexual Harassment or Assault sgrlaw.com ↗
- New York Enacts Greater Restrictions on Release Agreements Involving Claims of Discrimination, Harassment, or Retaliation (Littler) littler.com ↗
- Employers can't offer severance agreements with nondisclosure clauses, NLRB says (HR Dive) hrdive.com ↗
- Governor Hochul Announces Groundbreaking Statewide Pay Transparency Law Takes Effect (New York State Department of Labor) dol.ny.gov ↗
- Ringing In the New Year for New York Employers: What's Next for Compliance in 2026? (Duane Morris) duanemorris.com ↗
- New employment law trend: Restrictions on stay-or-pay provisions in employment agreements (McDermott Will & Schulte) mcdermottlaw.com ↗
- California Consumer Privacy Act (CCPA) — Office of the Attorney General oag.ca.gov ↗
- Ending Illegal Discrimination And Restoring Merit-Based Opportunity (Executive Order 14173) whitehouse.gov ↗
- Ames v. Ohio Department of Youth Services (SCOTUSblog case page) scotusblog.com ↗
- EU Reaches Agreement on 'Omnibus I' Impacting CSRD and CSDDD Compliance for US Companies (Cooley) cooley.com ↗
- Rule 1.13: Organization as Client (ABA Model Rules of Professional Conduct) americanbar.org ↗
- Associates Are Leaving Faster Than Ever (Above the Law, on the NALP Foundation's calendar-year 2025 attrition study) abovethelaw.com ↗
- Sartori & Partners — Building an in-house legal team ↗
- Sartori & Partners — General counsel salary 2026 ↗
- Sartori & Partners — Legal operations career path ↗
Three different bases appear in the prevalence figures and they measure different things: the 2016 field study reports company-level practice across 210 organizations, the 2025 Retention Report's 6% figure is a post-departure cadence on a panel of 88 HR leaders, and the S&P 100 survey counts published annual reports. The in-house population figure is a residual derived from Bureau of Labor Statistics data rather than a count of in-house titles, and the compensation figures are self-reported by individuals rather than submitted by employers.
No public source counts how many exit interviews United States corporate legal departments complete in a year, how many are coded into a reason taxonomy, or how many reach a chief financial officer. The nearest published measures are department-size medians drawn from general counsel interviews and voluntary-turnover rates by career stream drawn from an all-industry employer survey; they describe different populations and no publisher reports them as a product. The replacement multiples come from one publisher and the pay medians from another, which is why each seat on the axis above carries both inputs by name. The law-firm attrition figures belong to a different market and are reported for calendar year 2025.
Size the function. Price the seat. Then read the file.
The exit file sits on top of a department that was scoped by someone, paid against a band, and run by a function with a defined mandate. These are the adjacent pieces.
Building an In-House Legal Team
The build-side companion to this read: how the first seats are scoped, what the second hire is really for, and the sequencing that keeps a small department from becoming one person's single point of failure.
Read the build guideGeneral Counsel Salary 2026
The pay reference behind the replacement arithmetic here, seat by seat and band by band — the number the posted range for your backfill will be measured against.
See the pay bandsLegal Operations Career Path
Who ends up owning the coded read in practice: what a legal-operations lead is accountable for, where the role comes from, and why talent sits at the bottom of its scope.
Read the ops pathFor general counsel and heads of legal
Replacing a lawyer who left for a reason nobody wrote down?
We run in-house legal searches in New York and internationally, and the first conversation is usually about the seat rather than the shortlist — what it owned, why it opened, and whether the backfill should be the same job. Quiet, evidence-led, and just as willing to say the seat needs redesigning as to run the search.