Cristina Pérez, Innovation Director at Kantar, has spent more than 30 years researching the launch of new products. She has witnessed spectacular successes and unexpected failures, and above all maintains that consumers, both then and now, tend to behave in much the same way: “I always use the example of Brad Pitt. I’m married to my husband, and I’m faithful. But, of course, if Brad Pitt turns up… that faithfulness may not withstand certain stimuli,” she says with a smile. “You may find that a consumer has bought the same product on nine out of their last ten trips to the supermarket, but that doesn’t guarantee loyalty. They may simply be buying it because it’s readily available in their preferred store. But if one day it disappears, they won’t mind switching to something else.”

Human fickleness is perhaps the only thing that has not changed in the trial-and-error world of consumer goods. According to Promarca, around 20,000 product references are launched in Spain every year. Most are minor modifications, such as a new flavour or different packaging, to products that already exist rather than genuine innovations, but nearly all go through a verification process that begins with an idea discussed around a boardroom table; continues with concept testing; moves on to real-world consumer trials in controlled groups; and, with a bit of luck, ends up on supermarket shelves.

“Technology has completely transformed the testing of products and services,” says Débora Santos, co-founder of the Five Star Award, a Portuguese benchmarking system that has recently begun operating in Spain. “Advanced analytics allow us to track consumers’ behavior in real time. Consumers have become permanent testers.”

Where once relatively modest databases were used, companies now rely on massive panels. Kantar, for example, maintains a database of 1.5 million people willing to take part in brand experiments. Companies can also run tests using their own data gathered through loyalty programmes, such as store cards or newsletter subscriptions, or through an evolution of traditional sampling. What once meant handing out a small cube of cheese in a supermarket now involves sending free samples to carefully segmented and highly engaged audiences who provide feedback. Results can now be obtained in days rather than the weeks of work that were previously required.

Everything is both more complicated and easier at the same time. E-commerce provides an immediate picture of sales performance, but the routes to consumers have multiplied through social media, online reviews, and influencers. “Consumers are still the same, but they have more choice; they are more demanding, more critical, and they have many more touchpoints with products, such as an advertisement, a recommendation on their phone, a Google review or a photo on Instagram. All of this has changed the decision-making process,” Pérez explains.

Quantity is not quality

The history of marketing shows that carrying out multiple tests before a product launch does not necessarily mean testing well. Sometimes anxiety or budget constraints lead brands to skip important steps. Common mistakes can arise, for example, from failing to adapt consumer panels to cultural shifts in an aging population or in societies experiencing growing immigration. Errors may occur right at the beginning, by choosing the wrong idea, or at the end, through executives’ excessive confidence in a particular innovation. And the consequences can be disastrous.

One of the most infamous product unveilings in history was probably the presentation by U.S. tycoon Elon Musk and Tesla’s chief designer, Franz von Holzhausen, of the Cybertruck in November 2019. The vehicle, which looks like a cross between a tank and an alien spacecraft, suffered an embarrassing setback when its windows shattered after Holzhausen hurled two steel balls at them in an attempt to demonstrate, before hundreds of guests filming eagerly on their phones, that the armored glass was “unbreakable.” Tesla’s shares plunged in the aftermath.

Failures are inherent to innovation and business progress, but few companies have the courage to acknowledge them. Ikea is one of the rare exceptions willing to share a misstep for this article. Manuel Delgado, marketing manager at the Swedish giant, says that for Ikea, doing things better means “daring to experiment.” In the late 1990s, the company launched a range of inflatable sofas called Air, a project inspired by founder Ingvar Kamprad’s obsession with finding smarter ways to manufacture and transport furniture.

“They brought the Innerlig inflatable sofa to market. The concept was innovative, but the technology wasn’t ready. Customers used hot air from hair dryers to inflate it, and the valves developed leaks. The plastic covers were uncomfortable, and the price was high,” Delgado recalls. In 2026, Ikea launched the PS armchair, a revamped version of that earlier experiment. “It’s performing very well because the air-chamber material is better, it has a comfortable textile cover, and it doesn’t make noise like the previous one.”

Something similar happened at the Spanish food brand Helios, known for its jams and pantry staples. The brand identified an opportunity among so-called flexitarians, people who are not vegetarian but want to reduce their meat consumption, and believed they might welcome a pasta sauce tailored to their needs.

“Historically, we had operated in that category, but we were very small players and didn’t really know why,” explains Iker Asolo from Helios. The company studied format issues, product ranges and market trends. It launched a line of classic pasta sauces alongside three vegetarian options: arrabbiata, basil, and one called vegañesa, in which meat was replaced with a plant-based protein alternative that took two years to develop. “It was formulated from fava beans and peas. It turned out spectacularly. The concept tests and tastings were great,” says Asolo.

But consumers only embraced the classic flavors and largely ignored the experiment. “The impression we took away was that, at times, by focusing on the specific approach, we don’t see the consumer’s bigger picture,” says Asolo. The moral of the story? Replacing meat with naturally vegetarian alternatives proved easier than creating an imitation version that tried to win over fans of the traditional Bolognese sauce.

Some things have changed for the better, such as the cost of testing. José Luis Nueno, a professor at Spain’s IESE Business School, has been involved in the launch of hundreds of products. “In the past, if you made a mistake, you were in big trouble because the processes were expensive and required large samples,” he says. “People’s opinions are as unreliable today as they were before, but now research is more observational, based on what consumers actually did.”

Artificial intelligence is beginning to change everything, says Pérez. According to Kantar’s innovation director, “Now you can use it to run preliminary tests for relatively little money.”

Helios has already had the idea put to them by its analytics providers. “They tell us: don’t waste time doing tests with humans. Databases are faster and cheaper. But we haven’t tried it yet; we feel more comfortable doing it the way we always have,” says Asolo.

Will we eventually consume products that have been tested only by AI? That is probably already happening, but, according to Pérez, not just any AI will do. It needs to be based on models trained on data from hundreds of thousands of previous tests. “With AI we can search, anticipate and define future trends,” she says. “We incorporate it into several phases of our process. You can test solely with AI when you have a know‑how backed by thousands of data points. I’m in favor of AI, but always with a human layer to help interpret results.”

She contrasts this with what she often hears from disappointed clients: “Some tell you they test with ChatGPT, but what data is it basing its answers on? It can be terribly wrong.”

Débora Santos also sees advantages in predictive models, which are helping make product launches faster and cheaper. But, she warns, this creates a new challenge: “A larger strategic risk. We have an excess of data that we don’t always have the capacity to interpret. We don’t always distinguish noise from valuable insight. Many companies do not have the agility, concern or time to make that distinction, and that can lead to decisions made too quickly on data that aren’t real.”

Determining whether a toothpaste is likely to succeed is, ultimately, relatively straightforward. Other things, such as services or new technologies, are much harder to assess. Juan José Güemes, president of the International Entrepreneurship Center and vice-president of IE University in Madrid, points to the case of Meta, which invested nearly $70 billion in the metaverse, according to Bloomberg’s latest estimate, a virtual world that few people talk about anymore.

“This case is very interesting because they even changed the company name, convinced they would create a platform to commercialize digital content,” says Güemes. “But it didn’t work. When they abandoned that idea they regained stock market value.”

Selling the impossible

Rubén Ferreiro, one of the founders of the startup investment fund Lanai Ventures, has led numerous digital ventures and knows that the failure rate is extraordinarily high. “Product-development cycles are now much shorter thanks to AI, and they cost less. That means it’s cheaper to test digital products in real market conditions.”

That does not necessarily make success any more likely. A modest investment in highly targeted advertising aimed at potential customers may be enough to test a new app for group travel, but it can still fail for any number of reasons. “I think companies fail when they try to sell something nobody wants. I’ve seen plenty of sound business plans, but 90% fall short of expectations.”

Sometimes, he adds, the market simply is not ready to embrace a new service. “For example, when the metaverse launched, not even computers were ready. It wasn’t just that users weren’t willing to live a parallel life in that universe; the user experience was terrible. Now, by contrast, hardware is so small and efficient that it can fit inside the arm of a pair of glasses. For certain innovations, the technological capability has to be there first.”

Then there is the problem of overcrowding. “Shelf space is shrinking, there are more and more products… do we really need a new protein‑infused pasta? There used to be only a handful of pasta varieties; now there are dozens. And if you can’t connect with shoppers and stand out on that shelf, your product will fail. Retailers don’t give you second chances,” says Pérez.

But don’t be discouraged. Sometimes the magic comes from the most unlikely ideas. A couple of years ago, a well-known international brand tested a mayonnaise-scented perfume. The fragrance sold out within weeks.

New Coke (Coca‑Cola, 1985) and Crystal Pepsi (1992): Rivals even in failure

The world’s soft-drink giants know a thing or two about getting it wrong, which may be one reason they have remained dominant for so long. Coca-Cola, the brand that mastered the art of selling experiences as much as beverages, stumbled in 1985 when it attempted to reformulate its iconic recipe to compete more effectively with its arch-rival Pepsi.

New Coke was sweeter and smoother than the original and had performed strongly in consumer tests. Yet the company underestimated the emotional attachment customers felt toward the classic formula. In the United States, loyal drinkers launched a backlash campaign, with some even stockpiling cases of the original Coke at home. The Atlanta-based company withdrew the product less than 100 days after its launch.

Less than a decade later, Pepsi suffered a setback of its own when it tried to position its brand around a healthier image. Crystal Pepsi was a clear cola, transparent in appearance despite tasting broadly similar to a conventional soft drink. Consumers struggled to reconcile what they saw with what they expected a cola to be, and many perceived the product as a contradiction.

Since then, both companies have continued to innovate, with mixed results. Some products, such as Cherry Coke or Coca-Cola Jack Daniel’s, a ready-to-drink beverage containing 5% alcohol, have carved out a niche and remain on supermarket shelves. Others, including Pepsi Blue and Aquarius Cola, have faded into obscurity, remembered only by a small group of consumers and marketing enthusiasts.

Heinz EZ Squirt (2000): Multicolored ketchup

Some products simply go too far, and Heinz’s psychedelic ketchup range from the early 2000s may be a case in point. Its eye-catching colours, ranging from green to purple, generated plenty of curiosity, and the company initially reported strong sales for the new sauces. The quirky-history website Mental Floss notes that Heinz sold more than 25 million bottles of the condiment during its first three years on the market. But, like so many fads, its popularity faded almost as quickly as it had grown.

Betamax (1975). The best doesn’t always win

Those with a few grey hairs will remember the VHS-Betamax war. Betamax, the home-video format introduced by Sony in 1975, spent years battling VHS, launched a year later by JVC and its parent company, Matsushita (Panasonic). It is an example of how one technological standard can prevail over another without necessarily being superior in quality. In fact, experts generally regarded Betamax as an excellent product, but VHS ultimately overtook it by offering certain advantages, including longer recording times and broader commercial reach thanks to a larger network of licensing agreements.

Metaverse (2021): A spectacular crash

The metaverse is perhaps the most expensive misadventure in modern corporate history. A few months ago, Bloomberg estimated that Mark Zuckerberg’s company had invested more than $70 billion in the virtual-reality vision it began pursuing in early 2021, promising a digital world where people could work, socialize, entertain themselves and, of course, shop.

The market was projected to generate “hundreds of billions of dollars” by 2031. The company went so far as to rename Facebook as Meta following the announcement. Yet four years later, Horizon Worlds, the social platform designed to host those virtual experiences, had shut down.

LaserDisc (1978): Not even good for scaring away birds

Madrid-based arts and culture collective Basurama featured the format in some of its exhibitions on obsolescence, and many Spaniards may still remember it, although it never gained much traction in Spain as it did in markets such as Japan. The LaserDisc was one of the first disc-based systems for storing audiovisual content.

As tech website Xataka recalls, it was developed by MCA and Philips and launched in the United States in 1978 alongside the release of Jaws. It suffered from several drawbacks: its large size made it cumbersome and impractical, its storage capacity was limited, and it did not allow users to record or rewrite content.

Concorde (1969): Flying too fast

The supersonic passenger aircraft that made its first flight from Toulouse in 1969 passed into history in October 2003, brought down by a combination of economic pressures and the crash near Paris in July 2000 that claimed 113 lives.

It could cover the route between the French capital and New York in three and a half hours at Mach 2 (roughly 1,350 miles per hour). Its dazzling technological achievements were matched only by the cost of a ticket, which stood at around $9,000 in today’s money. Concorde’s retirement marked the end of one of civil aviation’s great technological adventures, making possible routine commercial flights at altitudes of between 55,000 and 60,000 feet.

Lotus 1‑2‑3: The Stone Age of spreadsheets

Lotus 1-2-3 was a pioneering spreadsheet program launched in 1983 by Lotus Development Corporation, which was later acquired by IBM in 1996. It was one of the first PC applications to prove genuinely useful for businesses and quickly became a cornerstone of the emerging office-software industry. It was eventually overtaken by Microsoft Excel, a reminder that being first to market does not always guarantee long-term success.

Segway (2001): A misstep in urban transportation

Considered the grandfather of modern electric scooters, the original Segway released by entrepreneur Dean Kamen was a two‑wheeled, self‑balancing electric vehicle that used the rider’s body weight to steer and maintain stability, allowing for safe travel over relatively even terrain.

It was billed as a revolution in urban transportation, but its high price tag, around $5,000 at launch, and its bulky dimensions made it impractical for everyday use. Even so, some sightseeing companies continue to operate fleets of Segways for guided tours. The company discontinued the product in 2020 and has since focused on manufacturing more conventional scooters.

Amazon Fire Phone (2014). A smartphone doomed from the start

Arriving too early can be a problem, but so can arriving too late. In 2014, Jeff Bezos announced that Amazon would enter the smartphone market with the Fire Phone, hoping to compete with the iPhone and powerful Asian manufacturers. Positioned in the upper-midrange segment, the device launched in the United States with a price tag of $649.

Just 14 months later, Amazon had slashed the price to $159 and cancelled plans for future versions. Losses were estimated at $170 million. A few months ago, Reuters reported that the e-commerce giant had revived the idea of developing a smartphone, this time designed to integrate closely with Alexa and provide users with a more direct way of interacting with the platform.

Google Glass (2013): Blurred vision

Thirteen years ago, Google’s engineers imagined a world in which technological innovation would be seamlessly layered onto reality through a simple pair of glasses. The device was unveiled in a promotional video showing a man moving effortlessly between the physical and digital worlds while having breakfast, meeting a friend and playing the ukulele.

The augmented-reality smart glasses launched for developers in the United States and the United Kingdom at a price of around €1,300. But technical limitations and privacy concerns prevented them from gaining mainstream acceptance, and they disappeared from the consumer market just two years later.

The torch has since been taken up by Meta, which last autumn unveiled its Meta Ray-Ban Display glasses. While outwardly resembling conventional eyewear, they incorporate a display that provides access to the artificial-intelligence features developed by Mark Zuckerberg’s company.

The DMC DeLorean (1981): ‘Back to the Future’

Almost as memorable as the legacy of Back to the Future, the futuristic film that captured the imagination of several generations, the DMC DeLorean was the only car ever produced by DeLorean Motor Company (DMC). Built between 1981 and 1982, the four-wheeled cult object featured the same gull-wing doors that helped make it famous on screen and rolled off the production line at DMC’s factory in Dunmurry, Northern Ireland.

The plant produced around 8,500 vehicles before the company went bankrupt. It was the steep price of building an entire business around a single product.

Terra (Telefónica, 1999). A digital David that never toppled Goliath

In June 2017, Terra, the internet portal launched by Spanish telecommunications company Telefónica with the ambition of becoming the leading Spanish-language destination online and competing with the U.S. tech giants, shut down in most of the countries where it still operated after 18 years in business.

Created under Telefónica chairman Juan Villalonga at the dawn of the internet age, Terra became one of the most visited websites in the Spanish-speaking world during the early 2000s. The company pursued an aggressive expansion strategy, making multibillion-dollar acquisitions, including the purchase of the U.S. portal Lycos for $12.5 billion. It would later sell the business for barely 1% of that amount. It failed to build a profitable, sustainable business.

Tuenti (2006). Spain’s home-grown Facebook

Founded in 2006 as a social network for teenagers, Tuenti grew to 13.5 million users and a workforce of 300 employees. Touted as Spain’s answer to Facebook, the platform, founded by Zaryn Dentzel, ultimately failed to keep pace with rival social networks and never managed to turn a profit.

After several strategic pivots, the service gradually disappeared. Tuenti launched a mobile operator in an attempt to carve out a place in Spain’s highly competitive telecoms market and later experimented with apps that allowed users to call and chat without using mobile data. Neither strategy succeeded.

Telepizza’s Pizzalada (2014): A strange Mediterranean twist

The Pizzalada was one of the strangest products ever to emerge from the fast-food industry. The hybrid of pizza and Caesar salad combined a bacon pizza base with lettuce, cherry tomatoes, rocket and breaded chicken. It failed to catch on, despite being conceived as a healthier offering built around more natural ingredients and intended as a nod to the Mediterranean diet.

Uterqüe (2008). Inditex’s stumble

Launched as a chain specializing in premium accessories, Uterqüe was intended to become the most upscale brand in the Inditex stable. Yet despite operating a network of standalone stores, it never accounted for more than 0.5% of the group’s revenue.

Thanks to the commercial muscle of the Spanish fashion giant, the brand expanded to more than 80 stores across 16 countries. But in 2022 those stores were absorbed into Massimo Dutti, another of Inditex’s flagship chains.

Together with Often, the menswear brand launched in 2003 and now largely forgotten, Uterqüe stands as one of the few notable failures in Inditex’s history.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition

A Donald Trump immigration policy has suffered another judicial setback. A federal appeals court rejected the administration’s strategy of detaining certain migrants without giving them the opportunity to seek release on bond and compared the scope of the measure to one of the darkest episodes in U.S. history: the mass internment of Japanese Americans during World War II.

In a 2-1 decision, the Fourth Circuit Court of Appeals ruled that the government cannot interpret immigration laws in a way that allows for mandatory detention of millions of people living in the country who are facing deportation proceedings. Judge Nicole Berner, who wrote the majority opinion alongside her colleague Barbara Keenan, warned about the consequences of mass detention without due process protections.

“Mass detention of marginalized individuals without due process harkens back to some of the darkest moments in our country’s history,” Berner wrote. The judge also noted that during World War II, the government ordered the mass detention of people of Japanese ancestry in the United States.

Berner said those detentions constituted a “historical stain” and argued that the interpretation of a law currently defended by the government could authorize a policy on an even larger scale based on an ambiguous legal provision approved three decades ago. She also cited the detention of Chinese immigrants in the late 19th century as another precedent for such measures.

The ruling represents the ninth rejection by a federal appeals court of the interpretation adopted by the Trump administration. Only two of the appeals courts that have considered the issue have backed the government’s position, creating a split that greatly increases the chances that the dispute will reach the Supreme Court.

The conflict stems from the different interpretation of a law adopted by U.S. Immigration and Customs Enforcement (ICE) since July 2025. For decades, administrations of both parties, including Trump’s first administration, had primarily applied mandatory detention provisions to people detained upon entering the United States or near the border. Those detained inside the country could generally request a hearing to determine whether they could be released on bond, with some exceptions provided by law.

The new interpretation considerably broadened that scope. ICE argued that people who entered the country without inspection could be considered applicants for admission even after having lived in the United States for years and therefore could be subject to mandatory detention without the right to a bond hearing while their immigration proceedings move forward.

The measure has affected even people with immigration applications that have been pending for years because of administrative backlogs. It also triggered a wave of lawsuits. More than 460 federal judges have rejected the government’s interpretation, while just over 50 have upheld it. The rulings against the policy have been issued by judges appointed by presidents of both parties.

The Trump administration nevertheless maintains that the 1996 Immigration Act allows it to apply mandatory detention beyond the border and that previous administrations simply did not exercise all the authority that, according to its interpretation, they already possessed. The Department of Homeland Security said it was confident in its legal position and noted that it had recently asked the Supreme Court to review a similar decision.

Judge Allison Rushing, a Trump appointee, was the only member of the Fourth Circuit to dissent. She argued that the fact that previous administrations had not used all the authority the administration considers available does not mean that the current exercise of that authority is unlawful.

Sign up for our weekly newsletter to get more English-language news coverage from EL PAÍS USA Edition