How to know if your employees are stealing from your gas station.

The hard truth

At most independent gas stations and c-stores, somebody on staff is taking small amounts of money or product occasionally, and often it's not the person you suspect. Industry data: employee theft is about 28.5–30% of total retail shrinkage. The National Retail Federation estimates internal theft costs US retailers $26 billion annually.

At a single c-store, employee theft typically runs $5,000–$50,000 a year. Below $5,000, you usually don't notice. Above $50,000, you usually do, and by then the dollar exposure is meaningful even before the time you spend investigating.

30% of shrink

Comes from employee theft, not customers

National Retail Federation data. Most independents underestimate employee theft because customer theft is more visible. Internal theft is quieter, slower, and bigger.

Why this is hard to detect

Employee theft at gas stations isn't usually dramatic. Nobody's walking out with thousands of dollars at once. It's small amounts repeated over many shifts: $15 here, two packs of cigarettes there, a $3 scratcher pocketed at the end of a slow Sunday.

The math is brutal. $20 per shift × 4 shifts per week × 50 weeks per year = $4,000 per year from a single clerk. Two clerks at that rate = $8,000. Three = $12,000. Spread across multiple categories and shifts, the annual cost compounds quickly without any single event being noticeable.

Five common employee-theft patterns

1. Sweethearting

Clerk rings up some items for a friend, family member, or just-a-friendly-customer at a steep discount or for free. The customer pays a small amount; the clerk pockets some of the difference or just lets the goods go for less than face value.

How it shows up in data: per-clerk average transaction value lower than peer average. Specific SKUs (often high-margin tobacco) move at higher quantities during that clerk's shifts. Receipt history shows unusually round-number totals.

2. Void abuse

The classic gas-station theft pattern. Clerk rings up a sale, customer pays cash, clerk hands them the product. After the customer leaves, the clerk voids the transaction and pockets the cash. POS shows fewer sales than actually occurred; cash drawer comes up balanced because the cash matched the (voided) sales.

How it shows up in data: void rate per clerk vs. baseline. Clerk-baseline is usually 1–2 voids per shift. A clerk running 5+ voids per shift, especially in patterns (voids clustered in 15-minute windows), is a strong signal.

3. Drive-off falsification

Clerk reports a drive-off (customer drove away without paying for fuel) when actually a paying customer paid cash and the clerk pocketed the cash. The drive-off log absorbs the loss; the owner doesn't notice because drive-offs happen sometimes.

How it shows up in data: drive-off frequency per shift vs. baseline. Plate numbers on drive-off reports that don't check out against state DMV records. Drive-offs that happened during low-traffic windows when manager attention was elsewhere.

4. Cash skim at handoff

Clerk closes their shift, counts the drawer correctly (or appears to), but removes cash before the next clerk takes over. The variance shows up on the next clerk's opening count or at end-of-shift two shifts later, making attribution unclear.

How it shows up in data: cash variance that appears consistently at the same shift transition (e.g., morning-to-afternoon handoff on Tuesdays). Drawer counts that don't match the prior shift's close.

5. Lottery pack manipulation

Clerk pockets scratcher tickets, sometimes winning tickets that customers throw out at the counter (and the clerk later cashes), sometimes individual tickets from active packs at the end of a slow shift. The missing serials show up later in pack settlement.

How it shows up in data: lottery serial gaps that cluster around specific shifts. A clerk who happens to be on shift for a disproportionate share of missing serial events. Lottery cash-out events near shift boundaries.

The three data sources you need to cross-check

No single report tells you what's happening. The signal lives in the overlap between three:

Cash variance by shift and clerk.Daily variance amounts attributed to who closed. Patterns over 30+ days surface the recurring shorts.

Void / refund / drive-off logs by clerk.Per-clerk void rate vs. that clerk's own baseline. Per-clerk refund rate vs. peer average. Drive-off frequency by shift.

Inventory variance by SKU and shift.Tobacco velocity per shift. Specific high-value SKUs dropping more than expected. Lottery serial gaps with shift attribution.

A pattern in one data source is a signal. A pattern across all three on the same shift and clerk is a confirmation.

Worked example: a confirmed pattern

30-day review at an independent c-store. Overnight clerk M.

Cash variance:
  Sun 11 PM–7 AM:  $-18 avg variance
  Tue 11 PM–7 AM:  $-22 avg variance  ← always over warning band
  Thu 11 PM–7 AM:  $-15 avg variance
  Compared to baseline (other overnight clerks): $-3 to $-5 avg

Void / refund logs:
  Clerk M overnight: avg 5.8 voids/shift
  Other overnights:  avg 1.2 voids/shift
  Refund rate: 2.1× peer average

Inventory by SKU:
  Marlboro Box velocity Tue overnight: 12% lower than other Tues overnight
  Newport Box velocity Thu overnight:  9% lower than other Thu overnight
  Pattern matches the cash-variance shifts

Cross-signal:
  Same clerk
  Same shift (overnight)
  Same days of week
  Three data sources all flagging in the same direction

Verdict:
  Strong pattern. Confirmed by camera review of three flagged
  evenings, clerk seen voiding transactions after customer
  leaves; tobacco pulls from display not appearing in POS.

Total dollar exposure across 30 days: ~$2,400
Annualized if unchanged: ~$29,000

What to do when you confirm a pattern

Document everything before any conversation. Camera footage of specific flagged events. Audit row exports. Variance history. The conversation goes differently when you have documentation than when you have suspicion.
Talk to an HR attorney before action.Termination for theft has legal exposure. Wage-and- hour rules, accusation defamation risk, and state-specific employment law all apply. A 30-minute consultation usually costs $200–$400 and saves considerably more than that.
Consider the criminal-vs-civil question.Some owners file police reports; some don't. The choice depends on dollar exposure, your willingness to testify, and local prosecutor policies. Many states won't prosecute under $1,000 in losses.
Decide what you want. Restitution? Termination? Prosecution? Quiet separation? Each choice changes how you handle the conversation and what you ask the attorney to draft.
Conduct the conversation professionally.Manager + owner together (witnesses matter). Specific evidence presented (no "I think"). Calm tone. Documented outcome.

How to make theft harder in the first place

Structured shift handoffs. Both clerks sign the drawer count together. Single- clerk drawer counts after handoff are where attribution dies.
Real-time variance alerts.Over-tolerance variance sends an SMS to the owner at submit. Clerks change behavior when they know the owner sees variance the moment it happens.
Per-shift cycle counts on tobacco.High-velocity tobacco SKUs counted every 1–2 days rather than monthly. Velocity drops surface quickly with shift attribution.
Camera coverage with clear angles.Register cameras with the cash drawer visible. Forecourt cameras with plate-number capture. Loss attribution requires evidence; cameras provide it.
Audit-trail integrity. Voids and refunds require a manager approval or a written reason code. Audit row preserves who, when, what variance, what reason, for every event.
Hire well. See how to hire a good gas station manager. Most theft prevention happens in the interview, not after hire.

Common mistakes when investigating

Acting on a single event.One bad night isn't a pattern. Document the event, watch for repetition, build the case across 30+ days before action.
Confronting before documenting.Once you talk to the employee, the pattern stops (which is good for losses but bad for getting documentation). Document first, talk second.
Skipping the attorney call.Wage-and-hour, accusation, and termination procedures all carry legal risk. $200–$400 for attorney advice usually saves more than it costs.
Trusting your gut over the data.The employee you suspect is often not the one taking. The employee you trust most has the most opportunity. Trust the data.
Trying to handle it through HR drama rather than process. Discipline, written warnings, and separations need clean documentation. Drama produces lawsuits.

Frequently asked questions

How do I know if my gas station employees are stealing?

You need to cross-check three data sources together: cash variance by shift and clerk, void/refund/drive-off logs, and inventory variance by SKU. A pattern in one alone is a signal; a pattern across all three on the same shift and clerk is confirmation. Single events aren't signal, recurring patterns over 30+ days are.

What's the most common employee theft at a gas station?

Void abuse is the classic pattern: clerk rings up a sale, customer pays cash, clerk voids after the customer leaves and pockets the cash. The drawer balances because cash matched the (voided) sales. Detection requires monitoring per-clerk void rates against their own baseline, most clerks run 1–2 voids per shift; theft patterns show 5+ in clustered windows.

How much do employees typically steal from a gas station?

At a single c-store, employee theft typically runs $5,000–$50,000 per year, usually small amounts repeated over many shifts rather than dramatic single events. $20 per shift × 4 shifts a week × 50 weeks = $4,000 from one clerk; two or three clerks doing similar compounds quickly across the year.

What should I do if I think an employee is stealing?

Document the pattern across 30+ days before any conversation. Pull camera footage of specific flagged events. Talk to an HR attorney before taking action ($200–$400 typically saves much more than it costs). Decide what outcome you want (restitution, termination, prosecution). Conduct the conversation professionally with documentation in hand.

Should I call the police?

Depends on dollar exposure, your willingness to testify, and local prosecutor policies. Many states won't prosecute under $1,000 in losses. For substantial losses ($5,000+) with clear video evidence, filing a report is usually worth it. For smaller amounts, quiet separation with restitution is often the practical outcome.

Can I just fire an employee I suspect of theft?

Legally risky without documentation. Wrongful termination claims, defamation, and labor-board complaints all become possible if the documentation is weak. Termination "for theft" requires evidence; "no longer a good fit" is the safer phrasing when documentation is thin. Always talk to an HR attorney first.

How can I make theft harder to commit?

Six things: structured shift handoffs with both clerks signing the drawer count, real-time over-tolerance variance alerts to the owner, per-shift cycle counts on tobacco, camera coverage with clear register angles, audit-trail integrity (voids require approval or reason codes), and disciplined hiring with reference checks. Most prevention happens before hire and at process level, not at the conversation.