什么？大数据是更好地吃零食的秘方……(译）

译者：北理大 数据 教育苏道强老师

吃零食是一种正式热潮。根据最近尼尔森的调查，超过 90 ％的美国人每天都吃零食 – 其中四分之一的人每天吃三到五次。但不只是美国人在吃零食。吃零食现在是 3740 亿美元的全球产业，它每年增长 2 ％。

欧洲人每年在零食上动用接近 1700 亿美元，而亚太地区占 500 亿美元，以及拉丁美洲 300 亿美元。美国整整落后于欧洲 1200 亿美元的年销售额。

在爱吃零食的人中对制造商和零售商最有吸引力的是新千年一代。新千年一代已占总人口的 14 ％，在 2015 年他们预计将有 24500 亿美元每年的消费能力，这甚至超过婴儿潮一代 2018 年的消费能力。但制造商们迅速认识到新千年一代不能被视为一个同质的群体。毕竟，新千年一代通常被定义为在 1982 年和 2000 年间出生的人，这意味着每个人都从一个 14 岁的高中生成为 32 岁的成熟人。

这就是为什么资本对零食的趋势以及迎合新千年一代的口味，会需要一种新的方法。人口统计数据，客户信息和反馈，以及其他数据在以前所未有的速度影响产品开发。其结果是，很多公司都不仅仅转向大 数据分析 来发现模式和可操作的见解，而且认识到以更迅速的计算来转变口味和日益增多的成堆的数据。

揭开不断变化的口味的模式

有了正确的数据，大数据分析工具可以帮助制造商了解更改的口味喜好以及哪些杂货店受消费者喜爱。这对制造商，零售商和消费者来说是一个三赢的局面。制造商可以在合适的地方放置合适的产品，零售商可以拉动销售，而消费者有机会获得更多他们喜欢的产品。这在世界各地都是真理。

大数据还可以帮助制造商更迅速地对不断变化的口味做出反应。例如，尽管很多西班牙食品已在美国成为主流，其他民族的食物在小众市场仍然很受欢迎。通过转向大数据分析，制造商可以更好地了解新出现的食品景观。

CPG 制造商的新模式

然而，对于许多制造商，这将意味着他们是如何运作的根本转变。根据埃森哲表示，到 2015 年， 85 ％的财富 500 强企业将无法使用大数据获得竞争优势。虽然技术一直在改善供应链等过程中发挥作用，许多公司需要重新考虑技术的作用，使用它更吸引客户而不是勉强维持新的效率。

这意味着要创造被称为 “ 数据供应链 ” 的东西 – 本质上确保数据不会躺在在筒仓里，而是会在整个组织中流转循环，这样公司就可以更方便地挖掘新方向的信息。

当它涉及到可以说是历史上最具衔接性的千禧一代时，这是特别重要的。通过整理所有新千年提供的数据，从他们的社交媒体供稿到他们超市的会员卡再到手机的数据资料，制造商可以更好地在这个人口及其各种零食偏好的不断增长中给自己定位。

CPG 制造商如何挖掘数据已给他们的客户匹配零食

现在最流行的小吃是巧克力（不要担心，紧随其后的是水果和蔬菜），口味因地区，年龄，收入等因素而不同。有些千禧一代在全职工作，而另一些则仍然在学校。有些和父母住在一起，而其他和自己另一半生活在一起。这里有五个方面的技术帮助制造商，零售商和消费者发现他们最渴望的小吃和口味。

1. 了解你的客户。大数据分析可以帮助揭示谁在购买什么样的产品以及更好地调整将来的符合他们口味的货物。猜测和假设客户想要的东西已不再是一种选择。现在的千禧一代购买什么小吃呢呢？吃零食的喜好是如何随着新千年一代的年龄的改变而改变的呢？如何才能将它关联到未来的产品计划中？

2. 监视店内体验。三分之一的购物者使用手机购物。不管是否检查价格或查找菜单，消费者都在积极寻找参与的商店。事实上，一项研究显示，近 73 ％的受访者表示，他们希望在他们的手机能有价格比较服务。同样的研究发现，超市的会员卡是一个友好地获取消费者数据的方式。近三分之二的受访者表示，只要他们的数据是安全的，他们对于零售商利用他们的购物习惯和购买历史记录来提供产品和服务，并不反感。

3. 点击进入社交媒体。从 YouTube 明星到 Instagrammers ，千禧一代通过社交媒体制造的名气并且支撑生活。这些影响者提供了一个在特定平台上向特定的受众确定的品牌主张的机会。另外，从消费者中获取社交媒体帖子，尤其是品牌的互动，这可以给消费者提供一定见解，然后可以导向生产新小吃。

4. 依靠认知计算。由于数据增长的速度和数量在不断增加，机器学习将会使企业能够更迅速地对包含在数据中的新思路做出反应。使用本体，可以发现关系并且增强理解，机器可以逐渐消化数据，从中吸取教训，并通过持续的互动磨练来发现新千年一代的口味喜好。在提供可操作的见解（即真正的理解）方面，认知计算超出了数据挖掘。

5. 使用数据来做。麦考密克＆ Company 的 FlavorPrint 活动是针对消费者的口味进行认知计算的一个例子。 FlavorPrint 提供基于口味偏好，食材成分和厨房用具的食谱。通过学习消费者的喜好，该平台随时间细化建议，这使得网站的重复使用率增加了一倍并且使得在网站上花费的时间增加了九倍。

随着世界各地吃零食的持续增长，并随着新千年一代的购买力日益支配所有的人口，看来明天的最重要的口味创造者可能是机器。

英语原文：

For better snacking, big data is the secret ingredient

Snacking is officially a craze. More than nine out of 10 Americans snack daily — a quarter of them three to five times every day, according to a recent Nielsen survey. But it’s not just people in the U.S. who are sneaking a bite. Snacking is now a $374 billion global industry, and it’s growing 2 percent annually.

Europeans spend nearly $170 billion annually on snacks, while the Asia-Pacific region accounts for $50 billion, and Latin America for $30 billion. The U.S. is right behind Europe at $120 billion in annual sales.

Among the snackers most attractive to manufacturers and retailers are millennials. Already representing 14 percent of the population, millennials are expected to have $2.45 trillion in annual spending power by 2015 and exceed baby boomers’ spending by 2018. But manufacturers are quickly learning that millennials can’t be treated as a homogeneous group. After all, millennials are commonly defined as anyone born between 1982 and 2000, meaning everyone from a 14-year-old high school student to a mature 32-year-old.

That’s why capitalizing on the snacking trend and catering to the tastes of millennials will require a new approach. Demographics, customer information and feedback, and other data are impacting product development at an unprecedented pace. As a result, many of these companies are turning not only to big data analytics to find patterns and actionable insights, but to cognitive computing to react more quickly to shifting tastes and increasingly unwieldy piles of data.

Uncovering the patterns in changing tastes

Armed with the right data, big data analytics tools can help manufacturers understand changing taste preferences and which grocery stores cater to niche consumers. It’s a win-win-win situation for manufacturers, retailers, and consumers. Manufacturers can place the right products in the right locations, retailers can drive sales, and consumers have increased access to the products they prefer. This is true around the globe.

Big data also helps manufacturers react more quickly to changing tastes. For example, although many Hispanic foods have gone mainstream in the U.S., other ethnic foods remain popular only in niche markets. By turning to big data analytics, manufacturers can better understand the emerging food landscape.

A new paradigm for CPG manufacturers

For many manufacturers, however, this will mean a fundamental change in how they operate. Through 2015, 85 percent of Fortune 500 companies will be unable to use big data for competitive advantage, according to Accenture. While technology has long played a role in improving supply chains and other processes, many of these companies need to rethink the role of technology, using it more to engage customers than eke out new efficiencies.

What this means is creating what’s been termed a “data supply chain” — essentially making sure that data doesn’t lie fallow in a silo, but is actively circulated throughout an organization so the company can more easily mine that information for new directions.

This is especially important when it comes to millennials, who are arguably the most connected generation in history. By compiling all the data that millennials offer, from their social media feeds to their supermarket loyalty cards to mobile phone data, manufacturers can better position themselves to tap into the growing clout of this demographic and their various snacking preferences.

How CPG manufacturers mine data to pair snacks with their audience

While the most popular snack right now is chocolate (don’t worry, it’s followed closely by fruits and vegetables), tastes vary by region, age, income, and other factors. Some millennials work full time while others are still in school. Some live with their parents while others live with their significant other. Here are five ways technology is helping manufacturers, retailers, and consumers uncover the snacks and tastes they most desire.

1. Know your audience. Big data analytics can help uncover who’s buying what products and better tailor future goods to their professed tastes. Guessing and assuming what the customer wants is no longer an option. What snacks are millennials buying now? How do snacking preferences change as millennials age? How does that correlate to future product plans?

2. Monitor the in-store experience. One in three grocery shoppers use a mobile phone while shopping. Whether to check prices or look up recipes, consumers are actively looking for engagement in stores. In fact, one study showed that nearly 73 percent of survey respondents said they wanted price comparison services on their mobile phones. The same study found that supermarket loyalty cards are a consumer-friendly way to pull data. Nearly two-thirds of respondents said they were fine with retailers using their shopping habits and purchase history to offer products and services as long as their data was safe.

3. Tap into social media. From YouTube stars to Instagrammers, millennials are making names and a living for themselves through social media. These influencers offer an opportunity to identify brand advocates to reach a specific audience on a specific platform. In addition, pulling in social media posts from consumers, especially brand interactions, can offer insights into consumer sentiment that can then be directed toward the production of new snacks.

4. Rely on cognitive computing. As the speed and volume of data continue to increase, machine learning will allow organizations to react more quickly to the insights contained in that data. Using ontologies that can discover relationships and aid understanding, machines can increasingly digest data, learn from it, and hone millennials’ taste preferences through ongoing interactions. Cognitive computing goes beyond data mining to provide actionable insights (i.e., real understanding).

5. Use data to engage. McCormick & Company’s FlavorPrint campaign is one example of cognitive computing targeting the tastes of consumers. FlavorPrint offers recipes based on taste preferences, available ingredients, and kitchen appliances. By learning a consumer’s preferences, the platform refines the recommendations over time, which doubled repeat usage of the site and increased the amount of time spent on the site ninefold.

As snacking continues to grow around the world, and as millennials’ buying power increasingly dominates all demographics, it appears the most important tastemakers of tomorrow are likely to be machines.