Most of the advice written about employee data is written for someone who does not exist in your business. It assumes an HR department, a headcount in the hundreds, and a reason to care about engagement scores and demographic dashboards. You have nine people, three of whom drive, and the employee data you own lives in a folder by the printer, a group chat, and a spreadsheet only you know how to sort.
That is a problem, but not the problem those articles think it is. The risk for a business that runs its own deliveries is narrower and much more concrete: the hours your drivers work are the single most heavily regulated thing about them, and the records proving those hours are the ones you are least likely to have. Route work is unpredictable by nature. A driver who was supposed to finish at four gets back at six, and whether that shows up correctly in your records is the difference between a routine complaint and a bill for back wages.
So this guide is about the employee data you actually have to keep, why each piece exists, and what the pieces tell you once you have them. Not the enterprise version. If you get to the end and decide you want something better than a spreadsheet holding all of it, an employee management system takes over the record-keeping and the dispatch side together, but the records come first, because software that automates a bad record is just a faster bad record.
The Bottom Line
- The Fair Labor Standards Act names 14 specific items you must keep for every non-exempt employee. Most of them you already have. The ones small operations miss are the day and time the workweek begins, and total hours for each workweek as a distinct figure (U.S. Department of Labor).
- Two retention clocks run at once: payroll records for at least three years, and the records wage calculations were built from (timesheets, schedules, rate tables) for at least two.
- If any vehicle you send out has a gross vehicle weight rating of 10,001 pounds or more and crosses a state line, your drivers need a federal driver qualification file on top of the wage records. Most cargo vans stay under that line. Box trucks often do not.
- The Department of Labor’s Wage and Hour Division recovered more than $259 million in back wages for nearly 177,000 workers in fiscal 2025, averaging about $1,465 a person (HR Dive).
- Separations in transportation and warehousing ran 4.2% in June 2026 and 3.8% in July — per month, not per year (Bureau of Labor Statistics). Your own turnover data is the only way to know whether you are better or worse than that.
Boost customer satisfaction with just a few clicks
Most-Loved Features:
- On-demand drivers
- Real-time GPS tracking
- Delivery confirmation photos
- Over 50% of customers report a smoother delivery experience
What counts as employee data in a delivery operation
Employee data is every piece of information your business holds about the people who work for it. The standard breakdown splits it into personal details, job details, pay, time and attendance, and performance. That framing is fine and every HR resource uses it, but it flattens something important: in a delivery operation these categories are not equally risky and they are not equally useful.
Sorted by how much trouble each one can cause you, the picture looks different:
- Time and attendance data carries nearly all the legal exposure. Hours worked, when shifts started and ended, breaks, overtime. This is the category that generates wage claims, and route work makes it hard to get right.
- Pay data is second, because it only makes sense alongside the hours. A rate table without the hours it was applied to proves nothing.
- Driver-specific qualification data is a category general HR advice does not have at all: licences, medical certificates, motor vehicle records, road tests. Whether you need it depends on your vehicles, covered below.
- Personal and identity data carries little operational value and most of your privacy risk. Social security numbers, bank details, addresses, emergency contacts. You need it, you must protect it, and you should not be casual about who can see it.
- Performance and route data carries no legal weight at all and is where the actual management value lives. Deliveries completed, time per stop, failed attempts, customer complaints, tenure.
The practical consequence: the data you are legally obliged to keep and the data that helps you run the business barely overlap. You need both, for different reasons, and conflating them is why so many operators end up with spotless payroll files and no idea which of their routes is losing money.
The employee records the FLSA requires you to keep
This is the list to print out. For every non-exempt employee, which in a small delivery operation means essentially all of your hourly drivers, packers and shop staff, the Department of Labor requires these 14 items:
- Full name and social security number
- Address, including zip code
- Birth date, if younger than 19
- Sex and occupation
- Time and day of week when the employee’s workweek begins
- Hours worked each day
- Total hours worked each workweek
- Basis on which the employee’s wages are paid
- Regular hourly pay rate
- Total daily or weekly straight-time earnings
- Total overtime earnings for the workweek
- All additions to or deductions from wages
- Total wages paid each pay period
- Date of payment and the pay period covered
Source: U.S. Department of Labor, Fact Sheet #21.
Read items five, six and seven together, because that is where small operations come apart. Item five asks you to have decided, in advance and in writing, when your workweek begins. Not when you pay people. When the week starts. Item six wants daily hours. Item seven wants the weekly total as its own figure.
If your workweek boundary is vague, you cannot calculate overtime correctly, because overtime is owed on hours past 40 in a defined workweek. A driver who works 30 hours Thursday through Saturday and 30 more Sunday through Wednesday has either worked two normal weeks or one 46-hour week with six hours of overtime, depending entirely on where you put the line. Operations that deliver on weekends get this wrong constantly.
There is no required format. A spreadsheet satisfies the law as long as it is accurate, complete and you can produce it. The regulation cares about content and availability, not sophistication.
What the FLSA does not require
The other direction matters just as much, because a lot of small businesses over-collect out of anxiety. The FLSA does not require you to keep performance reviews, disciplinary notes, training logs, route completion statistics, or any of the management data discussed later in this guide. It does not require a particular time-clock technology. It does not require employee signatures on timesheets, though getting them is a sensible habit.
Other rules may demand records the FLSA does not: I-9s for employment eligibility, state-specific sick leave accrual, OSHA logs above certain headcounts. But the wage and hour list above is the spine, and it is the one that gets audited.
How long to keep employee records
Two clocks, and people routinely collapse them into one. The Department of Labor sets a three-year minimum for payroll records and a two-year minimum for the underlying documents wage calculations were built from.
| Record type | Federal minimum | Practical advice |
|---|---|---|
| Payroll records, collective bargaining agreements, sales and purchase records | 3 years | Keep indefinitely; they are small files |
| Time cards, piece-work tickets, wage rate tables, work and time schedules, records of additions or deductions | 2 years | Keep 4 years to cover state law and claim windows |
| I-9 employment eligibility forms | 3 years after hire or 1 year after termination, whichever is later | Store separately from personnel files |
| Driver qualification file (if applicable) | 3 years after the driver leaves | Keep as its own file, not mixed into HR |
Two caveats that matter more than the table. First, states set their own floors and several are longer than the federal one, so the federal minimum is a floor and not a target. Second, the clock that matters in a dispute is the claim window, not the retention rule. A wage claim can reach back further than two years where a violation is alleged to be willful. Holding time records for four years rather than two costs you nothing and removes the worst version of this conversation.
The mirror-image mistake is keeping everything forever. Personal data you no longer need is pure liability, and holding a former employee’s bank details and social security number eight years after they left serves no purpose. Write down what you keep, for how long, and who deletes it.
Driver qualification files: when delivery staff need more than a timesheet
This is the piece of employee data that general HR guidance will never mention, and it catches growing operations at exactly the wrong moment.
Federal motor carrier rules attach to the vehicle, not the job title or the licence. The trigger is a gross vehicle weight rating of 10,001 pounds or more, used in commerce, crossing state lines. Above that line the Department of Transportation requires a driver qualification file for each driver — whether or not a commercial driver’s licence is involved.
The file generally holds the employment application, safety performance history from previous employers, motor vehicle records pulled on a schedule, a road test certificate or equivalent, and a valid medical examiner’s certificate.
What makes this a trap is that it has nothing to do with how big your business feels. A two-person florist that buys one used box truck and delivers over a state line has crossed into it. The same florist running three Transit Connects has not. Check the GVWR on the door jamb sticker of every vehicle you own or rent, not the weight of what you put in it. Rented and borrowed vehicles count.
If you are under the threshold and staying there, you still want a slimmer version of the same file for your own protection: a copy of each driver’s licence, a motor vehicle record checked at least annually, and a note of when you last checked. Your insurer will very likely ask for exactly this after any incident, and “we checked when we hired him in 2023” is a bad answer.
Employee data that isn’t legally required but pays for itself
Everything above is defensive. This part is where employee data starts earning money, and almost none of it is mandatory.
The delivery-specific numbers to capture per driver, per week:
- Stops completed and stops attempted. The gap is your failed-delivery rate, and it varies far more by driver than most owners expect.
- Actual route time against planned route time. Consistent overruns are either a routing problem or a training problem, and you cannot tell which without the data.
- Overtime hours by person and by week. Overtime concentrated in one or two people is a scheduling defect, not a labour cost.
- Customer complaints tied to a specific run. Complaints are usually filed against the business. Tying them to a route and a driver is what makes them useful.
- Tenure and exit reasons. Covered in the next section, and the one most operations throw away.
The reason to keep these per-person rather than in aggregate is that averages hide the thing you want. A business whose average route runs 12 minutes over is a business with one driver running 40 minutes over and four running on time. The aggregate tells you to fix your routing. The per-driver data tells you to ride along with one person on Thursday.
None of this needs to be elaborate. A weekly tab with one row per driver and six columns beats a dashboard nobody fills in.
What employee data reveals about driver turnover
Turnover is the expense most delivery operations carry without ever measuring. The Bureau of Labor Statistics puts total separations in transportation, warehousing and utilities at 4.2% in June 2026 and 3.8% in July. Those are monthly figures, which is the detail that gets misread. Sustained at that rate, a sector turns over a substantial share of its workforce every year.
You cannot benchmark yourself against that without three bits of data you may not be recording:
- Hire date and separation date for every person, kept after they leave. Purging former employees from your records destroys the only dataset that can tell you whether people leave at week three or year three.
- Reason for leaving, in your own words, written the week it happens. Not a formal exit interview. One line.
- Who trained them and on what route. Turnover often concentrates around a particular route or a particular trainer, and that only becomes visible with several data points.
The pattern this data most often exposes is early attrition. If your drivers overwhelmingly leave inside the first 90 days, you have a hiring and onboarding problem, and a retention bonus at 12 months will not touch it. If they leave at 18 months, you have a pay-progression problem, and better onboarding will not touch that. The remedies are nearly opposite, which is why guessing is expensive.
Replacement is not cheap either. Recruiting, screening, background checks, onboarding and training on your specific routes add up, and a route running short in the meantime costs deliveries. You do not need an industry estimate to price this. You need your own last three replacements, which is again a record you either kept or did not.
Where to keep employee data, and who should see it
Employee data management at your size is mostly three decisions, and they are about access rather than technology.
Decide what you actually need before you collect it. Every field you hold is a field you have to protect. Date of birth for someone over 19 is not a required record. Neither is a copy of a social security card once you have completed the I-9. Collecting less is the cheapest security control available.
Separate the sensitive from the operational. Identity and pay data needs to be locked down and seen by almost nobody. Schedules, route assignments and availability need to be seen by whoever dispatches. Keeping these in one undifferentiated folder forces you to choose between blocking your dispatcher and exposing bank details. Split them and the problem disappears.
Give access by role and review it. Whoever runs payroll sees pay. Whoever builds routes sees availability and qualifications. Nobody else sees either. The most common real-world breach in a small business is not a hacker. It is a shared login that a departed employee still has, or a personnel folder in an unlocked drawer in a room customers walk through.
Two practical additions. Keep medical information, including a driver’s medical examiner’s certificate, physically or logically separate from the general personnel file, because access rules for it are stricter. And write down who is responsible for all of this. Data governance in a nine-person business is one named person and a page of notes, but it has to be one named person.
Do you need a system to hold all of this?
Honestly, not at first. Up to roughly ten people, a locked cabinet for signed documents plus one well-built spreadsheet satisfies every requirement in this guide, and the discipline matters more than the tooling.
The point where it stops working is fairly predictable. It arrives when hours have to be collected from people who are not in the building, when the same figure is being typed twice (once into a timesheet and once into payroll), and when the person who knows how the spreadsheet works is also the person driving the Saturday route. That double entry is where payroll errors come from, and payroll errors in overtime are exactly the kind that turn into back-wage claims.
That is the transition an employee management system is built for: the records get captured once, at the source, and the hours, availability and driver qualifications feed the schedule instead of sitting in a file nobody opens until there is a dispute. Which features are worth paying for when most of your staff work off-site is a longer conversation, and that post has it.
Employee data mistakes that cost delivery businesses money
The recurring ones, in rough order of how expensive they get:
- Recording scheduled hours instead of worked hours. A driver scheduled 8 to 4 who gets back at 5:40 worked 9 hours 40 minutes. If your record says 8, your record is wrong in the direction that costs you.
- Treating a driver as exempt. Route drivers at a small business are almost never exempt from overtime. Paying a flat weekly rate does not make someone salaried in the legal sense, and this misclassification is a leading source of back-wage liability.
- Leaving the workweek boundary undefined. Covered above, and it silently corrupts every overtime calculation in a weekend-delivery operation.
- Not counting all of the working time. Loading the van, the pre-trip check, returning to base, reconciling undelivered items. If you require it, it is generally hours worked.
- Keeping route data only in aggregate. No legal cost, but it is why an underperforming route can lose money for a year without anyone noticing.
- Deleting records when someone leaves. This breaks both your retention obligation and your turnover analysis in one move.
Frequently asked questions about employee data
What employee data am I legally required to keep? For every non-exempt employee, the 14 items listed in the Department of Labor’s Fact Sheet #21: identity and address details, the workweek start, daily and weekly hours, pay basis and rate, straight-time and overtime earnings, additions and deductions, total wages, and payment dates. Other rules add to that list, but this is the core.
How long should I keep employee records? At minimum, three years for payroll records and two years for the documents wage calculations were built on, such as time cards and schedules. State law is sometimes longer. Keeping time records four years is cheap insurance against willful-violation claims that reach back further.
Do I need a driver qualification file for my delivery drivers? Only if a vehicle you use in commerce has a gross vehicle weight rating of 10,001 pounds or more and operates across state lines, and then yes, regardless of whether the driver holds a CDL. Check the GVWR sticker on every vehicle including rentals. Below the threshold, keep a licence copy and an annual motor vehicle record check anyway, because your insurer will want them.
Can I keep employee data in a spreadsheet? Yes. The law specifies no format, only that records are accurate, complete and produceable. A spreadsheet stops being adequate when hours come from people who are not on site and the same numbers get entered twice.
Who in my business should be able to see employee data? Access by role. Payroll data to whoever runs payroll, availability and qualification data to whoever dispatches, identity and bank details to as few people as possible. Keep medical records, including driver medical certificates, separate from the general personnel file.
What employee data is worth tracking beyond the legal minimum? Per driver and per week: stops completed against stops attempted, actual route time against planned, overtime hours, complaints tied to a specific run, and tenure with a one-line reason for leaving. None of it is required, and it is where the management value is.
Start with the two-week version
You do not need to fix all of this at once, and a plan that requires a quiet weekend will not happen. Two weeks of small steps gets you most of the way.
This week: write down the day and time your workweek starts, and put it where the schedule lives. Walk your vehicles and note the GVWR on each door jamb. Confirm that what you record as hours is what people actually worked, including loading and the trip back.
Next week: add one column to your existing spreadsheet for hire date and one for separation date, and backfill the last two years from whatever you have. Check who can open the folder your personnel files live in, and reduce it.
The employee data you already collect, arranged so you can find it, covers the legal requirement and answers the operational questions too. What most delivery businesses are missing is a defined workweek, hours recorded as they were worked, and records that outlive the people they describe. None of that requires a platform.
