Zara just made the same mistake it first made again in 2014. And we think the real lesson is about why no standing checkpoint survived twelve years and a full staff turnover.

———

Zara pulled a children's Halloween costume off its site this week after customers pointed out it resembled the striped uniforms worn in concentration camps during the holocaust. The company apologized, in effect, by quietly removing the listing and saying nothing further.

This would be less interesting had it been the first time it happened but it's not. The fact of the matter is that Zara has been here before. In 2014, the company faced almost identical criticism over a striped children's shirt bearing a six-point star. Different garment but the same pattern and same company, twelve years apart.

If you've ever shipped something, caught a problem after it went out, and spent the rest of the day wondering how it got past everyone, you already understand what the real question here is… and it's not "how did Zara miss this." It's "how did Zara miss this twice, with twelve years and you'd think an entirely different set of people in the room the second time?"

What Happened

Zara's striped children's "groom" costume went up for sale for £39.99 ahead of Halloween and came down in just a few days, after customers noted its resemblance to concentration camp uniforms worn during the holocaust. This is at least the second time this specific pattern has caused this specific complaint for this specific company. In 2014, Zara pulled a striped children's shirt featuring a six-point star after nearly identical criticism.

The kicker? Zara isn't alone this season, either. Target separately faced backlash this fall over a Halloween costume evocative of blackface. A whole other company with the same underlying failure.. a product moved from concept all the way through design, sourcing and final review, and shipped to stores and websites, without anyone in that chain stopping to ask an obvious question before it went live.

We're not saying anything in either case points to a single person deliberately trying to produce something offensive, but that's what we actually find to be the more useful fact here, not a mitigating one. Malice is relatively easy to root out once you find it. But it's harder to root out is the much more common failure mode behind both of these incidents: a review process that exists but doesn't actually catch anything, because nobody owns the specific job of asking the question that would have caught it.

This Isn't a Training Issue. More Like "What's Going On With The Systems?"

We think that the easy read on this story is that Zara needs better sensitivity training, or a more diverse design team, or clearer guidelines about historical imagery. Those things might all genuinely help… but none of them explain the twelve-year gap, because training and guidelines are exactly the kind of thing that gets refreshed, re-issued, and re-forgotten on a predictable cycle, and twelve years is more than enough time for an entire incident, and the lesson attached to it, to fully leave institutional memory.

What doesn't disappear though, albeit only if it's built correctly, is the system. The fix for a blunder like this can't just be a better-trained person. Instead, a checkpoint needs to be assigned that doesn't depend on any one person remembering the 2014 incident at all, because more likely than not, the people that were in the room in 2026 almost certainly weren't the people in the room in 2014.

There's a specific reason this kind of mistake survives inside otherwise careful companies, and it's worth naming because it has nothing to do with carelessness. The people closest to a piece of creative, the ones who've seen every version of it from concept sketch to final sample, are usually the ones least likely to notice what's wrong with it. Familiarity is exactly what erases our instinct to ask an obvious question. That's not a character flaw in any individual designer or buyer. It's a predictable failure of judgment under repeated exposure, and the only reliable thing you can do to fix that is a check that doesn't rely on anyone's eyes staying fresh after the fortieth review.

There's also a mundane, unglamorous reason this particular check keeps losing out to other priorities: it doesn't come with a deadline attached. A product launch has a ship date. A sourcing deadline has a calendar reminder. "Does this resemble a symbol? Bad or good? Cultural relevance?, and have we actually checked" doesn't have one of those unless someone puts in the effort to build one, which means it's exactly the kind of check that gets skipped first when a team is moving at retail-calendar speed with new lines shipping every week.

Here's the thing that a small brand would worry about more than it worries a Zara. Zara can take and absorb this. They'll pull the item, ride out the bad news cycle (ugg), and will very likely report ordinary numbers next quarter, but that's because it has the scale and balance sheet to treat a reputational hit as a rounding error. A business doing $1.5 million in annual revenue doesn't have that luxury. If that exact same failure happens to you.. a default nobody ever questioned, in a product, an ad, a piece of copy, can do lasting, specific damage to a business that size in a way scale simply can't absorb for it. The businesses with the least room for error are, in practice, usually the ones with the least formal process standing between a bad idea and a published one.

How We Build This Habit Into Client Work

We build a version of this checkpoint into every client relationship, and we don't treat it as a legal or compliance function, because almost none of our clients have one of those and most don't need one. We treat it as a named job: one person whose responsibility explicitly includes asking, before anything ships, could this be read badly, and have we actually checked, not assumed.

In practice that's as plain as a shared checklist that we go through for every piece of creative before it goes out: who reviewed it, what they were specifically checking for, and a sign-off date. It takes a few minutes per asset. Those few minutes are cheap. What they prevent usually isn't.

The same discipline applies to copy and claims, not just images, and it's often the part businesses skip first because text feels lower-risk than visuals. A testimonial nobody verified, a statistic pulled from a source nobody checked, a before-and-after claim nobody can actually substantiate: these cause the same kind of damage as a bad image, just on a longer fuse, because they tend to surface after a customer or a competitor checks, not the moment the content goes live.

None of this requires a big department or a some huge formal document nobody reads. It requires deciding, in advance, that this specific question has an owner and a deadline, the same way every other part of a launch does. That's the entire fix. It's also exactly the fix nobody at Zara appears to have built, twice.

Where This Leaves You

Zara will be fine. It almost always is, because that's what scale buys a company: the ability to survive the exact mistake a smaller business can't, and the ability to make the same one twice without it costing enough to force a real fix.

So the useful question this week isn't whether Zara needs better training. It's whether your own business has a specific, named, standing checkpoint before anything ships, with someone whose job it is to ask the obvious question out loud, or whether you've just been assuming, like Zara apparently did for twelve years, that someone already checked.