The sentence that ought to have been reassuring is the one that sent me back to our own schema. Landmark Group said card details were not compromised because they are not stored on its system. Good. Correct, even. And completely beside the point, which is what the last few days of coverage have been slowly establishing.
On Wednesday 11 November 2026, Landmark Group disclosed a cyber incident involving unauthorised access to Shukran, the loyalty programme behind a large part of its retail footprint in this region. The company said names, contact details and dates of birth may have been accessed, that debit and credit card details were not compromised, that passwords were not affected and that Shukran points were untouched. It said it had notified the relevant statutory and law enforcement authorities and had sent SMS messages to affected customers, warning them to be wary of unverified calls, emails, SMS or WhatsApp messages referring to Landmark or Shukran. It has not said how many members were affected.
Then on 15 November, The National published an analysis under the headline No, thank you: Shukran hack a reminder financial details not essential for fraud. That piece is the reason this article exists, because it quietly dismantles the thing every disclosure template leans on.
What Landmark Said, and What the Reassurance Hides
Read the statement as an engineer rather than as a customer and it splits into two halves. The first half is a genuine architectural achievement: card numbers are not in the system, so they cannot leak from it. That is tokenisation working exactly as designed, and the team that built it should be allowed to feel good about it.
The second half is the inventory of what was there. Names. Contact details. Dates of birth. Nobody tokenised those, because nobody ever thinks of them as sensitive. They are the fields you collect on a signup form without a second thought, and they are the fields that survive every migration because deleting them feels like losing something.
Our expert view, 1 of 3
A breach notice that lists what was not taken is telling you what the organisation considers valuable. Attackers do not share that definition. Mohamed Emam, a security researcher at Group-IB quoted in The National’s analysis, states it plainly: a data breach does not need to involve financial information to create opportunities for financial fraud. The practical test for your own product is simple and slightly brutal. If someone phoned your customer tomorrow knowing only the fields you store outside your payment vault, would that call be convincing? For most loyalty and retail systems in this market, the honest answer is yes.
Why a Date of Birth Outlives a Card Number
A compromised card is a solved problem with a known cost. It gets cancelled, reissued, and the loss is absorbed by an institution with a process for exactly this. The customer is inconvenienced for a week.
A date of birth cannot be reissued. Neither can a name, and a mobile number in this country is tied to an identity in ways that make changing it genuinely disruptive. These fields are permanent, and that permanence is what converts a one time exposure into a standing liability. Maher Yamout, lead security researcher at Kaspersky, made the compounding point in the same piece: stolen data can be combined with information from earlier breaches to build progressively more detailed profiles of individuals. Your incident is not evaluated alone. It is merged with everyone else’s.
The same analysis cites the data privacy firm Cloaked on how long that merged profile stays in circulation, and the timeline is the part I would put in front of any executive who wants to know when this is over.
Our expert view, 2 of 3
Most incident plans are written on a containment clock and measured in days. The lifecycle above runs on a different clock entirely, measured in years, and almost nobody staffs for the second one. Danny Jenkins, chief executive and co-founder of ThreatLocker, framed it in The National as containment being essential without meaning the risk to affected customers has gone away. For an employer, that gap has a concrete cost: your support and customer success teams will be the target surface for pretexting long after engineering has closed the ticket. If you are budgeting a response, budget the fraud awareness training for the people who answer the phone in month nine, not just the forensics in week one.
The 4 Changes We Made to What We Store
I spent two days on our own tables after reading the 15 November piece, and the uncomfortable discovery was not a vulnerability. It was an inventory. Here is what changed.
- Every personal field now has a named feature that breaks without it. We wrote one line per field. Date of birth failed immediately: it was collected for a birthday voucher that was switched off in a campaign rewrite and never switched back on. We had been protecting a permanent identity attribute to power nothing at all. It is gone from the write path and scheduled for deletion from the store.
- Precision was reduced where the field survived. Where a date of birth genuinely does drive something, an age bracket or a birth month almost always serves the same product purpose. A month is a far weaker impersonation tool than a full date, and the feature is identical. This is the cheapest change on the list and the one that most teams skip because it feels like a rounding error.
- Retention became a column, not a policy document. A field with no expiry is a promise to protect it forever. Every personal field now carries an explicit retention period in the schema, and there is a job that enforces it. The policy PDF had said this for three years. The database had never once agreed.
- The identity store was separated from the analytics store. Reporting queries had reach into the same tables that hold the fields that make impersonation credible, which meant the exposure surface was every system that had ever been pointed at the warehouse. Analytics now reads a projection with identity fields stripped. This was the only expensive change of the four, and the only one I would defend in front of a board.
None of this is novel. It is a reminder, and the UAE already supplies the regulatory half of the argument through its personal data protection framework, which we summarise in our guide to data protection in the UAE. The regulation tells you to minimise. The breach tells you what minimising is worth.
Not sure which of the four you need first
Tell us what your product stores and who it is stored for, and we will tell you which change moves your exposure most, and what the engineering role looks like if you need to hire for it.
Let us discuss itWhat This Changes in an Engineering Requisition
The reflex after any breach headline is to open a security engineer requisition. Sometimes that is right. After this particular disclosure, I think it is usually the second hire rather than the first, and the distinction is worth an argument.
A security engineer reduces the duration of an incident. They detect sooner, contain faster, harden the path the attacker used. That is real value and it is measurable. But the Shukran disclosure does not describe a detection failure that anyone outside Landmark can assess. It describes an exposure whose severity was fixed long before the attacker arrived, by whoever decided which columns the loyalty table would carry and for how long. That decision is data modelling. It happens in the first sprint of a product, it is never revisited, and it sets the blast radius permanently.
So the first hire, for most retail and loyalty products in this region, is the engineer who can redesign what is held and build deletion that actually executes. That is a privacy engineering skill set, and it interviews very differently from a security one: you are testing schema judgement and an instinct for what a product can do without, rather than threat modelling. We set out how we scope and test for that in hiring a data privacy engineer for UAE PDPL. The security hire follows, and when you do get to it, hiring a cybersecurity engineer in Dubai covers the screen.
Our expert view, 3 of 3
Here is the interview question that came out of these two days, and it has already changed who we shortlist. Hand the candidate a loyalty schema with twelve personal fields and ask them which four they would delete and what breaks. Weak candidates protect every field, because deleting data feels like losing capability and no engineer is ever praised for it. Strong candidates ask what each field earns and notice the ones that earn nothing. The second group is who you want designing anything that will hold a few million customer records in this market for the next five years, and they are rarer than the job title suggests.
The Part I Got Wrong
For about a year I used tokenisation as the headline answer whenever anyone asked how we handle customer data. Cards are not stored, therefore we are fine. It was true and it was a complete answer to the wrong question, and I repeated it often enough that other people in our organisation started repeating it too.
What this week has clarified is that a payment vault protects your liability, not your customer. The customer is exposed through the boring fields nobody thought to classify, and those fields were collected by product decisions rather than security ones. If there is one thing to take from the Shukran disclosure into your next planning cycle, it is that the question is not how well you protect what you store, it is why you are storing it.
Frequently Asked Questions
If no card data was taken, is the Shukran breach actually serious?
What should a product team actually change after a breach like this?
Does this mean we should hire a security engineer?
How long does stolen personal data stay dangerous?
Discuss what your product should stop storing
Bring us your schema and your roadmap. We will tell you which fields earn their place, which ones are pure liability, and what the engineer who fixes it should look like.
Let us discuss itRelated reading: Hire a data privacy engineer for UAE PDPL · Hire a cybersecurity engineer in Dubai · Build a retail loyalty app in the UAE
Sources: Landmark Group disclosure, 11 November 2026, reported by Gulf News and The National; analysis and expert comment from The National, 15 November 2026.
