Thoughts on Uber’s Surge Pricing

A few weeks ago, the following tweet popped up in my timeline.

thanks @Uber_Mumbai for showing me what #daylightrobbery means!I know I have a choice not to use it! Please RT #psa pic.twitter.com/llyqsggIxT

— Swathi Rishi (@Swathirishi) June 15, 2015

By now, everyone is probably familiar with the mechanics of Uber’s ‘surge pricing’. When the demand for taxis increases, the price of a taxi ride spikes. The company claims that higher prices encourage more drivers to get on the road. It also rations taxi usage by ensuring that only those who really need the taxi immediately and are thus willing to pay higher prices, continue to use it while others wait for the demand and supply to balance each other out.

The economic principles behind surge pricing are widely accepted by economists, which begs the following question – why don’t more companies adopt this pricing model to boost revenue? The answer as Megan Mcardle points out is that raising prices during periods of high demand is usually very unpopular with consumers who feel like they are being exploited. This is reflected in the sentiment behind the tweet at the top of this post. This negatively impacts the reputation of the brand and thus companies will usually forgo the short term boost in revenue in order to prevent the long term loss of reputation. Uber itself has faced tremendous criticism when it has (automatically) raised prices during natural disasters or emergencies.

So why does Uber unabashedly continue to persist with its aggressive surge pricing algorithm? I believe that one of the reasons is that its business model enables it to use it as a competitive advantage. Uber is a platform business, with taxi drivers on one side of the platform and taxi riders on the other. Most businesses fight over customers, but platform businesses have to fight over the suppliers (taxi drivers) as well. As Ben Thompson points out in this brilliant article, the fight over drivers is much more consequential than the one over riders, as having more drivers will enable Uber to lower the average wait time for a rider, which will attract riders to the service. The wait time is of course directly related to the number of available taxis on the road, which explains the mad rush to sign up drivers.

So how do you get drivers to sign up, especially when they have a choice between competing services? Like most other humans, drivers too will choose the company that they think will enable them to earn the most. Now read the tweet at the top of this post one last time, but this time, put yourself in a taxi driver’s shoes. Would you prefer to work for a company that paid you a base rate that remained constant at all times or one that gets raised aggressively during periods of high demand? The choice should be apparent, which explains why Ola was forced to abandon its initial stable pricing model and introduce peak pricing earlier this year.

Advertising revenue on LinkedIn and other ‘meet’ platforms

Earlier this month, the professional networking site LinkedIn, released an impressive earnings report for the fourth quarter of 2014. Revenue for the quarter was $643 million, an increase of 44% from the fourth quarter of 2013. It wasn’t just the top line numbers that impressed, all segments of the business grew at a healthy pace.

• Talent Solutions – Revenue earned from recruiters who pay a fee to access the database of professionals on the site. This accounts for 57% of LinkedIn’s revenue.
• Marketing Solutions – Revenue earned from the advertising that is on the site. This accounts for 24% of total revenue.
• Premium Subscriptions – Revenue earned from premium subscriptions. Accounts for 19% of total revenue.

This post is going to focus on the second revenue stream i.e. Marketing Solutions. But first, a little background information is required. In my social networks class, Profession Piskorski divided social networks into either ‘friend’ platforms or ‘meet’ platforms. Friend platforms such as FaceBook help users develop closer relationships with their existing network. Meet platforms such as LinkedIn or dating sites help users broaden their networks by introducing them to new people.

From a business model perspective, there is one key difference between the two. If a friend platform is successful, you spend more time on it i.e if everyone you know is constantly on FaceBook updating their status or uploading pictures or messaging, the more interesting content there is for you to see hence and the more likely you are to visit the site. On ‘meet’ platforms, the more successful the platform is, the less likely you are to visit. For example, if you are successful in meeting someone on a dating site, you stop using the site. For this reason, ‘meet’ platforms typically derive their revenue from subscriptions whereas ‘friend’ platforms tend to rely on advertising.

This is a challenge for sites such as LinkedIn that would like to get additional revenue from advertising. However, LinkedIn was initially seen as a place that you went only when you were looking for a job. Thus even though many users created profiles on the site, few visited regularly. This brilliant article details how LinkedIn addressed this issue by essentially turning itself into from a social network into a media platform. The entire article is worth reading but the key passage is the one below.

To get things started, the company launched LinkedIn Today in 2011 as a way of enticing professionals to log in every morning to catch up on industry news. The following year it signed up “influential thought leaders” to provide occasional commentary. Last year, the company bought Pulse, an app that focuses purely on content, and which the company would like professionals to check for a quick update on news relevant to them. And now, it runs a content behemoth that drives the internet’s manicured hordes to business publications around the web.

The strategy certainly seems to have paid off with advertising accounting for 24% of revenue in the most recent quarter as opposed to 22% a year ago. However, the company is not resting on its laurels. It plans to scale its B2B advertising revenue to one billion dollars by 2017 having acquired B2B marketing platform Bizio for 175 million USD last year. The company also recently launched its own ad network to allow companies to identify which LinkedIn user is visiting their site. If successful, this strategy could act as a template for other ‘meet’ platforms to make a play for a share of the lucrative online advertising pie.

Book Summary – Influence by Robert Cialdani

Influence by Robert Cialdini is widely considered to be a classic on the subject of human psychology. Released in 1984, the book has sold over two million copies worldwide. The book was initially written as a means of explaining the psychology behind common sales tactics, but is relevant to a far wider audience. I found the book enjoyable to read and have no hesitation in recommending it to anyone with an interest in human behavior.

The book focusses on the following six principles of influence.

1) Reciprocation

Humans are hardwired to reciprocate favors, gifts etc. Hence when someone offers us a free sample we may feel obliged to buy something even if that was not part of the original plan. For instance, companies will often send you sample offers or try before you buy schemes to exploit this tendency.

Failure to acknowledge this principle can result in bad design. For example, I sometimes encounter software products which demand that I register before I can use them and before they have proven their value. In some cases it is unfeasible to provide the service before I have registered, but in many cases it is not and in those cases, I am left with a negative impression of the product and often look for alternatives.

2) Commitment and Consistency

Once you state something in public or put it into writing i.e once you commit, it is hard to deviate from it as you feel the pressure to be consistent about it. This is the psychology behind bait and switch deals i.e. once you have committed to the idea of making the purchase it is hard to switch even after the terms of the sale change.

You can also use this tendency in a positive way. For instance, write down your goals and principles to strengthen your commitment to them. Written commitments are stronger than verbal commitments as it takes more effort to make a written commitment. In general, the greater the effort to make the commitment, the stronger the effect it has on the person making it.

3) Social Proof

One means that people use to determine what it correct is to to find out what other people think is correct. For example we use social proof to determine at what speed to drive on a highway or what to do with an empty popcorn box in a theater. Bartenders salt their tip jars with a few dollar bills to simulate tips left by prior customers. Companies use testimonials and phrases such as “best selling” or “fastest selling” to convey the impression that everyone else likes the product.

Companies prominently display the number of FaceBook ‘Likes’  that they have for this reason as well. Whenever I encounter a company that I am unfamiliar with, I usually feel more confident about their quality or service if several of my FaceBook friends have ‘liked’ them.

4) Liking

People tend to respond more favorably to requests from people who they like. E.g. Tupperware parties, Campaign volunteering by locals vs non locals. Factors that influence likability include physical attractiveness, similarity (personality mirroring), compliments (People like getting them) and Contact and Cooperation (People like those that they come into contact with frequently, but only in a cooperative setting and not a competitive one).

5) Authority

People respond to requests from authority figures. The authority figure does not even have to be real. E.g. Actors playing doctors in drug commercials. Other symbols of authority such as titles, and uniforms are also effective in generating responses from people.

6) Scarcity

Scarcity increases the attractiveness of a product. E.g. Limited engagement, limited stock etc etc. When the scarcity is a result of direct competition i.e. others are buying the product and hence there may be few left, the attractiveness is increased even more.

Thoughts on FaceBook and WhatsApp

Last week, it was revealed that WhatsApp had brought in revenues of around $10 million USD in the first half of 2014. This begs the question – why would Facebook spend $18B on a company that is on track to earn only $20 million dollars this year? At its current value of over $200B, FaceBook is currently valued at around 17 times its annual revenues, a valuation that would imply that WhatsApp is worth approximately $340 million dollars. When companies make acquisitions, they typically have to pay a premium of around 20% over the market value (40% if the acquirer is a private equity firm). So what makes WhatsApp so valuable that FaceBook was willing to pay such a huge premium?

Historically, companies have made acquisitions for one or more of the following reasons

• To exploit ‘synergies’ i.e one company has a good product, the other has great distribution, hence joining forces results in a situation that is beneficial to both
• To take advantage of tax benefits i.e the rumored deal between Yahoo and AOL
• To diversify and achieve seasonal balance i.e. a company that makes snowmobiles might acquire a company that makes beachwear to ensure that revenues are stable throughout the year
• Defensive – to prevent competitors from acquiring the company
• To expand their business by gaining access to new markets, customers etc.

With WhatsApp, it’s likely that the last two factors were the most important. The last thing that FaceBook would have wanted is for a competitor such as Google to get its hands on WhatsApp’s users. Furthermore, it has been noted that many of WhatsApp’s users were in countries such as India (It was the leading messaging service in India when it was acquired) and maybe it was really a means of expanding FaceBook’s user base. Or perhaps it was a means to reach the fickle teenager segment of the market which is not taking to FaceBook in the same way as other age groups have. These are all good reasons, however it is likely that there were additional factors that were in play.

In my Social Networks class, Professor Mikolaj Piskorski made the point that the frequency of posting on FaceBook decreases as the network size increases. This makes perfect sense if you think about it. Imagine that you are in a room with a few close friends. It’s likely that you will feel comfortable discussing nearly any issue with them. Now imagine that your family (parents, siblings) joins in. The conversation is likely to be slightly more guarded. Next imagine your co-workers and your manager join in. You now have to think very carefully about what you say as everyone can hear you.

I started using Facebook in 2004. At that time, it was only accessible to those in the United States with university email (.edu) accounts. Privacy controls did not exist (everyone in the University had full access to your profile) and yet, users shared content freely as they knew that the only ones viewing what you had posted were other students. As FaceBook opened up access to everyone, people started to become more cautious about what they posted. Photos that showed students drinking and partying were hastily removed as parents and potential employers were granted access to the site and users became much more careful about what they posted.

This is a huge problem for a company such as FaceBook. People visit Facebook several times a day because they expect to see updates from their “friends”. If people are posting less and the updates are less frequent, the time spent on the site will surely drop. One can still post status updates and photos and limit them to specific friends, but this process is unintuitive and tedious and there is always the danger that a friend will share a post with someone that it is not intended for. Furthermore, The less content you post on Facebook, the harder it becomes for the company to serve well targeted ads to you (FaceBook also tracks your behavior on other websites through cookies, the ‘Like’ button, and FaceBook Connect. This data is then used in order to show you relevant ads when you return to FaceBook).

It is likely that the acquisitions of WhatsApp and Instagram were aimed at alleviating this problem. In smaller groups, people are much more likely to revert back to the behaviors of the early FaceBook users. This is certainly true for me – I post much more frequently on WhatsApp than I do on FaceBook. Recently FaceBook relaunched Atlas, which is an advertising platform that will enable it to use the information it has on users to show them relevant ads on third party (non FaceBook) sites. Though the company currently does not scan and use the information from your WhatsApp messages to show you ads, that could very well change in the future and may represent a potential source of revenue that would justify the huge price paid for WhatsApp.

Book Summary – The Design of Everyday Things by Donald Norman

When I was studying for a Master’s degree in Computer Science, I took a course on user interface design. One of the assigned readings was a book by Donald Norman who is a noted academic in the field of cognitive design, and usability engineering. The book “The Design of Everyday Things” is a bestseller that remains hugely influential in the world of design and it had a big influence on me as well.

The book was written originally published in 1988, well before the Internet became mainstream, but it’s amazing how relevant it still is. Donald Norman has a website, which uses the design principles in the book to make recommendations on usability for websites and mobile apps (See here and here for some good examples). In preparation for my transition into the role of product manager, I recently re-read the book and thought it might be useful to summarise the key ideas from the book.

Affordances and Signifiers

I have grouped the first two concepts together as they go hand in hand. According to the book “Affordances represent the possibilities in the world for how a user agent interacts with something.” For example a cup has a handle that is big enough for the human hand, it therefore affords gripping. However not all affordances are easily identifiable. A signifier is something that signals the presence of an affordance. For example on a clear glass door without a handle, it may not be apparent which end of the door is to be pushed. Therefore the label “PUSH” is a signifier that can be added to signals which end of the door is to be pushed. Signifiers are important as they indicate how the system is meant to be used.

Mappings

Mappings specify the relationship between two sets of elements e.g the relationship between a set of light switches and the actual lights. Some mappings feel natural (e.g. turning the steering wheel left on a car to go left) while others do not (e.g Light switches where it is impossible to figure out which switch controls which light without resorting to trial and error). Good design makes use of mapping that is intuitive and feels natural. This may be dependant on culture and prior conventions. A few years ago, Apple tried changing the default setting on its trackpads so that moving your fingers down resulted in the page being scrolled up rather than down. This change in mapping was frustrating to many users who had grown used to the opposite mapping and many changed the setting to revert back to the original mapping. However, it is quite possible that future generations who use this new setting from day one, find it intuitive as it mimics the behaviour of a touchscreen.

Constraints

Constraints limit the number of actions that can be performed on the system. Good design uses constraints guide the user to the appropriate action. E.g. the corner pieces on Jigsaw puzzles where its obvious that the piece will only fit in the corner or the web application that only forces you to select from appropriate options by disabling the invalid ones.

Feedback

Feedback refers to the way that the user knows that the system has received their request and is currently processing it. Good designs feature immediate feedback – delays are disconcerting to the user. For e.g a user typing on a computer that has a small lag between the pressing a key and the display of the character on the screen will often press the key twice and make other errors.

Conceptual Models

Also referred to as mental models, this refers to our understanding of how a system works. For example, the file system in our computers is visually represented by icons of folders, sub-folders and files. In reality, this is not an accurate representation of how files are actually stored in computers. However, this mental model of folders is very effective in helping us gain an understanding of how we should organise our files and folders. Good designs use conceptual models that help the user understand how to use the product.

Final Thoughts

UI/UX design is a key part of a product manager’s responsibilities. A badly designed interface will frustrate users and ruin an otherwise good product and hence it’s important to focus on this as well. Design can also create a competitive advantage for a product as Apple has demonstrated over the last decade. Creating good design is not merely about aesthetics, it’s also about understanding how people think (This book was originally called “The Psychology of Everyday Things”). In my opinion, this book is a must read for a product manager and is a great introduction into user centric design. If you prefer a more hands on learning approach, the author also teaches the material from the book in a course on Udacity. You can view several of his talks and seminars on YouTube.

What I Read

This post was inspired by the “What I Read” columns on The Wire, which asks public figures about their daily reading habits.

A few years ago, I read this blog post by Robert Bruner and realised that I was spending a lot of time reading on the internet, but I did not feel that I was learning much. I then spent a lot of time seeking out high quality web sites and bloggers. There is a lot of information online and I think the best way is to find one website/blogger that fits into this category and then figure out where they get their information from. Twitter is great for this purpose as you can follow these individuals/sites and look at what they follow or retweet.

As soon as I’m awake I go to my RSS feed on feedly which I switched to after the demise of Google Reader. I’m also on Twitter all day and it serves as my second RSS feed. If there is an popular article on a topic that I am in interested in, it invariably is retweeted by a person that I follow on Twitter and I will read it or save it to Pocket to read later. I do not visit news sites such as The New York Times or the Economist as I usually find their best articles through Twitter or one of the bloggers that I follow.

Economics/Business:

Marginal Revolution – Marginal revolution is a blog run by Tyler Cowen, a professor of economics at George Mason University. He is a prolific blogger (several posts a day) and covers an impressive range of topics from politics, business, economics to chess and food. This is my favourite website and a majority of the articles that I read are usually referred to from this site.

Megan Mcardle – Blog on business and politics from a libertarian/conservative view.

Matt Yglesias – Blog on business and politics from a liberal view.

Robert Bruner – Robert Bruner is the former Dean of the Darden School of Business at the University of Virginia. Does not post very frequently but the posts are usually of a very high quality.

Greg Mankiw – Former Bush economic advisor and currently the economics professor at Harvard. This blog is now fairly inactive as he has taken on a more senior role at Harvard.

The FT – I started reading the FT about a year ago, when I was granted free access to it as an Oxford MBA student and will continue to read it until my access runs out.

CNBC – Stock market and other business news.

Money Control – The Indian version of CNBC; it’s where I go to get news on the Indian stock market.

Rediff – Indian news site. Probably not the best one out there, but I have stuck to it largely out of habit.

Politics:

RealClearPolitics – An American political news aggregator site. Also aggregates the latest polls and political news videos. A good resource on election nights as well.

Hotair.com – Popular conservative news blog.

Vox.com – Recently launched liberal news and opinion site.

Technology:

Engadget – Technology and Gadget news and reviews.

The Verge – Similar to Engadget.

TechCrunch – Similar to The Verge and Engadget, but a lot more comprehensive. Also has a lot of information about the latest technology startups and apps.

Anandtech – Similar to the first two, but the reviews tend to be more detailed and higher in quality. The site’s founder recently departed to go work for Apple.

Macrumors – Latest Apple news and rumors.

Sports:

ESPN – I follow ESPN during the NFL (American football) season. I also follow college football on here.

Walter FootBall – Serious and often humorous take on the NFL (American football). It’s also where I follow the latest mock drafts.

Yahoo Sports UK – This is where I get my football (soccer) news from.

Preparing for a Product Manager Interview

So you’ve worked hard on preparing the product manager application and you’ve managed to snag an interview invitation. You still have a lot of work ahead of you as PM interviews require a fair bit of preparation (especially if it’s your first PM job). If you have not already, I would urge you to read Ken Norton’s (ex Google PM) excellent advice on hiring PMs.

Every company is likely to have a different approach to interviewing and your best bet is to look at the company’s Glassdoor and Quora pages for interview information. This tends to be easier for larger companies and much harder for the smaller ones. In general, the questions that you will encounter can be roughly classified under the following categories.

Estimation Questions

• How many aeroplanes are in the sky right now?
• Estimate the annual revenue of Google in China?

These may seem like trick questions but they are meant to test your ability to quickly estimate the size of a market which is a useful skill to uncover new market opportunities or to test the viability of new features. These are also called Fermi problems, named after the physicist Enrico Fermi who was famous for making accurate estimations with very little data. The test is more about the method you use rather than the final number that you come up with. This type of question is very popular in the consulting interview and a google search will give you all the information that you need. For a software PM role, you should familiarize yourself with some of the common statistics from the software industry.

Smartphone ownership – 20% of the world, 60% of USA adults (45% overall)
Cell phone ownership – 90% of US adults (70% overall)
Desktop – 55% US
Laptop – 60% US
Tablet – 6% of the world
Facebook Users: 1.25B
Facebook Annual Revenue: 9B
Twitter Users: 255M
Twitter Revenue: 1B
Google Annual Revenue: 65B
Gmail Users: 500M
Instagram users – 200M
WhatsApp – 500 million users
Evernote – 100M users (5% of whom are paying customers)

Business Case Questions

• How would you react if a competitor entered your market?
• Should Google charge for gmail access?
• How much should Amazon charge for the Kindle HD?

These are some examples of the business case question. Popular books such as Case in Point and Case Interview secrets walk you through the entire process. Note that these books are geared towards the consulting case interview questions which though similar are slightly different from the PM case interview questions. The difference is that in the consulting interview, as you delve into the problem, you will be given additional data and asked to figure out the problem, which will test your quantitative and analytical skills. In the PM interview, it’s more about the approach that you adopt to crack the case. I used the first book and found it quite useful and there are plenty of practice cases that you can find online. Victor Cheng who is the author of “Case Interview Secrets” has an set of free video tutorials that were quite useful to me.

The Product Question

• How would you design an alarm clock for the blind?
• What is your favourite product and why?
• How would you improve Google Maps?

The product question is a perennial favourite of interviewers and is designed to test out one of the core requirements of a PM i.e a good product sense. If you are interviewing with a company that makes a product for consumers, expect to get questions on its design and how it can be improved. Surprisingly, I had a lot of trouble finding good resources for this question online. There were a few blog posts that I found useful, but most of the resources online were too brief. I eventually came across this framework on Jackie Bavaro’s quora blog that persuaded me to buy the book that she co-authored – Cracking the PM Interview – Gayle McDowell and Jackie Bavaro. This book is more than just interview prep, it introduces you to the PM role and is a great way to get up to speed if you do not have to time to read through all the resources that I linked to in my previous posts (See here and here). She covers all the different categories of questions that you are likely to encounter and my only complaint with the book is that I wished that there were more cases available for practise. Hence, I also ended up purchasing Decode and Conquer – Lewis Lin mainly to have a few more cases to practice. This book focusses more on the interview prep and is thus an excellent complement to the first book. I don’t really see them as competitors, and I recommend buying them both.

Technical/Programming Questions

You are probably not going to be asked to implement an algorithm to detect loops in a linked list on a whiteboard but you do want to research the company you are interviewing with to be sure. Companies such as Google and Facebook do have a programming round, but apparently this has been relaxed recently. You probably should be comfortable answering high level questions about the architecture of the applications that you worked on. If you have a programming background, you should be comfortable answering questions about the projects that you were involved with. It’s worth taking a day or so and trying to remember the details of your projects (tools used, problems faced etc) and writing them down. Some interviewers will drill down into the minute details and you will want to be prepared. Amazon is famous for this type of approach, where the entire interview is about a single accomplishment on your resume.

Behavioral Questions

• Walk me through your resume.
• Why do you want to be a PM here?
• What is your biggest weakness?

You should expect all the usual behavioral questions. It’s easy to underestimate this portion of the interview and most candidates tend to focus their preparation on the other portions of the interview, but most interviewers want to see a cultural fit. Hence, you should have prepared answers for all these questions and they should be tweaked to fit the culture of the company that you are interviewing at. Again, there is no lack of information online on how to tackle these questions. In general, the STAR framework (Situation, Task, Action, Result) tends to work well.

Final Words

As always the key is doing well is to practice as much as possible. If you are unsuccessful during the interview, try hard to get some actionable feedback from the company which you can then use to improve for future interviews. One of my complaints about the interview process is that you can get really good at interviewing by practicing and I am skeptical that this is a accurate predictor of your on the job performance. When I was interviewing for programming jobs, I was terrible during my first interview but improved steadily afterwards and finally ended up getting a job from the fourth company that I interviewed with. I am fairly certain that had I interviewed in the reverse order, I would have worked for Qualcomm rather than Riverbed Technology.