Skip to content
Tudor Hilcu THG Conseil

Maersk and NotPetya: ten days without IT

On 27 June 2017, the NotPetya malware brought down the IT systems of Maersk, then the world’s largest container shipping company. According to its chairman, it took ten days to reinstall 4,000 servers and 45,000 PCs. Maersk put its losses at between $250 million and $300 million. Two things kept it going: a copy of its user directory that happened to be offline, and teams who carried on by hand.

Author
Tudor HilcuFounder & CEO, THG Conseil
Published
Reading time
11 min
Topics
Crisis, Continuity, IT
Contents9

On 27 June 2017, Maersk was hit by NotPetya, a piece of malware that started out as a booby-trapped update to a Ukrainian accounting package. According to Adam Banks, who led the rebuild of the group’s IT, it reached around 55,000 devices in seven minutes. The team decided to switch everything off, which, according to him, took around seven hours. Some of the group’s port terminals closed, and several days went by before the booking website could take orders again.

In January 2018, at the World Economic Forum in Davos, Maersk’s chairman, Jim Hagemann Snabe, told the rest of the story. The company had to reinstall its entire infrastructure: 4,000 servers, 45,000 PCs and 2,500 applications, in ten days. A former head of the software company SAP, he added: “Normally […] I would say it’s gonna take six months. It took ten days.”

Here is the case in full, with its sources. Maersk is a shipping giant. Yet the question it raises applies just as much to an SME with twenty staff: if your computers stop tomorrow morning, what keeps running, and who knows how to do it by hand?

The copy that saved everything was there by chance

The teams rebuilding the network, brought together in England, soon found backups of almost every server, dating from three to seven days before the attack, according to the US magazine Wired. One was missing, and it was crucial: the backup of the domain controllers. These are the servers that hold the company’s user directory: who has an account, and what they are allowed to open. Without them, nothing starts up again.

Maersk had around 150 of them, each copying the others. Every one was a backup for the rest. That set-up had not allowed for one scenario: all of them being wiped at the same time.

After calls to hundreds of administrators around the world, the teams found one intact controller, in an office in West Africa. A power cut had knocked it off the network before the attack. It held the only known copy of the directory that the malware had not touched. Wired reports whoops of joy in the office when it was found. The connection was too slow to send it over, so, according to Adam Banks, a member of staff flew it in on a hard drive. Banks later described that drive as the yeast from which the whole network was rebuilt.

Remember this: the copy that saved Maersk was never planned. It was offline by accident. France’s national cybersecurity agency, ANSSI, asks companies to make that a rule: “Always make offline backups regularly”, disconnected from the rest of the IT system.

For ten days, 80% of the business by hand

In Davos, Jim Hagemann Snabe painted the picture: a company where a ship carrying 20,000 containers enters a port every 15 minutes, and which goes ten days without IT. According to him, volume fell by only 20%: the remaining 80% was handled manually. He spoke of “human resilience”, and praised the help customers gave.

According to Wired, staff took orders through personal email accounts, WhatsApp and spreadsheets. Adam Banks, for his part, made sure from day one that he had the whole executive team’s phone numbers on paper. To stop salespeople constantly calling the technical teams, he sent out an update video every 12 hours. The CEO chose the same openness with the outside world. Andy Powell, who later became the group’s chief information security officer, said that customers appreciated being told what was happening from the very first day.

One decision is worth noting. Rather than restore systems that had just been destroyed, Adam Banks proposed rebuilding everything from scratch. When the CEO voiced his concern, Banks gave him a cut-off point: if nothing was moving after eight to ten days and a patch or antivirus had come out, they would go back to restoring. A quick decision, with a date to review it.

Ten days is the chairman’s figure for the essentials. The rest took longer. According to Adam Banks, the core business was running again by around day 14, at reduced volume. It took another four weeks to get back to full scale: the time needed to source 17,000 new end-user devices. He says he did not go home for 70 days.

What the crisis cost, and where

Maersk published the cost. On 16 August 2017, its CEO, Søren Skou, said the attack had mainly hit container shipping, the port terminals and logistics, and that volumes had fallen for a couple of weeks in July. He expected an impact of $200 million to $300 million on results. In November, the group put it at between $250 million and $300 million, adding that the vast majority came from business temporarily lost in July and August. The 2017 annual report confirmed the figure, which covers, among other things, lost revenue, IT restoration costs and extraordinary operating costs.

One detail will ring true for any SME. At Damco, the group’s logistics arm, operating cash flow ended the year at minus $101 million, against plus $4 million the year before. The annual report puts this down in part to the temporary effect of the attack on operations and invoicing. An invoice that does not go out is a payment that does not come in.

After the crisis, Maersk strengthened its IT service continuity and its business continuity plans, and bought cyber insurance. In 2019, Andy Powell said everything was built with the aim of restoring service within 24 hours.

The same question, at SME scale

An SME does not have 4,000 servers. Yet it too has its own directory: administrator access, email, online banking, the invoicing and payroll software, the licences. And sometimes only one person knows where all of that is.

In Davos, Maersk’s chairman put it this way: the group had been “basically average” on cybersecurity, “like many companies”. So this is not a story about badly prepared giants. The case comes down to three lessons, which hold at any size:

  • the backup that counts is the one that stays offline, and holds the keys to get everything running again;
  • the business keeps going if someone knows how to do it by hand, with the paperwork ready;
  • most of what a crisis costs comes from lost business and invoices that do not go out. Fixing the IT is only part of the bill.

Five decisions to take on a quiet day

  1. Write down what has to keep going if the IT stops. Taking orders, delivering, invoicing, collecting payments, paying salaries. For each one, note who knows how to do it by hand, and get the paperwork ready: numbered order forms, printed lists of customers and suppliers.
  2. Keep an offline copy, and put the keys in it. The data, but also administrator access, software configurations, licences and installation files. ANSSI reminds you not to forget “configurations and setup media of your applications”. It recommends the so-called 3-2-1 rule: your live data plus two backups on two different media, one of them offline.
  3. Test a restore, for real. At least once a year, have a workstation or a server brought back up from the backup, and time it. ANSSI calls for regular restore tests, and for management to be involved in deciding what reduced level of service is acceptable during a cyber crisis.
  4. Write the recovery time into your IT provider’s contract. The recovery time, how often backups and tests are run, the report you receive after each test, and what happens if any of it is not met. A deadline that is not written down anywhere commits nobody.
  5. Decide in advance who talks to customers, and how often. Maersk gave an update every 12 hours. A fixed slot stops everyone calling everyone. The detail is in the article on the first 48 hours of a crisis.
Fictitious example: the “if the IT stops” page of a 30-person wholesale SME
What has to keep going Who knows how to do it by hand What is ready
Taking orders Léa and Marc, by phone Numbered paper order forms, a printed list of the top 50 customers with their prices
Shipping Paul, warehouse manager Blank delivery notes, a list of carriers and their numbers
Invoicing and collecting Sophie, accounts A paper log of deliveries to invoice as soon as the software is back, a list of customers who pay by transfer
Paying salaries The managing director, with the accountant The latest payroll summary, printed every month

At the bottom of the page: “Last restore test: 12 March, 3 hours 40 minutes to get the server running again. Next test: September. Offline copy: drive kept in the safe, updated every Friday.”

The calculation that goes with it, in round figures. The SME invoices about €40,000 per working day. One week without its software, with 60% of the business kept going by hand: 5 days × €40,000 × 40% = €80,000 of revenue lost or pushed back. And the sales kept going by hand, 5 × €40,000 × 60% = €120,000, are invoiced late: that much cash comes in late. Both figures hit cash: over the following weeks, up to €200,000 of receipts are missing or late, while salaries and overheads fall due on time. That total is what the cash flow plan has to be able to absorb.

What I learnt

On my assignments for large groups, I have led critical go-lives, with a rollback plan ready before go-live day, and tracked remediation plans in live operation. What saves hours is decided beforehand: who decides, what you know how to bring back up, and how long it takes.

On every assignment, I also handled the supplier contracts, with service levels written down in black and white, so that penalties could be claimed or a contract exited. My sports club, closed overnight by the government in 2020, taught me the rest: when business stops, cash and the relationship with customers decide what happens next.

The limits of this case

The cost and its components come from Maersk’s own publications. The rebuild figures are the ones Jim Hagemann Snabe gave in Davos. The account of the backups comes from Wired, which relies on Maersk staff who were mostly anonymous, and from the talk Adam Banks and Andy Powell gave in 2019.

The sources differ on three points. The office holding the intact copy was in Ghana according to Wired, and in Lagos, Nigeria, according to Adam Banks: I write “an office in West Africa”. Wired also describes a relay of two staff members to bring the drive to London, where Adam Banks talks about a single trip. The duration depends on what you count: ten days for the essentials according to the chairman, four more weeks to reach full scale according to Adam Banks, close to two months of work for some teams according to Wired. Finally, in 2019, Adam Banks gave a total cost of $350 million, including around $30 million for the rebuild; I use the figure Maersk published.

I do not deal with the technical side of IT security: those choices are made with your IT provider. For the wording of contracts, insurance and the obligations that apply to you, go through them with your lawyer and your accountant. The ANSSI recommendations quoted here were checked on 24 October 2026.

Where to start this week

  • Ask your IT provider, or whoever looks after your IT: when did we last restore a backup, and how long did it take?
  • Check that at least one copy is disconnected from the network, and that the emergency access details are not stored only in the IT they are meant to rescue.
  • Write down on one page what has to keep going if the IT stops, and who knows how to do it by hand.

Half an hour, and you know whether your backup is a certainty or an assumption.

A backup you have never restored is only an assumption.

To write down who decides, who speaks and who does what on the day everything stops, read the first 48 hours of a crisis, on one page. The Crisis on one page module adds an Excel workbook to spot your critical suppliers, watch ten warning signs and keep a log of decisions. When invoices stop going out, the 13-week cash flow plan tells you how many weeks you can hold out. All the documents are free to download on the Resources page, and the rapid assessment takes two minutes, with nothing sent. Coming through a crisis explains how I work. The first conversation is free, by phone and in strict confidence, within 48 hours: call me or write to me, on +33 6 74 86 88 70 or at contact@thgconseil.fr. In a crisis, call directly: the conversation and the intervention contract happen within 2 hours.

Sources

Facts checked on 24 October 2026 against the original sources.

Is something holding you back?

Call me or write to me. The first conversation is complimentary, and I reply within 48 hours.

Take the rapid assessment

Four questions, two minutes. Nothing is sent or stored.

By phone:+33 6 74 86 88 70WhatsAppEmail