River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

Employee Handbook Template and Tracker

Five documents and four sheets, including a coverage sheet that reports each policy against the employees who acknowledged a version actually containing it.

Free download  ·  No account needed

What Effective Coverage Is

One number says 99.2 percent. Seven policies are under half.

Marloe Logistics, 240 employees across seven states. Handbook v5 effective 1 July 2026. The last company-wide re-acknowledgment campaign ran at v3 in March 2025, and v4 and v5 went out by email.

What is being measuredQuestion it answersFigure
Signed acknowledgment on fileDid this person ever sign something?238/240  99.2%
Acknowledged the current versionDid this person sign v5?22/240  9.2%
Effective coverage, per policyFor this policy as it reads today, did this person acknowledge a version that contained this text?0.0% to 100%

The seven policies below half

PolicyApplies toLast amendedCoveredCoverage
Colorado paid sick leaveColoradov50 of 90.0%
Arbitration and dispute resolutionEvery statev522 of 2409.2%
Remote work expense reimbursementEvery statev522 of 2409.2%
Anti-harassment and reportingEvery statev441 of 24017.1%
Washington paid family and medical leaveWashingtonv41 of 520.0%
Pay transparency in postingsCA, NY, WA, COv412 of 4427.3%
California meal and rest breaksCaliforniav47 of 1838.9%

Every one of the seven was amended in v4 or v5, and that is the finding rather than a coincidence: the most recently rewritten policy always has the lowest coverage, and it was usually rewritten because it was the exposure. Marloe's arbitration clause was rewritten in v5 because counsel had called the old one probably unenforceable. Twenty-two people are bound by the new one. Invoke it against any of the other 218 and the clause that governs is the one counsel already said would not hold.

The other 27 policies are the reason this is not just a re-sign

Most of a handbook does not change between versions. Twenty-seven of Marloe's 34 policies read identically in v3 and v5, so a v3 signature still binds on all of them. That is what makes the gap small enough to close: seven policies, not thirty-four.

Every handbook report ends at the same number. Some share of employees have a signed acknowledgment on file, it is nearly always above 95 percent, and it stays there because collecting it is on the onboarding checklist. That number answers whether somebody ever signed something. It does not answer the question a dispute asks, which is narrower: for the policy now being invoked, did this employee acknowledge a version that actually contained this text?

A handbook is a sequence of versions, and an employee is bound by the one they signed. Marloe Logistics, a fictional 240-person warehouse software company, reports 99.2 percent acknowledged. Per policy, coverage runs from 0.0 percent to 100 percent. Its arbitration clause was rewritten in v5 because counsel called the old one probably unenforceable, and 22 people are bound by the new one. Its Colorado sick leave policy was rewritten for nine employees, none of whom have acknowledged it.

The useful part is what that does to the fix. Most of a handbook does not change between versions, so 27 of Marloe's 34 policies read identically in v3 and v5 and a v3 signature still binds on all of them. The real gap is seven policies, and the ask differs by state: three in Ohio, four in New York, five in California. Scoping is a legal question, not a formatting one, which is also why pay range disclosure is not a national paragraph.

Thirty-four policies, seven states, five versions, one coverage figure each

The four sheets in the pack. Every figure comes from the same 240-person roster.

Policy Register

One row per policy, not per section. A section cannot carry a scoped acknowledgment, so the register is what makes the rest of the arithmetic possible. Thirty-four policies, five versions, seven states.

PolicyApplies toApplies to
employees
Last
amended
What changed in it
Arbitration and dispute resolutionEvery state240v5Clause rewritten after counsel flagged the old one as probably unenforceable
Remote work expense reimbursementEvery state240v5A fixed monthly allowance replaced case-by-case approval
Colorado paid sick leaveColorado9v5Accrual and carryover rewritten after the state's own requirements changed
Anti-harassment and reportingEvery state240v4Reporting routes rewritten to add a second channel that does not run through the manager
California meal and rest breaksCalifornia18v4Premium pay for a missed break added, and the waiver conditions written out
Pay transparency in postingsCA, NY, WA, CO44v4Range disclosure extended to internal postings
Washington paid family and medical leaveWashington5v4Contribution split and the top-up interaction corrected
FMLA general noticeEvery state240v3Rewritten around the Department's prototype text after an audit found the summary too thin
At-will employmentEvery state240v1Not amended since
34 policies: 8 last amended in v1, 7 in v2, 12 in v3, 4 in v4, 3 in v5. Nine are scoped to fewer than all seven states.

A policy's scope is a legal question and it is answered per policy. A federal rule is a floor, and state law giving greater rights is not displaced by it, so a single national text either understates rights in the strictest state or applies that state's rules to people who are not in it. Marloe's four-person Washington population gets the same treatment as the 128-person head office.

Effective Coverage by Policy

Covered means the employee acknowledged a version at or after the version that last amended this policy. The denominator is the employees the policy applies to, which is why a nine-person policy can read 0.0 percent.

PolicyAmendedApplicableCoveredCoverageNot covered
Colorado paid sick leavev5900.0%9
Arbitration and dispute resolutionv5240229.2%218
Remote work expense reimbursementv5240229.2%218
Anti-harassment and reportingv42404117.1%199
Washington paid family and medical leavev45120.0%4
Pay transparency in postingsv4441227.3%32
California meal and rest breaksv418738.9%11
Equal employment opportunityv324023196.2%9
FMLA general noticev324023196.2%9
Nevada paid leavev3615996.7%2
At-will employmentv124023899.2%2
New York paid family leavev31212100.0%0
Every policy under 50 percent was amended in v4 or v5. The 27 policies whose text is unchanged since v3 all sit above 96 percent, because 190 people acknowledged v3 in the March 2025 campaign.

The v3 floor of nine people is the same nine everywhere: two on v1 from January 2023, five on v2, and two with no acknowledgment on record. Seven of the nine show as signed on every report the company runs.

Acknowledgment Tracker

Grouped by state and by the version each group acknowledged, because those two together determine what a person still has to re-sign. Sorted by size of the ask.

StatePeopleVersionTo re-signWhich ones
Nevada1none28Whole handbook
Texas1none27Whole handbook
Ohio2v120Whole handbook
Nevada1v214Whole handbook
Ohio4v213Whole handbook
California11v35Anti-harassment, CA breaks, pay transparency, arbitration, remote expense
Colorado9v35Anti-harassment, pay transparency, arbitration, remote expense, CO sick leave
Washington4v35Anti-harassment, pay transparency, WA leave, arbitration, remote expense
New York8v34Anti-harassment, pay transparency, arbitration, remote expense
Ohio105v33Anti-harassment, arbitration, remote expense
Nevada49v33Anti-harassment, arbitration, remote expense
Texas4v33Anti-harassment, arbitration, remote expense
Five states19v42Arbitration, remote expense
Six states22v50Nothing outstanding
209 people on a targeted ask of 664 employee-policy pairs, plus 9 on a full re-sign. A blanket re-sign of the whole handbook would be 240 people times 34 policies, or 8,160 pairs.

Same state, different ask: the 105 Ohio employees on v3 owe three policies and the four on v2 owe thirteen. Same version, different ask: v3 in Ohio is three policies and v3 in Colorado is five. Neither difference is visible on a report that counts signatures.

Review Calendar

Cadence set by what actually ages: a state requirement, a contribution rate, a case-law question, or a figure in the text. Each row carries the coverage it would reset.

PolicyCoverage
today
CadenceNext dueWhy this cadenceCoverage reset
if amended
Colorado paid sick leave0.0%Annually, and on any state amendment15 Jan 2027State requirement changed once already inside this handbook's life9
California meal and rest breaks38.9%Annually, and on any state amendment15 Jan 2027Premium pay exposure per missed break18
Pay transparency in postings27.3%Annually, and whenever a new state is hired into15 Jan 2027Four states in scope today and the list grows with remote hiring44
Washington paid family and medical leave20.0%Annually, and on any contribution-rate change15 Jan 2027Contribution split changes by rate year5
Arbitration and dispute resolution9.2%Annually with counsel30 Jun 2027Enforceability turns on case law, not on the text staying still240
FMLA general notice96.2%Annually31 Mar 2027A defective notice can itself be treated as interference, so the text is the exposure240
At-will employment99.2%Every two years with counsel30 Jun 2027Wording that survives review does not need touching, and touching it restarts coverage240

The last column is the point of the sheet. Amending a policy resets its effective coverage to zero for everyone it applies to, so a small clarification to an undisputed policy can be worth holding until the next scheduled campaign, and a rewrite of the clause counsel is worried about cannot. That cost belongs in the decision to amend rather than in a discovery three months later.

What is in the pack

01

Policy Register

One row per policy rather than per section, carrying the version that last amended it, what changed, and the states it applies to. A section cannot support a scoped acknowledgment, which is why the register is the unit everything else is computed from, the same unit the underlying policy set itself is maintained against.

02

Effective Coverage by Policy

Applicable employees, employees who acknowledged a version at or after the amending one, and the rate. Both of the numbers normally reported sit at the bottom of the same sheet so the gap between them is visible rather than argued about.

03

Acknowledgment Tracker

Grouped by state and by acknowledged version, because those two together decide what each person still owes. Same version in two states is a different ask, and the sheet names the policies rather than the count.

04

Review Calendar

Cadence set by what actually ages in each policy, with the coverage each amendment would reset in the last column. Reviewing the performance policy annually is different from reviewing a contribution rate that changes by rate year.

05

What Effective Coverage Is

Three pages to read before the coverage sheet. The three things that have to be joined, why the most recently rewritten policy always has the lowest coverage, and why amending a policy is a coverage decision as well as a legal one.

06

Handbook

The assembled handbook by section, with each policy stamped with the version that last amended it and its jurisdiction scope. State-scoped policies sit in their own part rather than being folded into a national paragraph.

07

Acknowledgment Form

Two forms. A full-handbook form for a new hire or anyone below the last campaign, and a scoped form naming specific policies and versions, which is what makes a three-policy request answerable rather than another whole-document click.

08

Change Summary and Campaign

What changed and why, written per policy so it can be attached to a scoped request. The campaign document runs the targeted ask, including chasing it, which is the same problem as collecting compliance documents.

09

Why the reporting policy is usually the least covered

Marloe's anti-harassment policy was rewritten in v4 to add a reporting channel that does not run through the manager, and 199 people are still bound by the routes it replaced. What that costs shows up in a live investigation rather than on a compliance report.

How it works

  1. 1

    Send the handbook and its earlier versions

    Every version you still have, not just the current one. The version history is what makes a v3 signature meaningful on 27 policies and worthless on seven, so a handbook with no history gets a coverage figure nobody can defend.

  2. 2

    Send acknowledgments with the version, not the date

    Which version each person signed. If your records only say signed, reconstructing the version history is the real first task and the space will say so rather than computing a number it cannot stand behind.

  3. 3

    Scope each policy to the states it applies to

    One row per policy, with the rows that need counsel left open rather than guessed. The applicable population becomes the coverage denominator, which is why a four-person state matters as much as the head office.

  4. 4

    Coverage comes back per policy, then a targeted ask

    Expect the policies you rewrote most recently to be the least covered. The campaign that follows names the specific policies each person owes, which is a request that gets returned rather than ignored.

Frequently asked questions

Is the acknowledgment rate not already the compliance number?

It answers whether somebody ever signed something. Marloe reports 99.2 percent while its arbitration clause is acknowledged by 22 of 240 people, because the clause was rewritten in v5 and most signatures are on v3. In a dispute the version the employee signed is the one that governs.

Why not just have everybody re-sign the whole handbook?

Because it is 240 people against 34 policies and it produces one more click on a document nobody reads. Coverage says the real gap is seven policies: three in Ohio, four in New York, five in California and Colorado. A three-policy request with the changes named gets returned.

Why does the same version mean a different ask in two states?

Because scope is per policy. FMLA is a floor and state or local law providing greater family or medical leave rights is not superseded by it, so an employer complies with both. Marloe's v3 employees owe three policies in Ohio and five in Colorado.

Does anything have to be in the handbook by law?

The FMLA general notice does, if you have one. An FMLA-covered employer with eligible employees must include that notice in employee handbooks or other written guidance, where such materials exist, or hand it to each new hire. Failing the notice requirements can itself be treated as interference.

Our records only say signed, with a date. Can we still use this?

Partly. A date plus your version effective dates usually reconstructs the version, and that reconstruction is the first task rather than a blocker. Where it stays ambiguous the policy is reported as uncovered, because a coverage figure built on a guess is worse than a gap you can see.

What does the coverage figure change about amending a policy?

It puts a number on the amendment. Rewriting a company-wide policy resets its coverage to zero for all 240 people, so a small clarification to something undisputed can wait for the next scheduled campaign while a clause counsel is worried about cannot. The review calendar carries that column.

What format are the downloaded files, and is it free?

Five Word documents and four CSV sheets in one zip, free and with no account. The documents open in Word, Pages and Google Docs, the sheets in Excel, Numbers and Google Sheets. Edit with AI installs the same pack as a private Space and computes coverage against your own roster.

Find out which policies your signatures do not cover

Send the current handbook, whatever earlier versions you still have, your acknowledgment records with a version rather than just a date, and where your employees work. What comes back first is the coverage figure per policy, and the group whose signatures predate your last campaign.

Edit with AI