Q4 Record Sales & Starbucks’ One Problem + Two Opportunities

Starbucks posted a stellar earnings result today, with net revenue at record level of $6.3bn, beating expectations in many aspects.

Starbucks up after hours with earnings beats | Source: CNBC

While a single beat doesn’t indicate a new stage of growth, in the long run there are three issues I think matter the most and should be watched closely.

1 | Problem: Frappuccino’s decline

It was an problem Starbucks facing over the years. People are leaning towards healthier products (at least a very obvious trend in Cali), which usually means less sugar and less calories. That’s a problem for Starbucks’ Frappuccino, as explicitly mentioned in CEO’s presentation in June.

How did/will Starbucks address this? Product Mix & Innovation

Strategy a) Big push for healthier product lines – Nitro Cold Brew & Refresher series etc.

Starbucks Nitro Cold Brew | Source: Starbucks

– Nitro Cold Brew “is expected to be available in nearly 1,500 stores in 26 markets by the end of 2017″ (SBUX Jul. 2017 Press Release) -> “[Starbucks is] accelerating this platform to more than 2,800 stores by the close of fiscal year 2018, up to more than 6,000 stores by year-end fiscal year 2019″ (SBUX Q3 Earnings Call Transcript). A 4x availability expansion for cold brew in 3 years.

– Several new products in the Refresher category were introduced, e.g. “Dragonfruit” introduced Jun. 2018, “Pink Drink” introduced last year, etc.

Dragon Fruit | Source: Starbucks
Pink Drink | Source: Starbucks

Strategy b) Instagramable & limited edition within the Frappuccino category

– Unicorn Frappuccino (Apr. 19-23, 2017), Zombie Frappuccino (Oct. 26-31, 2017), Christmas Tree Frappuccino (Dec. 7 – 11, 2017), Crystal Ball Frappuccino (Mar. 22-26, 2018), Witch’s Brew Frappuccino (starting Oct. 25 for a limited time while supplies last)… among many others.

– Plus, Starbucks’ new Frappucino recipe has fewer calories and less sugar, part of its efforts to reduce sugar by 25% by 2020.

Unicorn Frappuccino | Source: time.com
Zombie Frappuccino | Source: Starbucks
Christmas Tree Frappuccino | Source: Starbucks
Crystal Ball Frappuccino | Source: Starbucks
Witch’s Brew Frappuccino | Source: Starbucks
2 | Opportunity – Digital Interactions

Starbucks Rewards could serve a similar role as Amazon Prime. In past last 3 month, loyalty program accounts for 14% of all transactions and US loyalty members contributed 40% of US sales. That’s what happened in the Amazon case, where its Prime members out-spend non-members significantly.

Digital relationship makes it easier to incentivize purchases, market new products/initiatives, bring in more collaborations (e.g. Spotify, Pokémon GO), expand membership offerings and more.

Starbuck’s push for afternoon consumptions is also facilitated by the digitalized promotions.

The room to grow digital relationships is still large – currently 15.3 million global active members, only representing ~22% of its 70 million global customers base. “Additionally, drive-thru, out-the-window and Mobile Order and Pay combined grew to more than 50% of the way customers are ordering, up more than 10 percentage points in just two years” according to COO.

3 | Opportunity – China Growth

It’s more debatable on Starbucks’ China future. Just want to highlight a few sure things.

Starbucks took full ownership in China East & opened flagship Roastery in Shanghai in 2017 – definitely the right moves here.

China’s coffee consumption will explode, even considering major cities alone. Younger generations will consume more coffee and they will represent an increasing proportion of the overall urban population.

China coffee consumption potential | Source: Starbucks 6/19 Presentation

Coffee even has a role to play in China’s GDP growth by boosting workers’ average productivity and the “culture” of working overtime.

Herding effect is stronger in China and “Instagramable & Limited” strategy may provide better outcomes if properly implemented.


Other things worth noticing – Starbucks’ food, packaged goods with Nestlé, next-generation store design, coffee supplies, SKU outside the Starbucks’ core products, etc…

140 -> 280: How Twitter’s Length Limit Shapes The Language Used

Social network platforms such as Twitter and Weibo, are where we post our words/expressions, and is taking an essential role in todays’ Modern Communication.

Interestingly, but not surprisingly, they are also providing a feedback loop to change our ways of communication (and our thinking – but that’s another different story). After all, we are one of those animals with natural herding inclinations.

Taking a very simple example – as Twitter was becoming more popular, people started to lean heavily towards the use of new abbreviations, partially due to its 140-word limit. (Similarly, when Blackberries were popular, certain abbreviations were created, used and spread among professionals in texts & emails)

In 2017, Twitter officially made its “140-word limit” a history [Weibo ended the limit in 2016] and doubled the character limit to 280. One resulting impact is that people are spelling out abbreviations and acronyms more often, according to Twitter’s report.

Source: TwitterData

TwitterData also touted that “people are saying ‘please’ (+54%) and ‘thank you’ (+22%) more.” in the first year of doubling limit.

[Something worth noting on the data – 1) need to consider the difference in total number of tweets 2) need to consider other ways of saying ‘please’ and ‘thank you’…]

It is really nice that people could be more friendly and polite (not excessively tho). It would be even nicer if people could inherit their concise and efficient use of language in the 140-word era & all the other positive impacts Modern Communication made on us.

Apple’s Pricer New Products and 2-Way Strategy

Today Apple announced new products of three lines of business – MacBook Air, iPad Pro and Mac mini.

MacBook Air 2018 Gold | Source: Apple
iPad Pro 2018 11-inch and 12.9-inch | Source: Apple
Mac mini 2018 | Source: Apple

Besides the new features and superior performance, Apple seems to be more addicted to aggressive price tags. Today’s event is the latest episode of a series, starting from the 10-year anniversary model iPhone X last year.

[discussion below is based on the most basic consumer version of each product]

9/12/2017 – iPhone X introduced @ $999, a 30% increase from the high-end phone (iPhone 7 Plus @ $769) one year before

5/27/2018 – updated iPad introduced @ $329, flat with the new iPad debut price in Apr. 2017.

7/12/2018 – updated MacBook Pro @ $1,799; the price seems to be the same as Oct. 2016’s debut price with Touch Bar, but is a 20% increase in terms of available low-end option (MacBook Pro with no Touch Bar @ $1,499).

9/12/2018 – iPhone XR @ $749, increased by 7% ($50) compared to the low-end phone last year (iPhone 8 @$699) ; Apple Watch Series 4 @ $399, increased by 21% ($70) from a year ago @ $329.

10/30/2018 – new MacBook Air @ $1,199, a 20% increase from the old model years ago @ $999; updated iPad Pro @ $799, a 23% ($150) increase from previous Jun. 2017 version @ $649; new Mac mini @ $799, a 60% increase from the old model years ago @ $499; redesigned Apple Pencil 2 @ $129, increased by 30% from $99 when it was introduced 3 years ago

The 2-way strategy looks clear: 1. use entry-level new products (e.g. iPad & iPhone XR) to attract new users and compete with other firm’s often lower-priced products. Educational market and some international markets are important here, especially for new-user growth, ensuring an increasing number of total active users in Apple’s ecosystem and getting more market shares in terms of shipment. So there is barely any increase in price. 2. raise prices by 20-30% in other product lines to keep margin (for example Apple Watch price increase might be more associated with costs) or to compensate lower margins in entry-level products.

Apple’s overall average gross margin from 2016 to present is actually lower than the average from 2014 to 2015 (38.44% vs. 39.63%, down ~120bp), even with higher gross margins from service revenues weighing in more recently.

Nasdaq Composite and Value Creation

[This blog post is mainly focused on NASDAQ]

1. How big is today’s loss & Are we really panicking?

The market rout on 10/24 (NASDAQ Composite dropped 329.14 points to 7,108.40) made it the largest retreat so far this year, surpassing the previous 316-points loss merely 2 weeks ago (10/10).

US stock index 10/24 tumble | Source: WSJ, Dow Jones Market Data

While the 4.43% retreat is widely “touted” as the biggest one-day loss since August 2011, it’s also the 3rd largest daily point loss in NASDAQ history (the other two happened in 2000 during the dot-com bubble).

Top 10 NASDAQ Largest Point Decrease as of 10/24/2018
Date Point Change
04/14/2000 -355.49
04/03/2000 -349.15
10/24/2018 -329.13
10/10/2018 -315.96
04/12/2000 -286.27
02/08/2018 -274.82
02/05/2018 -273.42
04/10/2000 -258.25
01/04/2000 -229.46
03/27/2018 -211.73
Reference: NASDAQ Largest Point and Percentage Decreases in the NASDAQ Composite Index

However, in percentage wise, we could see NASDAQ’s 10 biggest single day decrease ranging from -11.35% to -7.23% (3 of which related to the 2008 financial crisis, 3 related to the dot-com crash, another 3 related to the 1987 crash).

“-4.43%” is not a big deal & that’s not how investors react if they are really panicking.

2. All about inflation?

It’s crucial to incorporate inflation into analysis.

NASDAQ climbed above 5,000 in Mar. 2015, and it took another 7 weeks before it reached the dot-com bubble peak of 5048.61 (intra-day record was 5,132.52).

NASDAQ Composite climbed above 5,000 in Mar. 2015 | Source: Quartz, FactSet

However, when we adjust for inflation, things are different.

Doing a simple math – U.S. CPI (urban, all items ex. food & energy) is 181.3 in 2000 (annual average) and 256.5 in 2018 (H1 average); so a 8,109.69 peak in August should be around 5,732 in 2000-price-level.

So NASDAQ Composite spent more than 18 years to grow 13.5% (from peak to peak), while fundamentally technology has gone through tremendous evolutions.

CPI – urban consumers, ex. food & energy (Jan. 1999 – Sep. 2018) | Source: U.S. Bureau of Labor Statistics

Similarly, WSJ reported in January this year that “Nasdaq Tops Inflation-Adjusted High [7269.89] from Dot-Com Boom”; and the chart below shows the inflation-adjusted path.

Inflation-adjusted Nasdaq Composite | Source: MacroTrends
Why did we wait for 18 years just for a come-back? So value creation has little to do with the price? Is this all about inflation in the end?
3. Sit back and relax – It’s NOT all in vain

It’s a sure thing that companies can be over- or undervalued over time. But the benefits and growths are also real.

We have (unprecedented) iPhone in 2007 and (finally) massively produced Tesla Model 3 in 2018; we have 1 TB cloud storage for $9.99/month on Dropbox/Google Drive and unlimited storage of business account for $15/month on Box or $10/month on OneDrive.

Development in biotech is equally impressive. Sequencing cost dropped from ~$100 million per genome in early 2000s to ~$1,000 in 2016; and we are curing/curbing more types of cancer with unprecedented success rate and less harmful methods. Other advancements, for example those in neuroscience or surgical robots, are no less exciting.

And we can see their reflections in stock price (e.g. AAPL TSLA ILMN ISRG).

While the NASDAQ Composite in 2000s may be overvalued, comparatively in 2018 it is more supported by concrete revenue and earnings. Promises and expectations are still built in the price, but things are a lot better.

From a similar discussion in Mar. 2015 when NASDAQ reached 5,000 again after 15 years – P/E is not even close to insane levels.

Nasdaq historical LTM P/E | Source: CNN, Nasdaq, Bloomberg

When we call it a bubble (e.g. dot-com bubble), it is usually characterized by an increasing [absurdly large] difference between price and actual value created.

Revenue or earnings can rarely be doubled consecutively between regular quarters but prices can. That’s what happened in 1999 when Nasdaq Composite rose 85.59% (vs. 28.24% in 2017) and 13 large-cap stocks rose over 1000% (vs. Amazon +56%, Netflix +55%, Facebook +53%, Apple +46%, Alphabet +33% in 2017).

While a 10-fold rise in stock price is tough to catch up by revenues and earnings, I see a 50% rise doable and reasonable (e.g. could be a combination of 80% EPS growth + 20% P/E decrease, 50% EPS growth + P/E unchanged, or 25% EPS growth + 20% P/E increase), as long as the new P/E is justified by the future prospect.

I am not saying we shouldn’t anticipate a correction; but NASDAQ Composite hasn’t been super crazy neither.

I would like to conclude this post here by saying that – I believe in the future and the real benefits of tech [if correctly used]. Some areas might seem to be more over-promised than others, and some risks are looming on the horizon [to negatively impact global economy in a nontrivial manner], but true value creation should and will always be valued.


Appendix – what does S&P look like

S&P forward P/E as of Sep. 2018 | Source: JP Morgan Asset Management

Facebook’s Home Device with Camera and Microphone

上周 Facebook 发布了新款硬件 Portal – 一个以 FB 好友为电话簿的新型视频座机..

作为科技巨头中最后一个入场的,在 Amazon Echo 和 Google Home 开打很久、Microsoft 去年发布与 Harman Kardon 合作的 Invoke、Apple Homepod 今年早些时候入场后,Facebook 也是挤了进来,成为最后一个加入的美国科技巨头。

先说说各家的优势。

Amazon 是先发优势,提前入场熟悉用户优化 Alexa 拓展平台;还可以跟 Amazon 以后发布的各项智能家居配合(比如那个微波炉)。Amazon 之前也有做硬件和系统的基础。

Google 肯定是要抢这个未来 smart home device 核心位置的,nest 已布局多年,推广 google assistant(智能语音),顺带连接 google 本身各项服务,以及推广 youtube (premium) music。Google 的硬件实力近几年也突飞猛进,系统平台的强势更不用说。

Microsoft 这个略尴尬,也不算它自己做的 device,主要是推广了 Microsoft 的 Cortana 智能语音。

Apple 不用说,音乐和硬件一直是强项,Siri 也是有基础的,iOS 本身也有 Homekit。虽然目前没有带屏幕,但我相信就是日程上的事。已经有 FaceTime 和 FaceID,想象空间绝对够。

Facebook 强在社交网络 (FB Messenger) 的嫁接,自动生成电话薄。

然而,Facebook Portal 细想一下,在很多方面都很尴尬。

首先,Facebook 没有自己的智能语音,Portal 是直接用的 Amazon Alexa。在这一群科技巨头面前直接弱三分 (Alexa, Google Assistant, Cortana, Siri)。

其次,Facebook 没有平台能力,没有智能家居想象空间。本身社交跟智能家居的契合度(目前)就很小。未来跟 Oculus AR/VR 有多少契合很难说,至少短期没有。

三者,作为音箱用,FB 没有音乐类服务,仅是接入 Spotify/Pandora;有视频类的服务,但才开始,FB Watch/Show。另一方面说,由于 FB 内部 app 的分割,Instagram TV / 直播的想象空间又暂未实现。

第四,从 monetization 的角度,FB 逃不出 ads。在巨头之中,由于上述局限,FB 是最容易/最早会在 Portal 里加 fb ads 的.. Google 也差不离,但由于它更想做成大平台,可能离直接投放 ads 很远。

最后,还有 FB 的隐私和数据问题。今年以来一直是热点。考虑到美国对于家庭隐私的重视,估计 Portal 有点悬。

但从视频通话的角度,FB Portal 是很可以的。但放到大的背景中,突出两个字 – 尴尬。

New Art by Banksy & Sotheby’s Performance

10 月 5 日,在苏富比拍卖的 Banksy 2006 年作品 Girl with balloon 在落锤后被作者本人设计的碎纸画框碎了一半。

整个作品被重新命名 Love is in the bin。

Banksy B. 1974 GIRL WITH BALLOON signed and dedicated on the reverse spray paint and acrylic on canvas, mounted on board, in artist’s frame 101 by 78 by 18 cm. 39 3/4 by 30 3/4 by 7 in. Executed in 2006, this work is unique.
BANKSY, LOVE IS IN THE BIN, 2018

作者 Banksy 在 Instagram 上 po 出了自带碎纸的画框制作小视频,并 quota 毕加索

“The urge to destroy is also a creative urge” – Picasso

原本约 1 百万英镑的拍卖成交价,在一波戏剧性的操作后作品将大幅增值。

虽说 Banksy 说拍卖行并无参与 – 但无论怎样,这已经不单单是一幅画作,整个戏剧性故事都离不开 Sotheby。根据苏富比的 blog,这是首次在拍卖中卖出了行为艺术 (it marked the first time a piece of live performance art had been sold at auction, read more here)。

这是 Banksy,Sotheby,买方,以及现场观众共同完成的艺术。

期待多年以后的 Love is in the bin 的拍卖。

Tesla Has New Chairman, Facebook and More?

美国传统的董事会管理层双层结构在硅谷和科技行业不太流行。

给予创始人(们)极大自由和权力后,now funds are fighting back。

在公司蒸蒸日上的时候,换 board chair 的声音几乎听不见或被忽略;但一旦公司出现非短期问题,换 board chair 也被更多提及,并等待一些 “catalyst”

Tesla 要换 board chair 的声音不是一天两天,直到 Elon Musk 的 Twitter 被 SEC 盯上。 9/29,Musk 和 SEC 和解条款中包括了卸任董事长。

今天,几家基金要 Facebook 换 chairman 的提案再次被关注。过去几年 FB 表现已经很好,在 Zuckerberg 带领下社交巨头进入收入利润高增长。然而今年 Zuckerberg 的日子并不好过。一系列隐私方面的问题和激进的广告策略,让用户国会等各方都找到很多机会来 dis FB。

提案方包括 Trillium Asset Management (53,000 shares),以及纽约市 Comptroller 和三个州的 State Treasurer –  Illinois (190,712 shares) , Rhode Island (168,230 shares) and Pennsylvania (38,737 shares), New York City Pension Funds (4.5 million shares),大约近 500 万股(占 0.2%)。

现在 FB 的双层股权基本保证靠投票很难的;提案更多的为了公开讨论引起关注,为以后的契机做铺垫;如果 FB 或 Zuckerberg 出现一些现象级问题,可能也会有跟 Elon Musk 类似的场外和解。

GL to all

Netflix’s 130 Million 付费用户

Netflix 第三季度财报还是很可以的。

用户/付费用户的超预期增长支撑了 Netflix 的 growth story,股价反应很积极。

但同时我也感觉,高企的估值即将触碰到一些天花板。

  1. 美国付费用户已有  5696 万
    • 算上每个付费账户可以多人 access,这个数字接近美国付费 cable 用户数约 9000 万 – 1 亿
  2. 海外付费用户 7346 万
    • 虽说海外市场更广阔,算上印度、巴西等国家人口是相当多。但付费能力的用户依然有限,除去中国美国,除去网速欠佳,适合 Netflix 且能盈利的地区并不多。
  3. 其它公司如 Disney,  Amazon, Youtube 的 streaming service 将抢夺用户
  4. 在内容上,成本将提升。上游版权制作能力等整合聚集,bargain power 增强
  5. 商业模式上局限,不像爱奇艺可以采用被广泛接受的广告模式

与 Amazon 的关系 – 作为 AWS 的最重要客户之一,以及 Prime Video 的竞争对手,微妙;还有未来 Netflix subscription 可能与 Amazon Prime 终极会员体系的竞争。

在泛娱乐 streaming 领域,Facebook TV, Instagram TV, Snapchat 等争夺用户屏幕时间。

这么看来,Netflix 在 2019,有很多挑战需要正面对待。

Money-Printing Skydeck & 帝国大厦

芝加哥地标之一的 Willis Tower 是相当有江湖地位的一座高楼。

1973 年建成的时候叫 Sears Tower,成为当时全球第一高楼,1450 ft;2009 年更名为 Willis Tower;2015 年,BlackStone $1.3bn 拿下。

Willis Tower 第 103 层的 Skydeck 是相当著名的景点,是美国 highest observation deck

Image Source: theskydeck.com

周末去了一趟,也顺道好奇了一下 Skydeck 的盈利水平 –

The 103rd-floor Skydeck Chicago, which draws 1.7 million visitors per year…

In 2016, Skydeck revenue totaled $35.5 million, up from $28.7 million in 2014.

Operating income at the Skydeck rose to $29.9 million in 2016, up 22 percent from 2014, according to the Morningstar report. The observatory also is an extremely lucrative business, generating an operating profit margin of 84 percent last year.

部分报道的原文 link 已经打不开了,不确定数字是否属实;先把搜索的截图存下来

money printing…

简单 run 一下数字:平均 5000人/天,350天/年,$20/人 -> $35m,差不多

由于整个建筑的 modular 设计,skydeck 所在的第 103 层可以粗略估计一下,为 75 * 75 * 2 = 11,250 sqft,assume 是 10k sqft。

稍微查了一下附近 west loop 的 office 租金,assume $45/sqft/yr

skydeck 的收益用 $35m/10k sqrt 算一下,约是 $3500/sqrt/yr

money printing…


对比一下 Empire State Building – 有 86 层和 102 层两个 observatory,2017 年和 2016 年分别有 4.05 million 和 4.25 million 人次的 visitors,年收入分别为 $127.1 million 和 $124.8 million,人均约 $30。

考虑到纽约对比芝加哥、帝国大厦对比 Willis Tower,一些 premium 完全合理。不过帝国大厦的 operating margin 较低,2017 年约 76%,2016 年约 77%。

帝国大厦的 observatories 的收入也是占到帝国大厦总收入约 40%…

Source: Empire State Realty Trust 2017 Annual Report


Takeaway – 相比于租写字楼的公司,游客的钱好赚,好赚太多。

 



Gig Economy & Employment Issues in the 21st Century

今天公布的美国上一周 (9/23 – 9/29) Weekly Jobless Claims 降至 20.7 万例,接近 49 年低点。49 年最低点在两周前触及,9/9 – 9/15 那周的 Initial Jobless Claim 为 20.2 万例,是 1969 年 11 月以来的最低点。

昨天公布的 nonfarm private sector employment 数据也很出色,增加 23 万例。市场对于明天将公布 non-farm payroll 数据期望相当之高。

8 月的 unemployment rate 已降至 3.9%。Full-employment 社会的话题今年以来获得更多关注…

但在经济数据的背后,不应该忽视的是新型商业模式对于这些数据的影响 – uber, lyft, airbnb, grubhub, etc… so called “gig economy”。同时,原先简单粗暴的 employed/unemployed 划分已经不能充分反应新经济的就业问题。

把旧的思维模式强加在当前经济模式下,必然会有各式各样的问题:

  1. 原本 unemployed 人口通过 uber 等公司作为收入来源,没有 actively looking 或者没有 desire for or availability for work;然后一部分也不会当自己是 employed,由于是数据来源是 survey,最终导致部分人口没算在 labor force 里,不算做失业人口。一个参考是 08 – 13 年 labor participation rate 的下降。
  2.  gig economy 中的 “employer” 各方面没有保护,比如最低工资保障 mini wage,健康保险 health insurance,养老 pension,失业保护 unemployment insurance 等
    • 各州对于相关纠纷的判决有所不同,但都开始关注 independent contractor / employee 的问题。
    • 参考 case 1:加州 Lawson v. GrubHub,2 月 Judge Corley 支持 independent contractor [trial court];在下面的 case 之后,5 月 Lawson 上诉至 U.S. Court of Appeals for Ninth Circuit
    • 参考 case 2:加州 Dynamex Operations West, Inc. v. Superior Court,4 月底 California Supreme Court 支持按 employer 的划分,并采用 ABC test
      • (A) that the worker is free from the control and direction of the hiring entity in connection with the performance of the work, both under the contract for the performance of the work and in fact;
      • (B) that the worker performs work that is outside the usual course of the hiring entity’s business; and
      • (C) that the worker is customarily engaged in an independently established trade, occupation, or business of the same nature as the work performed.
    • 参考 case 3:纽约州 The New York State Unemployment Insurance Appeal Board 于 7 月支持 uber driver 在失业保险方面应算作 uber 的 employee
    • 相关 case – 领取失业福利的同时开 uber:Lowman v. Unemployment Compensation Board of Review [Pennsylvania Commonwealth Court],“by occasionally accepting offers of work such as that afforded by Uber, an individual does not make a ‘positive step’ toward establishing an independent business”,参见 Bruce L. Baldwin 文章
  3.  under-employment 问题。当前 gig economy 中的 independent contractor 更偏向于 lower-skilled 工作;以及拉低整体的 productivity。这是个更长期的问题。

BLS 有一项统计,在 2017 年 5 月,美国约有 1.6 million electronically mediated workers, accounting for 1.0 percent of total employment;uber 今年 9 月在自己的 blog 中称已有超过 90 万的 driver,两个数据也算 consistent。

下图来自今年 9 月 JPM 的一篇 research,更直观,但有 bias,在 JP Morgan 开户的人更可能参与在 gig economy 中

从整体看,这还不是一个大规模群体,但不难想象这是一个持续增长的群体。在 AI 和娱乐的未来,gig economy 的作用将更加明显。