MIT Technology Review
by Jamie Condliffe
June 30, 2016
AI systems are modeled after human biology, but their vision systems still work quite differently.
Computer vision has been having a moment. No more does an image recognition algorithm make dumb mistakes when looking at the world: these days, it can accurately tell you that an image contains a cat. But the way it pulls off the party trick may not be as familiar to humans as we thought.
Most computer vision systems identify features in images using neural networks, which are inspired by our own biology and are very similar in their architecture—only here, the biological sensing and neurons are swapped out for mathematical functions. Now a study by researchers at Facebook and Virginia Tech says that despite those similarities, we should be careful in assuming that both work in the same way.
To see exactly what was happening as both humans and AI analyzed an image, the researchers studied where the two focused their attention. Both were provided with blurred images and asked questions about what was happening in the picture—“Where is the cat?” for instance. Parts of the image could be selectively sharpened, one at a time, and both human and AI did so until they could answer the question. The team repeated the tests using several different algorithms.
Obviously they could both provide answers—but the interesting result is how they did so. On a scale of 1 to -1, where 1 is total agreement and -1 total disagreement, two humans scored on average 0.63 in terms of where they focused their attention across the image. With a human and an AI, the average dropped to 0.26.
In other words: the AI and human were both looking at the same image, both being asked the same question, both getting it right—but using different visual features to arrive at those same conclusions.
This is an explicit result about a phenomenon that researchers had already hinted at. In 2014, a team from Cornell University and the University of Wyoming showed that it was possible to create images that fool AI into seeing something, simply by creating a picture made up of the strong visual features that the software had come to associate with an object. Humans have a large pool of common-sense knowledge to draw on, which means they don’t get caught out by such tricks. That’s something researchers are trying to incorporate into a new breed of intelligent software that understands the semantic visual world.
But just because computers don’t use the same approach doesn’t necessarily mean they’re inferior. In fact, they may be better off ignoring the human approach altogether.
The kinds of neural networks used in computer vision usually employ a technique known as supervised learning to work out what’s happening in an image. Ultimately, their ability to associate a complex combination of patterns, textures, and shapes with the name of an object is made possible by providing the AI with a training set of images whose contents have already been labeled by a human.
But teams at Facebook and Google’s DeepMind have been experimenting with unsupervised learning systems that ingest content from video and images to learn what human faces and everyday objects look like, without any human intervention. Magic Pony, recently bought by Twitter, also shuns supervised learning, instead learning to recognize statistical patterns in images to teach itself what edges, textures, and other features should look like.
In these cases, it’s perhaps even less likely that the knowledge of the AI will be generated through a process aping that of a human. Once inspired by human brains, AI may beat us by simply being itself.
By GREG BENSINGER
Wall Street Journal
April 13, 2016
Hana Pugh, a 29-year-old event planner and new mother, buys most of her household items and baby goods online. But she doesn’t use her computer to shop. Instead, Ms. Pugh taps away at her iPhone.
“It’s quicker to pull out my phone and click ‘buy’ than to log on to my computer,” said Ms. Pugh, of Bowie, Md., who relies on Amazon.com Inc.’s app for essentials such as diapers and wipes.
Shopping on a small screen used to be a pain. But as consumers spend more of their days glued to smartphones, retailers are getting savvier with apps that ease browsing, offer rewards, suggest the right products and simplify the purchase to one click.
The so-called appification of shopping promises to radically transform the retail industry by creating new shopping habits, reshaping sales tactics and carving out winners and losers. Instead of placing one big order from a computer, people are increasingly making smaller purchases in short bursts throughout the day on their phones, a phenomenon retailers call “snacking.”
Mobile sales are booming, especially compared with sales gains from desktop computers. Last year, U.S. sales from mobile devices jumped 56% to $49.2 billion, doubling the previous year’s growth, according to comScore. Desktop sales still dwarf mobile, reaching $256.1 billion last year, but annual growth slowed to 8.1% from 12.5%.
The retailers that are succeeding are training customers to think of their smartphones like an all-day impulse aisle. Apps are able to capture data available on handsets and push consumers to buy when they have a spare moment, whether in line for a morning coffee, or, as in Ms. Pugh’s case, nursing her child.
But merchants say many shoppers on phones still shy away from buying big-ticket items such as sofas, preferring larger photos, expanded reviews and product descriptions, as well as price comparisons available on a desktop computer.
And retailers have to be careful not to seem invasive. Amazon’s shoe retailer, Zappos.com, is testing technology that highlights different products based on a phone’s operating system. Target Corp. pushes coupons to customers who have the app open in its stores. Fashion retailer ModCloth Inc. suggests some products based on a customer’s location.
Amazon’s dominance on the Web extends to mobile, and the online retailer has the top-ranked app, according to Apple. Among the top-ranked apps on both Apple and Android devices are two, young mobile-centric marketplaces, OfferUp Inc. and ContextLogic Inc.’s Wish, which last year were valued at $800 million and $3 billion, respectively, by investors.
“Mobile devices are driving demand,” said Andrew Lipsman, a comScore vice president, who has studied mobile shopping. “They can create an impulsive buying moment at any point in the day because they are with you all the time, right in your pocket.”
Selling merchandise on smartphones still poses challenges. The ease of buying single items instead of building a shopping cart can drive up retailers’ shipping costs. And consumers are more likely to window shop for products but not necessarily buy them.
People still make most of their online transactions on desktop computers. Ryan McIntyre, chief marketing officer of men’s fashion retailer JackThreads Inc., said the average cart is about $5 higher on his company’s website than its app. But he says customers spend about 10% more on average on mobile devices in a given six-month period.
Mobile devices also drive Web sales: Nearly 40% of desktop transactions in the fourth quarter took place after a customer visited the retailer’s app or mobile site, according to consulting firm Criteo.
Olivia Bryant, a 19-year-old Starbucks barista in Bakersfield, Calif., said she spends up to two hours a day shopping on her iPhone through apps such as marketplace Etsy Inc. and fashion retailer Poshmark Inc. “It’s much simpler to shop on my phone,” she said. “There aren’t a lot of distractions.”
The average U.S. consumer last year spent 3 hours and 5 minutes a day using apps, compared with 51 minutes surfing the mobile Web, according to eMarketer. Such devotion to apps isn’t lost on retailers. Target has shifted resources and staff away from Web development to its app—which it hopes to make central to all of its digital design.
“With smartphones we’re able to reach you all the time” through texts or lock-screen messages, known as push notifications, said Wish CEO Peter Szulczewski. The San Francisco startup has created an experience that mimics the mall, with a seemingly endless inventory to scroll through and bargain-basement prices.
Some apps, such as those from Zappos and online auction site Tophatter Inc., identify a user’s device to judge a customer’s possible buying power. “Our data shows that the more expensive device you have, the more you might spend,” said Tophatter CEO Ashvin Kumar.
“Millions of Amazon customers shop exclusively on a mobile device all year,” said an Amazon spokeswoman.
Retailers may be most excited to cater results based on a customer’s location, whether it is knowing a user has entered their store or is vacationing in warmer climes.
“If you’re in Australia, we might serve you swimwear when it’s winter in the Northern Hemisphere,” said Matt Kaness, CEO of San Francisco fashion retailer ModCloth, which has about 1.2 million users. EBay Inc. will highlight generators to customers shopping in areas affected by a big storm, said Hal Lawton, the company’s North America senior vice president.
Other retailers such as Wal-Mart Stores Inc. and Nordstrom Inc. have explored ways to tie in mobile shopping with stores. Customers could check in to a store through the app, providing salespeople with their purchase history, even encrypted payment information, for a quicker and more personalized shopping experience.
But the main catch for apps may be the impulsive shoppers. “On bar nights, we see drunk shopping, which is very interesting,” said Alan Tisch, CEO of fashion app Spring Inc. “Maybe there’s an opportunity there.”
3/18/2016 – by Monica Watrous
ANAHEIM, CALIF. — Nearly a decade ago, the Coca-Cola Co. found itself at a crossroads. The Atlanta-based beverage company, along with the rest of the carbonated soft drink category, had come under fire for the negative health consequences of sugar consumption. That is when the company launched its Venturing and Emerging Brands (V.E.B.) unit to identify and partner with disruptive startups in the beverage category.
“When we started V.E.B. in 2007, we certainly were challenged on innovation and finding new growth in areas that were unfamiliar to us,” said Matthew Mitchell, vice-president of portfolio strategy and ventures for V.E.B., during a March 11 panel discussion at Natural Products Expo West in Anaheim. “We talk about acquisitions and deals, but for us this was about behaving in a relationship. We knew that this was important for us to go out and not only buy brands but to invest in people because they were helping us change that dialogue.”
Coca-Cola has since invested in and partnered with such brands as Honest Tea, Suja Juice, Zico coconut water and fairlife ultra-filtered milk. Mr. Mitchell said the missions and values of the brands have influenced Coca-Cola’s business model.
Seth Goldman, co-founder of Honest Tea, said: “When we did the deal, Muhtar Kent, the c.e.o. of Coke, said, ‘At the end of the deal, if Honest Tea becomes more like Coca-Cola and we don’t become more like Honest Tea, then we failed.’”
Coca-Cola acquired Honest Tea in 2011 after an initial 40% investment in 2008. Mr. Goldman said the partnership has opened new doors for his organic beverage brand, including recent distribution in Wendy’s and Chick-fil-A restaurants.
“There isn’t anything we have done with the brand that we wouldn’t have done independently, and quite frankly wouldn’t be able to do without Coke, whether it’s getting into Wendy’s or Chick-fil-A with fair trade organic product… or upgrading to fair trade sugar in our (plastic-bottle teas),” Mr. Goldman said. “That decision came out of Atlanta, and for me, that means our DNA has penetrated.”
For Jeff Church, co-founder and chief executive officer of Suja Juice, which last year received a minority investment from Coca-Cola, the partnership is a step towards making his company’s organic cold-pressed juice beverages more affordable and accessible to mainstream consumers.
“For us, when we felt we had proof of concept for conventional channels, we really wanted to go, and in order to do that, we had to make sure our cost structure was right, and we had to make sure we had proper funding to do that because we’re not going to win the day with an $8 bottle of juice,” Mr. Church said. “It’s got to be a product where the quality and integrity is there, but the price delta between our products and traditional products is not that high, and the way to do that is to partner with someone who can help us with that.”
Suja Juice leverages Coca-Cola’s scale, distribution and access to cost structure while retaining its entrepreneurial spirit and speed to market, Mr. Church said. Honest Tea also has maintained control over all major decisions, Mr. Goldman said.
“(We had) a three-year runway to demonstrate that we could scale the business the right way and do it with integrity,” he said. “By the time we got to the three-year point where Coke had the option to buy the rest of the company, they said this is working and bought the company.”
Coca-Cola’s primary objective through V.E.B. is to seek and develop the next billion-dollar brand. The business unit tracks startups through four phases of growth: experimentation, proof of concept, pain of growth, and scale to win. V.E.B. also analyzes consumer trends to predict how the beverage marketplace will evolve over the next five to 10 years.
“For me, the exciting part is not only have we shifted the way we’re looking at the brands, but we’re looking at how we operate,” Mr. Mitchell said. “Much more at the street level … (We have become) cognizant of segmentation and cognizant that what people buy in the store on one street may be very different than what we’re buying five blocks away. How we market that, how we divide our products in each one of those stores is very important.
“That’s not easy for us. That’s a long road for us, quite frankly.”
From drug discovery to price optimization, across virtually every industry, more companies are using predictive analytics to increase revenue, reduce costs, and modernize the way they do business. Here are some examples.
Disrupt An Industry
Drug discovery has been done the same way for decades, if not centuries. Researchers have a hypothesis-driven target, screen that target against chemical compounds, and then iteratively take them through clinical trials. As history has shown, a lot of trial and error is involved, perhaps more than is necessary, particularly in this day and age. According to industry association PhRMA, it takes an average of more than 10 years and $2.6 billion to develop a drug. Pharmaceutical company BERG Health aims to change that. It is using predictive analytics and artificial intelligence (AI) to discover and develop lifesaving treatments.
“There’s no way a human can process the amount of data necessary to dissect the complexity of biology and disease into form-based discovery,” said Niven Narain, founder and CEO of BERG. “We use human tissue samples to learn about as many biological components as we can and we include that patient’s clinical and demographic data.”
Its platform builds a model of healthy individuals and then compares that to individuals with a disease. The AI then builds a model of the genes and proteins that pinpoints the core differences between health and disease. The model helps BERG target its drug discovery process. The company also uses the same process to identify which patients are the best candidates for a certain drug.
Using a single tissue sample, its platform can create more than 14 trillion data points that collectively become a “patient signature.” The patient signature indicates whether or not the individual will likely respond well to a treatment that, for example, is far more precise than first-line pancreatic cancer treatment. First-line pancreatic treatments fail 90% of the time, Narain said.
(Image: bykst via Pixabay)
Meet Customer Demand
Handmade photo product company PhotoBarn has increased its throughput 500% by creating warehouse software and lean manufacturing processes that are built around predictive analytics. Before its transformation in 2015, the company struggled to balance supply and demand.
About halfway through 2015, the company started using predictive analytics to forecast sales, inventory, and raw materials to anticipate what it would need before and during the holiday season. That and its new lean manufacturing process enabled the company to move five times more product using the same number of people.
“The spikes and volumes in the holiday period are hard to handle. In 2015, we reimagined our supply chain from suppliers to customers,” said PhotoBarn’s business analytics and marketing chief Ryan McClurkin. “We were able to handle the order volumes without hiccups [because] we’re anticipating versus reacting, and it pays huge dividends.”
Predictive analytics has helped Alabama’s Birmingham Zoo more accurately forecast attendance. As a result of that, the company can make more informed staffing and marketing decisions.
“The number of people who attend the zoo affects staffing, marketing and events planning. You could look at historical averages, but we pulled historical data and correlated that with weather data, school calendars, national holidays, [and other] variables to predict how many people would show on a given day,” said Joshua Jones, managing partner at data analytics and data science consulting firm StrategyWise.
The information is displayed on a digital dashboard that provides a much more accurate forecast. Instead of guessing that 10,000 people will come to the park based on historical information alone, Birmingham Zoo can now see it is likely that 7,131 visitors (or whatever the number happens to be) will attend on a particular day.
Create The Perfect Game
Success in the lottery industry is all about finding the right payout levels. Two of the most important factors are the sizes of the prizes and the frequency of payouts, which is why prize values and odds vary significantly in a single game. However, some games are more popular than others.
“Lotteries want to maximize their revenues so they can [contribute to] education and whatever social programs the state wants to support,” said Mather Economics director Arvid Tchivzhel. “We’ve measured responses in tickets purchased due to changes in the payout structure. You can almost build the perfect game based on where you set the payout levels and the frequency.”
Sell More Effectively
Jewelry TV (JTV), like many luxury goods retailers, was hit hard by the recession. The company tried a number of tactics to improve sales that didn’t work as well as hoped, so it eventually embraced predictive analytics.
“A regression model helps you understand what’s impacting your revenue. When you start building a predictive analytics model, it can tell you why what you’ve been doing isn’t working — customers don’t care,” said Ryan McClurkin, former director of strategic analytics at Jewelry TV and currently chief of business analytics and marketing at PhotoBarn. “Predictive analytics can tell you your customers care about this [instead].” That’s the power of predictive analytics. It allows you to see the variables you can innovate around.
By JACOB BUNGE
Feb. 1, 2016
Several startups are racing to be the first to fill U.S. consumers’ plates with laboratory-developed hamburgers and sausages that taste just as good as the kind from cattle and pigs.
Memphis Meats Inc., a San Francisco company founded by three scientists, aims in three to four years to be the first to sell meat grown from animal cells in steel tanks. Rivals including Mosa Meat and Modern Meadow Inc. also aim to bring such “cultured meat” to market in the next several years.
The competition highlights how these efforts have expanded since the 2013 taste test of a burger grown in a lab through a multiyear, $330,000 project funded by Google Inc. co-founder Sergey Brin and spearheaded by physiologist Mark Post . Reviews of the patty were mixed, but encouraged Mr. Post, who co-founded Netherlands-based Mosa Meat, to press on.
The startups’ lofty goal is to remake modern animal agriculture, which the United Nations estimates consumes one-third of the world’s grains, with about a quarter of all land used for grazing. The companies say that growing meat with cells and bioreactors—similar to fermentors used to brew beer—consumes a fraction of the nutrients, creates far less waste and avoids the need for antibiotics and additives commonly used in meat production.
“The meat industry knows their products aren’t sustainable,” said Memphis Meats Chief Executive Uma Valeti, a cardiologist and medical professor at the University of Minnesota. “We believe that in 20 years, a majority of meat sold in stores will be cultured.”
The potential payoff could be enormous—American spent $186 billion on meat and poultry in 2014—and this month, Memphis Meats plans to announce its strategy and about $2 million in funding from venture-capital firms including SOSV LLC and New Crop Capital.
Some in the meat industry are skeptical that consumers, many of whom are demanding “natural” or organic food made without additives or genetically modified ingredients, will embrace meat grown from animal cells. Representatives for major meat suppliers Tyson Foods Inc., Hormel Foods Corp. and Perdue Farms Inc. declined to comment, saying the technology was still too new.
But enthusiasm for new technology to satisfy consumers’ hunger for meat is high among venture-capital firms and Silicon Valley investors. Microsoft Corp. co-founder Bill Gates and Twitter co-founders Biz Stone and Evan Williams have invested in plant-based protein companies Beyond Meat and Impossible Foods Inc.
Memphis Meats grows meat by isolating cow and pig cells that have the capacity to renew themselves, and providing the cells with oxygen and nutrients such as sugars and minerals. These cells develop inside bioreactor tanks into skeletal muscle that can be harvested in between nine and 21 days, Mr. Valeti said.
While the source cells can be collected from animals without slaughtering them, Memphis Meats and others have relied on fetal bovine serum, drawn from unborn calves’ blood, to help start the process. Mr. Valeti said Memphis Meats will be able to replace the serum with a plant-based alternative in the near future, and Mr. Post says he also expects to be able to eliminate its use. Without the serum, there will be no need for antibiotics, according to the researchers.
Mosa Meat, which Mr. Post started with Maastricht University food technician Peter Verstrate, aims to sell cultured ground beef to high-end restaurants and specialty stores in four to five years, and is fielding interest from potential investors, Mr. Post said. Though the method’s efficiency and environmental aspects strike a chord with some consumers, “it will take time and early adopters” to catch on, he said.
Modern Meadow is working on cultured leather, which could be for sale in two to three years, according to Sarah Sclarsic, business director for the Brooklyn, N.Y., company. Meat, she said, “is a longer-term mission for us.”
The meat startups say their main challenge will be scaling up production while keeping costs low enough that cultured meat costs—and tastes—about the same as meat sliced from animals. Currently it costs about $18,000 to produce a pound of Memphis Meats’ ground beef, compared with about $4 a pound in U.S. grocery stores, according to the U.S. Department of Agriculture. Eventually Memphis Meats and Mosa Meat aspire to sell more-complex products like steak, and make meat healthier by growing cells that contain less saturated fat.
Memphis Meats officials say they have had discussions with the U.S. Department of Agriculture and the Food and Drug Administration on how their food will be regulated. The FDA would likely review the cultured meat before the USDA Food Safety & Inspection Service would begin regulating the product and how it is processed, a USDA spokesman said.
Memphis Meats plans to eventually unveil its meat at restaurants and retailers, including several Memphis-area barbecue restaurants that are co-owned by William Clem, a tissue scientist who teamed up with Mr. Valeti and Nick Genovese, a stem cell biologist, to start the company.
Mr. Clem said he has been pitching the cultured meat idea to regular guests of his chain, Baby Jack’s BBQ, some of whom are skeptical and others interested.
‘We’ve got a road map to start small and introduce it to people…’
—William Clem, Baby Jack’s BBQ and Memphis Meats
“This is probably the toughest market you can imagine for something like this. It’s Memphis, Tenn., it’s all about tradition,” Mr. Clem said. “We’ve got a road map to start small and introduce it to people and get some feedback.” Memphis Meats has discussed its product with food service distributors U.S. Foods Inc. and Sysco Corp., he added.
Steve Lieber, global brand head of BurgerFi, a Florida-based chain that serves burgers from grass-fed beef on tables made from recycled milk jugs, said his company would consider using cultured beef for a seasonal special if it tasted as good as BurgerFi’s current meat.
“We do want to be a cutting-edge company in everything we do,” he said. But “right now for millennials, the tendency toward natural is ingrained.”
SEPTEMBER 23, 2015
- By 2025, the share of tasks performed by robots will rise from a global average of around 10 percent to about 25 percent across all manufacturing industries.
- Wider robotics adoption will boost manufacturing productivity by up to 30 percent.
- As a result of higher robotics use, average manufacturing labor costs are projected to be 33 percent lower in South Korea and 18 to 25 percent lower in China, Germany, the US, and Japan than they otherwise would have been.
It has been roughly four decades since industrial robots—with mechanical arms that can be programmed to weld, paint, and pick up and place objects with monotonous regularity—first began to transform assembly lines in Europe, Japan, and the U.S. Yet walk the floor of any manufacturer, from metal shops to electronics factories, and you might be surprised by how many tasks are still performed by human hands—even some that could be done by machines. The reasons are simple: economics and capabilities. It is still less expensive to use manual labor than it is to own, operate, and maintain a robotics system, given the tasks that robots can perform. But this is about to change.
The real robotics revolution is ready to begin. Many industries are reaching an inflection point at which, for the first time, an attractive return on investment is possible for replacing manual labor with machines on a wide scale. We project that growth in the global installed base of advanced robotics will accelerate from around 2 to 3 percent annually today to around 10 percent annually during the next decade as companies begin to see the economic benefits of robotics. In some industries, more than 40 percent of manufacturing tasks will be done by robots. This development will power dramatic gains in labor productivity in many industries around the world and lead to shifts in competitiveness among manufacturing economies as fast adopters reap significant gains.
A confluence of forces will power the robotics takeoff. The prices of hardware and enabling software are projected to drop by more than 20 percent over the next decade. At the same time, the performance of robotics systems will improve by around 5 percent each year. As robots become more affordable and easier to program, a greater number of small manufacturers will be able to deploy them and integrate them more deeply into industrial supply chains. Advances in vision sensors, gripping systems, and information technology, meanwhile, are making robots smarter, more highly networked, and immensely more useful for a wider range of applications. All of these trends are occurring at a time when manufacturers in developed and developing nations alike are under mounting pressure to improve productivity in the face of rising labor costs and aging workforces.
To assess the potential impact of the coming robotics revolution on industries and national competitiveness, The Boston Consulting Group conducted an extensive analysis of 21 industries in the world’s 25 leading manufacturing export economies, which account for more than 90 percent of global trade in goods. We analyzed five common robot setups to understand the investment, cost, and performance of each. We examined every task in each of those industries to determine whether it could be replaced or augmented by advanced robotics or whether it would likely remain unchanged. After accounting for differences in labor costs, productivity, and mix by industry in each country, we developed a robust view of more than 2,600 robot-
industry-country combinations and the likely rate of adoption in each.
The following are some of the key findings of this research:
- Robotics use is reaching the takeoff point in many sectors. The share of tasks that are performed by robots will rise from a global average of around 10 percent across all manufacturing industries today to around 25 percent by 2025. Big improvements in the cost and performance of robotics systems will be the catalysts. In several industries, the cost and capabilities of advanced robots have already launched rapid adoption.
- Adoption will vary by industry and economy. Among high-cost nations, Canada, Japan, South Korea, the UK, and the U.S. currently are in the vanguard of those deploying robots; Austria, Belgium, France, Italy, and Spain are among the laggards. Some economies, such as Thailand and China, are adopting robots more aggressively than one might expect given their labor costs. Four industrial groupings—computers and electronic products; electrical equipment, appliances, and components; transportation equipment; and machinery—will account for around 75 percent of robotics installations during the next decade.
- Manufacturing productivity will surge. Wider adoption of robots, in part driven by a newfound accessibility by smaller manufacturers, will boost output per worker up to 30 percent over the medium term. These gains will be in addition to improvement from other productivity-enhancing measures, such as the implementation of lean practices.
- Savings in labor costs will be substantial. As a result of higher robotics use, the average manufacturing labor costs in 2025—when adjusted for inflation and other costs and productivity-enhancing measures—are expected to be 33 percent lower in South Korea and 18 to 25 percent lower in, for example, China, Germany, the U.S., and Japan than they otherwise would have been.
- Robots will influence national cost competitiveness. Countries that lead in the adoption of robotics will see their manufacturing cost competitiveness improve when compared with the rest of the world. South Korea, for example, is projected to improve its manufacturing cost competitiveness by 6 percentage points relative to the U.S. by 2025, assuming that all other cost factors remain unchanged. High-cost nations—such as Austria, Brazil, Russia, and Spain—that lag behind will see their relative cost competitiveness erode.
- Advanced manufacturing skills will be in very high demand. As robots become more widespread, the manufacturing tasks performed by humans will become more complex. The capacity of local workers to master new skills and the availability of programming and automation talent will replace low-cost labor as key drivers of manufacturing competitiveness in more industries. There will be a fundamental shift in the skills that workers will need in order to succeed in advanced-manufacturing plants.
Few manufacturing companies will be left untouched by the new robotics revolution. But getting the timing, cost, and location right will be critical. Investing in expensive robotics systems too early, too late, or in the wrong location could put manufacturers at a serious cost disadvantage against global competitors.
The right time for making the transition to advanced robotics will vary by industry and location. But even if that time is several years away, companies need to prepare now.
To gain competitive advantage, companies need to adopt a holistic approach to the robotics transition. We recommend that companies take the following actions:
- Understand the global landscape. First, companies need a clear picture of the trends in robot adoption around the world and in their industries. They need to know how the price and performance of robots are likely to change in comparison with the total cost of labor in each economy where they manufacture—and how this comparison is likely to change in the years ahead. They must also factor in other considerations, such as the flexibility of labor rules and the future availability of workers, that support or hinder wider robotics adoption in a given economy. It is important to keep in mind that these are moving targets.
- Benchmark the competition. Companies need to be well aware of what their competitors are currently doing and understand what they will do in the future. If robotics adoption is expected to rapidly increase in their industry, they should assume that the total cost of systems will fall. This knowledge will help companies more accurately estimate the cost and timing of investments as well as make decisions about where to locate new capacity.
- Stay technologically current. Rapid advances in technology mean that companies must stay abreast of the evolving capabilities of advanced robotics systems. They should have a clear view of whether and how quickly innovation is resolving technical barriers that so far have inhibited the use of robots, such as the ability to manipulate flexible or oddly shaped materials or to operate safely alongside workers. Just as important, when will these new applications be cost-effective? As they take stock of the new capabilities and the improvements in price and performance, many companies—even small and midsize manufacturers—may discover that installing robotics is more cost-effective than they once thought. In some cases, having a view on the evolution of robotics and automation can help a company determine whether it is better to wait for a better technology to emerge or to implement a new process that allows them to upgrade technology without having to duplicate what they have already done. In many ways, timing is crucial.
- Prepare the workforce. As more factories convert to robotics, the availability of skilled labor will become a more important factor in the decision about where to locate production. Tasks that still require manual labor will become more complex, and the ability of local workforces to master new skills will become more critical. The availability of programming and automation talent will also grow in importance. Companies and economies must prepare their workforces for the robotics revolution and should work with schools and governments to expand training in such high-compensation professions as mechanical engineering and computer programming.
- Prepare the organization. Even if the economics don’t yet favor major capital investment, companies should start preparing their global manufacturing operations now for the age of robotics. They should make sure that their networks are flexible enough to realize the benefits of robotics as installations become economically justified in different economies and as suppliers automate. They should get themselves up to speed on new advanced-manufacturing technologies and think about how they will transform their current production processes so that these technologies can achieve their potential. For many manufacturers, adapting to the age of robotics will require a transformation of their operations.
Manufacturers do not have the luxury of waiting to act until the economic conditions for robotics adoption are ripe. Our projections show that when the cost inflection point arrives, robotics installation rates are likely to accelerate rapidly. This will provide the opportunity to create a substantial competitive advantage. Companies and economies that are ready to capitalize on the opportunity will be in a position to seize global advantage in manufacturing.
Food Industry & Consumer Trends
By Elaine Watson
Jan 27, 2016
The top 10 branded processed food companies in the US have lost 4% of their market share in the past five years as smaller, more innovative brands have seized the initiative, says Rabobank.
Big food manufacturers are also sitting on outdated assets optimized to produce a small number of products at huge volumes, at a time when consumers are shunning ‘mass produced’ goods in favor of personalized, artisanal and local products, he points out.
“Some companies have already begun to shed assets, including Kraft Heinz, General Mills, Mondelez and Kellogg…. Yet probably more could be done as they reassess the risks of owning the means of production in a market where consumers are becoming increasingly needy and fickle…
“The idea that one size no longer fits all is a fundamental challenge to their business model and is being felt across the food chain… Flexibility is the future.”
Cleaning up labels, acquiring sexier brands
So what’s to be done?
When it comes to ingredients, big food companies have already made significant steps to ditch artificial colors, flavors and preservatives, embrace cage-free eggs and even label GMOs in 2016, says Fereday.
We can expect more fast-growing, entrepreneurial brands such as EPIC to get snapped up by strategic investors in 2016, says Rabobank
The problem is that consumers, particularly Millennials, don’t always give big corporations credit for such moves, so it’s not always clear what the ROI is, he observes.
“Our big fail for food companies around this strategy lies with their marketing, and how they have struggled to find their voice and target audience in the age of multimedia.”
As smaller, more entrepreneurial and mission-driven brands are seen to be more ‘authentic’ by many consumers, they will also be snapped up at an earlier stage by large CPG companies in 2016, he predicts.
“Some of these smaller companies become attractive targets surprisingly quickly.”
IRI: Categories offering quick, healthier solutions for on-the-go consumers are driving growth
His comments came as Chicago-based market researcher IRI published its latest ‘Times and Trends ‘ report observing that the US consumer packaged goods (CPG) market is characterized by declining volumes (-1.7% in 2015), and very modest dollar growth (+0.6% in 2015) that is driven largely by inflation.
However, some food and beverage categories are growing, says IRI, notably in “categories that provide quick and healthier solutions for on the go consumers”.
The top performers in terms of unit sales growth in the year to Nov 1, 2015, were refrigerated lunches (+14.2%), refrigerated tea/coffee (+10.5%), ready-to-drink tea/coffee (+10.3%), spirits/liquor (+8.3%), energy drinks (+8.1%), refrigerated salads/coleslaw (+7.8%), bottled water (+7.1%), sports drinks (+7%), other sauces (+6.8%), and bakery/snacks (+6.1%).
Healthy eating and easy-preparation trends are helping to support growth across a number of categories
It adds: “Within edibles, refrigerated lunches posted the strongest unit sales growth for the year, up 14.2%, versus overall refrigerated department sales increases of 1.1%.
“Unit sales growth came despite significant price increases, which were spurred by inflationary prices …and only minimal increases in merchandising activity.
Unit sales of bottled water surged 7.1% in US retail outlets in the 52 weeks to November 1, 2015, says IRI
“The bottled water category saw volume sales increase 7.8% during the past year, amid relatively flat (+0.5%) prices and the proliferation of enhanced bottled waters. Healthy eating and easy-preparation trends are helping to support growth across a number of categories, including refrigerated salad/coleslaw, which saw unit sales increase 7.8% for the year.”
Performance was weakest in the frozen foods sector, where unit sales declined 1.5%, and strongest in beverages, where unit sales increased 2.9%, adds the report: “Frozen dinners/entrees and frozen pizza saw sharp unit sales declines during the past year (4.6% and 3.6%, respectively).”
‘Explosive’ online sales growth
While online sales of consumer packaged goods account for less than 2% of overall industry sales, growth has been “explosive”, notes IRI.
“In fact, average annual growth of online CPG spending has topped 15% since 2010. Between 2013 and the end of 2018, the Internet will account for about 50% of industry growth, or $28bn.”
On a hot, swampy November afternoon at IMG Academy in Bradenton, Florida, the private school’s boys’ lacrosse team is huddling around the bench draining water bottles after a round of drills. It’s a scene like any other on high school fields around the country . . . except here there’s a small crew of men sitting off to the side with laptops, running diagnostics on the team’s fluid intake, tracking each player individually through smart chip–enabled refillable water bottles.
Each “smart cap” bottle is digitally linked to a specific player. It works with an app to calculate how much he sweats in an average practice, how much sodium he loses, and how much he needs to drink to maintain optimal performance. Each bottle is filled with a drink formula that corresponds to an individual player’s sweat type. A microchip and a small turbine in the spout measure how much he takes with every sip. LED lights on the cap help him pace his drinking, showing whether he’s ahead of or behind his target. This is the future of athletic hydration. It’s also the future of Gatorade.
The brand’s new high-tech focus can be traced back to senior vice president and general manager Brett O’Brien’s decision in 2014 to build an internal innovation unit to look beyond bottle shapes and new flavors and toward a higher mission. After all, something had to be done. Gatorade sales in the first half of 2009 had fallen 18% year over year; new competitors such as VitaminWater, Red Bull, and Monster had gained influence and market share; and the product, born 50 years ago in a University of Florida lab as the original sports specialty drink, had become known more as a hangover helper than a high-performance elixir. O’Brien and the innovation group set about getting Gatorade back into shape, transforming it into an elite sports brand on par with Nike and Under Armour. “It’s not just about capturing a bigger part of a marketplace,” O’Brien says. “It’s about filling a void for your consumer, and ours are athletes.”
Gatorade started by looking at how athletes were already using its products. For instance, pro athletes had long been mixing sodium powder packs and other additives into their Gatorade (the brand even makes its own, called Gatorlytes). Now the company is looking to cut out the extra steps by creating 12 different formulas served in small egg-shaped pods that mix with water—to be used in new bottles featuring the smart-cap design. The formulas will range in carb, calorie, and electrolyte levels to optimize fluid recovery. To know which formula is right for which athlete, the brand has developed a suite of products and technologies that work together to measure and track individuals’ data. A smart scale linked to a tablet and software, for example, catalogs a player’s weight, along with training time and intensity, to make fluid and carb-intake recommendations. A patch, like a near-field communication chip-enabled Band-Aid, will analyze a player’s sweat and communicate with the digital platform to identify his sweat type—which will determine sodium, electrolyte, and additional fluid-intake needs.
All of this technology was developed in conjunction with the brand’s nutrition-minded Gatorade Sports Science Institute, fueled by PepsiCo’s 40% increase in R&D spending between 2011 and 2014. Now Gatorade is testing it on the field with top high school, college, and pro athletes, including the Brazilian national soccer team, the Kansas City Chiefs, the Boston Celtics, FC Barcelona, and the University of Florida. For Kansas City Chiefs wide receiver Jeremy Maclin, the idea of customized hydration is intriguing. “I don’t think there’s any athlete on this earth who wouldn’t take an advantage if proven to work,” he says.
Xavi Cortadellas, Gatorade’s global innovation and design senior director (who joined the brand in 2011 after 11 years at Nike), believes the move into customization and personalization was inevitable, given the rise of data-driven insight in sports in general. “It’s about whether you want to be a premium brand or not,” he says. “If you want to play in the premium space, then you have to be delivering personalization.” As part of Cortadellas’s plan to scale these new ideas, a simple version of the bottle will be available to the public by mid-2016—it can be customized with digitally printed cap graphics including team and player names, numbers, and logos. “Here at IMG, one of the things the kids were really excited about was having the ability to customize the bottle ‘just like NikeID personalized my shoes,’ ” says Cortadellas. “That’s what you want to hear.”
Ultimately, Gatorade wants to move beyond the field. The brand is developing and testing a new line of food products expected to hit shelves over the next several years—protein bars, vegetable-based nitrate boosts, protein-enriched yogurt—that will embody what nutritionists have long been preaching to athletes: that they are athletes all day, even at rest. “In all these spaces where we have no or minimal share, the potential for growth is gigantic,” Cortadellas says.
Not long ago, a Gatorade ad tagline asked, “Is it in you?” If Cortadellas and his colleagues are right, the question will soon become moot. They’ll already know.