Breachvaultops • Updated November 2026
📝 Pricing & Premiums • 9 min read

Seven Mistakes Companies Make When Choosing a Cyber Limit, and How to Size It Instead

Peer benchmarks and revenue rules of thumb produce cyber limits that look sensible in a board pack and fall short at claim time. Here's how a CFO can size a limit from realistic loss scenarios in one afternoon.

SPONSORED CONTEXTUAL PLACEMENT
💼
Evaluating Enterprise Deployments in 2026?

Request our confidential total cost of ownership matrix and vendor RFP templates.

Launch Calculator →

Borrowed Benchmarks: Copying Peers and Using a Revenue Percentage

Ask a broker "how much cyber insurance do I need" and you'll usually get one of two answers: what similar companies buy, or a number tied to revenue. Both are fast and easy to defend to a board. Both tend to produce a limit that has little to do with what an incident at your company would cost.

Mistake 1: Copying what peers buy

Benchmarking data shows what companies bought. It doesn't show what they lost or needed. Peers set limits around budget, available insurer capacity that year, and whatever their last broker suggested. When the market hardened around 2021 and 2022, many buyers cut limits to hold premiums flat, and those decisions still sit in benchmark sets.

Advertisement

Your peer group also doesn't share your exposure. Take two manufacturers that each earn $80M. One runs a single plant that depends on networked controls. The other has five sites with manual fallback. Their worst-case outages are nowhere near each other, even though revenue matches.

Do this instead: Use benchmarking as a sanity check at the end, not the starting point. If your scenario-based number is well below the peer median, find out why. If it's above, that alone isn't a reason to cut it.

Mistake 2: Applying a revenue percentage

Rules like "$1M of limit per $10M of revenue" are popular because they need no analysis. But revenue doesn't drive cyber loss. These four things do:

  • Daily gross profit at risk when systems go down, and how fast you can work manually.
  • Records held: how many, how sensitive, and which countries the data subjects live in.
  • Payment flows that fraud can redirect.
  • Contractual liability to customers whose data or operations depend on you.

A $40M payments processor holding two million cardholder records can face a bigger realistic loss than a $400M construction contractor. Size the limit on these four drivers, not the top line.

RECOMMENDED SOLUTION

Looking to streamline vendor RFP selection?

Access pre-negotiated tier discounts through our independent institutional network.

Compare Pricing →

Contract Minimums and Sublimits That Hollow Out the Headline Number

Mistake 3: Buying exactly the contract minimum

Customer agreements often require $1M, $2M or $5M of cyber cover. The customer's procurement team wrote those figures to protect the customer, not to match your losses. And the limit is shared: if that customer sues you, the claim draws down the same aggregate that must pay for your forensics, restoration and lost income.

Do this instead: Treat the largest contractual requirement as a floor. Pull every master services agreement and data processing agreement. Note the required limit, whether the customer must be named as an additional insured, and any primary and non-contributory wording. Then size from scenarios and confirm the result clears the floor.

Mistake 4: Ignoring sublimits

The gap between the aggregate and its sublimits is where many claims go wrong. A $5M policy may cap social engineering fraud at $250,000 and dependent business interruption at $1M. Those caps apply however much aggregate remains. The headline number is the most the policy pays in total; each sublimit is the most it pays for one type of loss.

Commonly sublimited coverTypical mid-market sublimit (indicative, varies by insurer)What to check
Social engineering / funds transfer fraud$100,000 to $500,000Whether callback verification is a condition of cover
Dependent (contingent) business interruptionOften a fraction of the aggregate, or a fixed amountWhether it covers only named IT providers or all suppliers
Regulatory fines and penaltiesAnywhere from the full limit to a fixed cap, "where insurable by law"Which jurisdiction's law decides insurability
Payment card (PCI) fines and assessmentsCommonly $100,000 to $1MWhether card brand recovery charges are included
Hardware replacement ("bricking")$100,000 to $500,000Whether betterment (upgraded replacements) is excluded

Do this instead: Line up each sublimit against the matching line in your loss scenarios, then negotiate the sublimits your scenarios actually reach. Raising the aggregate won't fix those gaps.

Quieter Errors: Eroding Limits, Single-Event Thinking and One Policy for Many Countries

Mistake 5: Forgetting what sits inside the limit

Many cyber policies pay defense costs from within the limit. Every dollar spent defending a regulatory investigation or class action reduces what's left for settlements and other costs. The retention bites too: a $250,000 retention on a $2M policy means you carry the first quarter-million yourself. Business interruption cover usually adds a waiting period, often a number of hours, before income loss counts.

Do this instead: Add a defense cost line to each scenario. When you compare scenario totals to the limit, take the retention out of the total first.

Mistake 6: Sizing for one event when the limit is annual

Cyber limits are usually annual aggregates. Two unrelated incidents in one policy year, say a funds transfer fraud in March and ransomware in October, draw from the same pot. Reinstatement of the limit is rarely offered, and when it is, it's expensive.

Do this instead: Check that your worst-case scenario plus a moderate second event both fit within the limit. If they don't, decide openly whether to accept the gap or pay to close it.

Mistake 7: One global limit for a multinational footprint

Regulators differ by country, so a single number set at headquarters can fall short elsewhere. Notification duties and enforcement styles vary across the EU and UK GDPR, Brazil's LGPD, India's DPDP Act, Singapore's PDPA, Australia's notifiable data breaches scheme and US state laws. Whether fines can be insured varies even more. In the UK and many EU member states, regulatory fines are generally treated as uninsurable; elsewhere the answer depends on the type of penalty. Some countries also restrict cover from insurers that aren't admitted locally, so a global policy may not respond there the way you expect.

Do this instead: Map where your operations, data subjects and revenue sit. Ask your broker whether you need locally admitted policies under a master program, and how the limit is allocated across them. Check fines coverage with local counsel in each major jurisdiction.

Running Cyber Loss Scenario Modeling in an Afternoon

Scenario modeling doesn't need actuarial software. You need your finance lead, your head of IT or security, and three to four hours. The method:

  1. Pick four scenarios: ransomware with an operational outage, a data breach, funds transfer fraud, and an outage at a critical cloud or software provider.
  2. Set severity at "severe but plausible." Use your real recovery capability, not your target. If your last restore test took six days, don't model two.
  3. Price each cost line using internal figures (daily gross profit, record counts, payment run sizes) and broker or panel vendor estimates for forensics, legal and notification.
  4. Map each line to a policy section and its sublimit.
  5. Subtract the retention and add a defense cost line.

Here's a hypothetical example. Say a 250-person logistics company earns $120M a year at a 25% gross margin, with operations in the UK, the Netherlands and the UAE. Daily gross profit is roughly $82,000. IT estimates 10 days of full outage after ransomware, then 20 days at about 40% degraded.

ScenarioMain cost linesEstimated range (hypothetical)
Ransomware with outageLost gross profit ~$1.5M; forensics and response $250k to $500k; restoration $300k to $600k; legal $100k to $200k; extortion, if legally payable and paid, $0 to $2M$2.1M to $4.8M
Data breach, 400,000 recordsNotification and call center at an assumed $2 to $5 per record; forensics; regulatory defense in two jurisdictions; customer claims$1.5M to $3.5M
Funds transfer fraudOne redirected supplier payment run$400k to $900k
Cloud provider outage, 3 daysLost gross profit after the waiting period; extra expense$250k to $600k

The ransomware scenario sets the ceiling at about $4.8M. Add a defense buffer and the company needs roughly $5M for its worst case, before any allowance for a second event. The fraud scenario exposes a separate gap: a typical $250,000 social engineering sublimit would leave most of that loss uninsured.

Turning Scenario Numbers Into a Limit, a Tower and a Budget

Start with the largest severe scenario total. Subtract the retention, then add 15% to 25% for defense costs and a second-event cushion. Round up to the nearest increment the market sells. In the example, that points to a $5M limit at minimum and a case for $7.5M. Price the extra layer if the company's largest customer contract carries uncapped data liability.

Above a few million dollars, limits are usually bought as a tower: a primary policy plus excess layers that follow its wording. Excess layers generally cost less per million than the primary, so adding the last $2.5M can be cheaper than buyers expect.

For indicative pricing, these are typical 2026 market ranges. They vary widely by industry, controls, region and contract terms:

  • Small companies under about $10M revenue: a $1M limit often runs $1,500 to $7,500 a year.
  • Mid-market companies with $50M to $250M revenue: a $5M primary limit commonly runs $30,000 to $120,000.
  • Excess layers of $2.5M to $5M added to that program: often $10,000 to $50,000 each.

Pricing in Europe, the Middle East, Asia-Pacific and Latin America can differ a lot, because insurer capacity and local admitted requirements vary. Confirm with your broker and insurers before you set a budget.

Strong controls lower the price of a given limit. Phishing-resistant MFA, tested offline backups, endpoint detection with 24/7 monitoring and payment callback procedures all help. If budget forces a trade-off, raise the retention before you cut the limit, provided your balance sheet can absorb the larger first loss.

Decision rule: If your current limit is below your largest scenario total net of retention, or any sublimit falls short of its matching scenario line, take the scenario table to your broker before renewal. Ask for quotes at two limit levels and two retentions. Have a licensed broker and counsel in your main jurisdictions review the final structure before you bind.

Common Questions

How much cyber insurance do I need if my company is small? +
What's the difference between an aggregate limit and a sublimit? +
Should we raise our retention to afford a higher limit? +
Are regulatory fines covered by cyber insurance? +
BR
Breachvaultops Editorial Team
Independent research desk
Guides on Breachvaultops are produced by our editorial team. Pricing, features and regulations change often - confirm details with the vendor or a licensed professional before making decisions.
SPONSORED 2026 Vendor Pricing Benchmarks: Model exact seat pricing, setup fees, and ROI across our verified database.
Launch Calculator →