Kill Switches Required in California


Starting July 1, 2015, every smartphone sold in California will be required to have a kill switch, which allows the user to make the phone unusable if stolen.

The phone itself is valuable but data such as pictures and account information are even more important. If the phone is lost, it can easily be replaced but if the content is not backed up, the information will not be recoverable. The best case scenario is losing the phone itself but in the worst case possible, the thief may use the data from the phone maliciously. Technology evolves so quickly that it is always steps ahead of the law.

I am both surprised and impressed at how fast the law was instituted in order to mitigate ubiquitious phone crimes. I have heard many anecdotal stories where phone thieves target people on public transportation in the city. At the very least, I believe that this law will make thieves think twice before committing crimes. After a couple of years, it will be interesting to look at the theft data to see if the kill switch discouraged phone thieves from stealing.

 Kill Switches


The Challenges of Big Data


Big data is a buzz word popular in the valley that refers to a large collection of data that is mined to provide insights that allows for better decisions in various fields. Currently, when combining different sets of data, one of the challenge is standardizing data formats.

Palantir, a CIA funded start-up, has been solving this problem for over 10 years, well before the buzz word became mainstream. It solves world issues with a platform that analyzes unstructured data sets to draws relationships. Palantir provides solutions for government and large enterprises but with how ubiquitous data is now, other start-ups are attacking the same problem for commercial purposes.

Companies are trying to make data more accessible for the masses so that even individuals without a computing background can easily comprehend data and ask the right questions. But as data becomes easier to collect and analyze, this will bring privacy concerns. It might take a high profile lawsuit to trigger further government regulation. It is just a question of when this will grow to be of greater concern to everyone.

Big Data


Seaworld – Rebuilding Brand


As an animal enthusiast, I had to watch Blackfish, a documentary that explores Seaworld’s treatment of killer whales and the detrimental effects of keeping them in captivity. With boycotts against the theme park, Seaworld experienced declining attendance and profits since 2013. This shows how branding is extremely important to a business. I believe that the park is only redesigning the whales’ tank due to lower profits. They are trying to desperately rebuild their image so they can recoup the losses.

However, a company needs time and resources to build a reputation in order to develop trust with customers. For example, the BP Gulf oil spill was one of the largest environmental disaster in recent history. BP spent over millions in advertising after the oil spill to restore their brand. Despite Seaworld’s investment, their image may not be easily recoverable. If they ever do, it will take years before people forget about the documentary.



The Rivalry Between Lyft and Uber


I am an occasional user of Uber and I have taken Lyft once. The customer experience is very different between the two services and they emphasize that through the drivers’ mannerisms. The rivalry has just escalated a level up.

Uber and Lyft accused each other of sabotaging each others’ businesses by ordering and cancelling rides as well as poaching drivers. This is Uber’s second time of being charged of disrupting a rival’s service. The constant accusations between the two is akin to children fighting on the playground. Instead of spending time on sabotaging one another, wouldn’t both be better off if they focus on making the companies fulfill the mission of bringing people together? Or perhaps this is a PR stunt that is timely enough to promote Lyft Line, the new carpooling feature that allows riders to share fees.

Uber V. Lyft


A Beta Look: Google Shopping Express


On my way to work, I saw a little van with a blue parachute painted on it. This sparked my interest in Google Shopping Express.

Google is breaking into e-commerce with Google Shopping Express. Google Shopping Express allows local retailers such as Target and Walgreens to sell their goods through Google’s website. After the user places an order on the website, Google will send their couriers to pick up items from the retailer and deliver them during the same day.

Amazon is almost synonomous with e-commerce. I wanted to find out how Google compares with the e-commerce behemoth so I did a business study by first finding the relationship between marketing spend and sales. Then using Amazon as a benchmark, I analyzed Google’s profit and losses if Google were to gain Amazon’s market shares in three different cases: 5%, 10%, and 15%.

The only scenario where Google is profitable is the optimistic case where it has 15% of Amazon’s market share. However, even if Google Shopping Express were unprofitable, it would still have positive externalities in terms of ad revenues. The shopping service would lead to an increase in the volume of product searches, which translates to greater product listing ad revenues. Through the service, the search giant can collect even more personal data that allows them to improve on how they target ads. Although Google Shopping Express seems to be Google’s attempt at competing with Amazon, it can have a significant impact to Google’s core search business.

Download (PDF, 472KB)


Social Experiments



A month after FaceBook disclosed their experiment on altering users’ News Feed to elicit different emotional expressions, OKCupid revealed that it has done multiple social experiments in where the company intentionally mismatched users.

Although the ethical aspect is debatable, I do not think that this revelation should come as a surprise. There is a reason why the companies show fine prints before allowing access to the service. I doubt that many people spend the time to read the detail. Despite the users’ frustration, I think many people will still use Facebook because it has been integrated into many peoples’ daily lives. This begs the question of how and if these popular services should be regulated. Who owns the data and what can companies do with it? How much can the government do to protect people on these services?

Facebook Experiment

OKCupid Experiment


A Beta Look: Facebook Ads


I have always been curious as to how Facebook makes money from ads so I decided to run my own campaign for my puppy named Miko.

I created a page for Miko and my goal was to gain page likes with the lowest cost. I chose four different visuals with four distinct phrases for two pricing methods: optimized CPC and CPA. I allocated $5 a day for each of the pricing mode. After running the campaigns for a couple of days, I noticed a significant difference in terms of the metrics even when comparing between the two pricing methods and that Facebook channels a majority of the traffic to the most popular ad leaving the others with low impression, click-through, and conversion rates.

Due to the low CPA and high conversion rate for optimized CPA, I lowered the budget from optimized CPC so that most of the resources were invested in the former. Although the ads in optimized CPA had lower impression rates, it had high conversion rates and low CPA and CPC. This indicates that the quality of users are higher and thus, are more likely to like the page.

Download (PDF, 256KB)


A Beta Look: Online Dating


During dinner, my friends were discussing the popularity of all the online dating services. This piqued my curiosity about the difficulty of finding love. I decided to take a statistical look at the difficulty of finding a soul mate as well as ideas that can help with this age-old challenge.

I took a look at the chances of a 25 year old female professional finding true love. Assuming that she is only interested in men in the Bay Area along with several trait preferences, there are only 116 potential matches out of 7 million people, which is a mere .002%! To improve the odds, one idea is to leverage the existing personal data that is stored in different online services. For example, if someone enjoys watching Game of Thrones on Netflix, we can recommend potential matches that also like Game of Thrones. Although this can be a great way to solve this problem, privacy will definitely be of concern.

Download (PDF, 881KB)