OneNym.idDiscuss deployment
English

THE BOARDROOM FIREWALL / Chapter 7

Spending Wisely — Budget and the Return on Security

Security spending is easy to see and its benefit is mostly invisible, because the best outcome is that nothing happens. This chapter gives leaders a way to judge whether the money is aimed at the right risks and whether anyone can show it made the company safer.

This page summarises the chapter. The full text is in the book.

There is no correct percentage

The first question leaders ask is how much to spend, and the honest answer is that no magic percentage exists. Surveys report figures from a few percent to around ten percent of the IT budget, but they average companies of very different size, sector, regulation and risk appetite. A benchmark can sense-check a budget; it cannot produce one.

The principle is to spend in proportion to what you could lose and how likely you are to lose it. The book names two ways to get this wrong: spending too little, which is obvious only after an incident, and spending without direction, where generous budgets go on overlapping or unused tools. In the book's example, the board chair of the fictional Meridian Foods sets aside a benchmark slide and asks the security head for her three biggest worries. The weakest turns out to be ransomware recovery: backups exist, but no full restore of the production systems has been tested in over two years, and fixing that costs far less than the benchmark gap implied.

Budgeting from risks, not products

Risk-based budgeting starts from the short list of things that would truly hurt the business, such as ransomware stopping operations, payment fraud, stolen customer data, a compromised supplier or a misconfigured cloud service. For each, the team describes today's exposure and what reducing it would take, the cost is compared with the reduction in expected loss, and leaders choose. The total budget then follows from those decisions instead of being fixed in advance.

Estimates will be rough, and that is acceptable. Accepting a risk and spending nothing can also be the right call, provided it is a deliberate, recorded decision rather than an accident. A practical tip is to ask for three options per risk: a minimum, a recommended level that brings the risk within appetite, and an enhanced level. For technical readers, the chapter outlines quantification in the style of the FAIR model, and warns against single-point figures that suggest false precision and against modelling controls as if their benefits simply add up.

Where the first money should go

The first money belongs on the basics, which are less glamorous than what vendors want to sell. The book names five: multi-factor authentication, backups that have actually been restored from, prompt patching of known weaknesses, practical and regular training, and logging and monitoring so that intrusions are noticed. None is cutting-edge, and all address the ways attacks most often succeed. If a budget funds advanced tools while any of these remain weak, it is fair to ask why.

Prevention against the cost of a breach

Every security budget is a comparison between the cost of prevention, including the friction new controls add to daily work, and the cost of an incident. Breach costs rarely appear on one invoice: they include response specialists, revenue lost during downtime, fines and notification costs, customers who drift away, and the attention of the leadership team.

The book works through an illustrative case. If ransomware stopped Meridian's ordering and production for five working days at around 0.5 million of revenue a day, lost sales alone would be about 2.5 million, before overtime, spoiled stock and retailer penalties. Separating and regularly rehearsing backups would cost a few hundred thousand. The numbers are invented, but the method is real: build the breach figure from your own revenue, operations and contracts, because generic figures can be inflated by vendors or shrunk by stretched finance teams.

Reading requests and measuring a quiet year

Leaders cannot judge the technology in a spending request, but they can judge the reasoning. Ask which risk it addresses, what will be measurably different in six or twelve months, what the full cost is including the people to run it, and whether something the company already owns could do the job. Guard against shelfware, software bought but barely used, by asking for a dated deployment plan, short usage reports at three and six months, and evidence of value before renewal.

When success looks like nothing happening, measure readiness instead: coverage of multi-factor authentication, time to apply critical updates, the date and duration of the last restore, how quickly suspicious activity is spotted, and the results of exercises and penetration tests over time. In the book, Meridian answers a director who suspects overspending with evidence: every account now uses a second factor, critical updates take under ten days instead of six weeks, and a full restore that first took four days and failed in three places now takes thirty hours.

KEY POINTS

Key points

  • Use industry benchmarks only to sense-check a budget built from your own top risks.
  • Ask for three options per major risk: minimum, recommended and enhanced.
  • Confirm the five basics are in place and tested before funding advanced tools.
  • Build breach-cost estimates from your own revenue, operations and contracts.
  • For every request, ask about the full three-year cost, including the people needed to run it.
  • Write down the risks you choose to accept by not funding them.

QUESTIONS

Common questions

How much should a company spend on cybersecurity?
There is no correct percentage. Industry surveys show where peers sit, but they average companies with very different risks. The budget should follow from the handful of risks that would truly hurt your business and what it costs to bring each within the level of risk the board is willing to accept.
What is risk-based security budgeting?
Risk-based budgeting starts with the risks rather than the products. For each major risk, the team estimates current exposure and what reducing it would cost, and leaders compare that cost with the reduction in expected loss before choosing. Some risks may be deliberately accepted with no spending, as long as that choice is written down.
What should we spend on first in cybersecurity?
Spend first on the basics: multi-factor authentication, backups that have been tested by actually restoring them, prompt patching, regular practical training, and logging and monitoring. These measures are inexpensive compared with advanced tools and address the ways most attacks actually succeed. Advanced protections make sense only once the basics are in place.
How do you measure the return on security spending when nothing happens?
Measure readiness rather than waiting for incidents. Useful measures include how many accounts use multi-factor authentication, how quickly critical updates are applied, when systems were last restored from backup and how long it took, and how results of exercises and penetration tests change year on year. These show the effect of spending before an attack tests it.
What is shelfware and how do we avoid it?
Shelfware is software that has been bought but is not used, or is only partly used. It is common in security because products are sold on what they could do and are sometimes bought in a hurry after an incident. Avoid it by requiring a dated deployment plan, usage reports a few months after purchase, and evidence of value before any renewal.