Tag Archives: Google

The data-driven approach is changing the world

Posted on

I was reading this profile of Esther Duflo in the New Yorker last night when I realized that rural development and the tech sector aren’t too different in their approach to problems. In each one you’re trying to convince the customer that what you’re selling – immunizations, software – will make their lives better. Sometimes you’re not reaching out to the customer that well, in which case you try experiments to make things better. In tech, this is pretty low cost – you segment the customer base, show the treatment group a revised page and then measure the differences. In development this is more expensive – you have to create a randomized controlled trial, divide your population in two and then measure the results later.

The basic principles are the same, though. Most successful web designers and aid workers/economists believe the following:

1. We can be doing a better job selling our product
2. Our assumptions could be right or could be wrong, but we have no way to tell without testing them
3. Good data is essential
4. Opinions without data are meaningless
5. Experiment constantly to collect more data and find things that work

The success of this approach has become so clear, and such a part of my approach to problems, over the past few years that it’s hard to imagine doing things differently, or having a discussion with people that rely on their gut. Even having discussions with people is pointless, when you can go look up some papers or evidence that proves or disproves your point, but most people aren’t willing to trust the data over their intuitions. There are two types of people who have opinions about social promotion in schools: people who have looked at the data and people who haven’t. Overwhelmingly the people who look at the data are in favor of social promotion (as are school administrators).

Institutions push back against the data-driven approach, however, because of its implications. Underlying the hierarchical system at most companies is a belief that the people at the top are better at making decisions, or their assumptions are somehow better than the people just starting. If we’re trusting the data, then your guess as the CEO is just as good as mine, as the intern; the people at the top must be comfortable ceding power. Plus we have all these biases, like the ability to discern patterns out of random data and anchoring, and if you’re not familiar with cognitive bias by this point go read this.

In both the tech sector and in development, there are companies that – for the moment – aren’t randomizing, and are still doing okay and collecting money. That’s going to change, though, and soon, because relying on the data gives you such a clear advantage. Google’s made billions with it. J-PAL’s gaining on everyone else in development, only because their experiments are demonstrating results. If you’ve got nothing else to go on, evaluate the future success of an organization by how it makes decisions, and how open it is to change.

CMC Silicon Valley Trip: Thursday

Posted on

Google

  • The trip started off with a very nice young lawyer showing us the campus and telling us about all of the Google perks and quirks. For example, every worker must be within 150 feet of free food, the building numbers start at 42, some guy built a vending machine that displays prices based on how healthy things are, we take lots of awesome trips, here’s a screen that shows what people are searching for in real time, etc. These are traps. The message that everyone should have taken from all of this is, We are fucking good at selling ads. They only afford all of the perks because the people that work there are unbelievably talented, and they don’t want those people worrying about anything besides organizing and archiving all of the world’s information, and otherwise doing really cool shit. All of the perks are like flashy traps. Never forget that Google is really good at making money.
  • Google is also very good at a meta level; they’re not only good at delivering relevant search results and selling ads but they’re also good at being a company. Everything at Google is well thought through, and works well. While you’re on the toilet, you can read a daily 1-page tutorial on good coding practice. The company is constantly re-evaluating what they are doing and the sort of proceses they use. The word several employees used is “It’s a mess around here right now.” It was a similar to the practice of the best teachers in the Atlantic article on great teachers from a few days ago – when evaluators want to come see them, they all say that the evaluator can’t come in right now, because they’re revamping their whole math curriculum or implementing a new module. The theme is constant improvement. Google is trying desperately hard to stay nimble and maintain the ethos of a small startup.
  • I observed that Google employees are very good at getting things done. When something should be done, like the chairs are uncomfortable or the recycling program stinks, Google people are very likely to just do it. Respect.
  • When you assemble the world’s greatest talent in one area and create an amazing culture, you can do unbelievable things. Most of the world’s greatest works of art were produced in two Italian towns in a period of about 100 years, during the Renaissance. The University of Chicago had pretty much every good economist and finance professor in the 70′s and 80′s – Fama, French, Coase, Milton Friedman, Gary Becker, Black and Scholes, Harry Markowitz, Kenneth Arrow and Friedrich Hayek were all there. I would argue there’s a similar concentration of talent in Silicon Valley right now.
  • Job titles are irrelevant. No one at the company does exactly the same thing for very long – people get shifted around within the company, they work on different projects, they learn about new things. The needs of a company change rapidly, so the idea of training people for specific tasks (or even trying to centrally manage an economy) is a little silly.
  • Monetary compensation isn’t as high as at other companies but Google outspends everyone else on perks, and rewards its top talent very well (In the range of baseball player salaries, according to Jonathan Rosenberg). Most people in SV are very focused on work-life balance; short commutes, and doing a job that doesn’t drive you up a wall with boredom or frustration. SV people recognize the importance of quality of life.
  • Jonathan Rosenberg is one of the top people at Google. He compared career hunting to surfing; your goal is to catch a big wave and ride it. Rosenberg continually tried to figure out what the next new thing in technology was, from creating information systems within companies, to helping companies collect outside information, to creating fast Internet connections, and finally search, moving from company to company as the hot new product changed. Larry Page and Sergey Brin convinced him of the money in search when they said, “Search is the moment that the user tells the computer what you’re looking for.”
  • It’s interesting to see how companies divide up the workspace. At Meebo, everyone was out in the open, with three or four people all facing each other in clusters. Same at Atlassian. At Google and EA, most people had three-person cubicles. EMC had tall individual cubicles you couldn’t see over, and corner offices. Google also had these cool yurt things that blocked out sound and had individual temperature control. Pretty cool.
  • Everyone has at least two screens, usually giant ones. Those aren’t cheap; companies must realize that having so many screens helps productivity.
  • The question ramble: In the face of silence, people get nervous. I love when someone asks a question and then starts to ramble on and on, trying to fill the silence. Often they provide a justificat
  • When you’re sitting around a conference table, you need to pick your seat well. The best seats are closest to the speaker, on the sides of the table. The next best seats are at the other end of the table, directly facing the speaker. The next best seat is in front of the speaker, so that if you were facing the table your back would be to the speaker (if chairs are there). The worst is on the sides, away from the speaker. Often the speaker takes the side of the table closest to the door. It’s a tactical mistake to walk all the way into the room and take the furthest seat; you want to take the best seat and leave the stragglers in the worst spots.
  • The opposite strategy applies for sitting in the backseat of a crowded car. The last person in the car never sits in the middle seat. Thus, delay moving towards the car by all possible means.
  • Google Labs are a way of telling employees that there aren’t any rules about which products get chosen; if you have a product you can put it in Labs. To get a product out of Labs, it needs to get used a lot. Google doesn’t care much about profitability for their products, because they make so much from ads. They have the luxury of time that many other startups don’t.
  • Rosenberg told a great story about how an internship he applied for came down to the final two applicants, himself and Mr. Perfect, who was tall and handsome and beat him at everything. Rosenberg hit it off with the employer’s administrative assistant and Mr. Perfect didn’t, and Rosenberg got the job. “If you want to know which first-year bankers are going to make it, ask the assistants which ones they like. The ones they like are the ones who are successful.” The lesson is be nice to the people that set your schedule; they work harder and for lower pay.

Youtube – Steve Grove ’00

  • Steve Grove is a great example of someone who gets stuff done. He proposed to the Kennedy School of Government that he would fly around the world and film their graduates doing all of these amazing things, and they agreed to fund him. So he got to go around the world for free and get some experience filming people.
  • Grove also wrote up YouTube and pitched a politics channel to them. He pretty much built the Youtube politics division from the ground up, and has interviewed many candidates. He took initiative and landed two really cool jobs.
  • It’s pretty cheap to generate video and post it on Youtube. How long until we have 24-hour coverage of a candidate? The Barack Obama Youtube channel, giving you 24 hours of Barack, commentary on Barack, testimony from voters, donor drives, etc.
  • Youtube’s starting to transcribe videos. This allows you to get an idea of the speech, and then read the rest of the transcription. You can also search within the video by keyword.
  • Before the JK Wedding video, Chris Brown’s “Forever” was #270 on the Billboard charts. After the video it shot to #4. Youtube begged UMG not to pull the video, now they share the revenues from ads based on the video. There’s money to be made there, instead of pulling content that you produced.

CMC Silicon Valley Trip: Monday

Posted on

I’m currently in Silicon Valley on a networking trip sponsored by my school’s Information Technology Advisory Board. Each day, we visit two companies, and each night I will post summaries and thoughts. Here’s the recap of Monday’s action.

Microsoft

  • A common misconception about people from Silicon Valley is that they don’t care much about their appearance. That’s crap; most people that I’ve seen care very much about their appearance. They just care about it in different ways than we are used to. Case in point: our host, Scott Mauvais ’90, wore a ponytail to his mid-back, which is probably a very credible signal for non-tech types.
  • Mauvais earned points for opening the floor up to questions right away. For most speakers this is an effective tactic, and we had lots of questions about Microsoft.
  • At the same time, it quickly became clear that while Mauvais was knowledgeable about Microsoft and cared about the company, his area of expertise was limited to what he worked on, which was very much about enterprise software and very little about competing with Apple or putting together Windows Vista. Many answers started off with “I only know what I’ve been reading in the paper.” Students continued to ask detailed questions about other Microsoft departments.
  • Many students also enjoy using questions as signaling. If you hear anyone open a question with “I spent last summer doing X” or “I did a computer science assignment on X,” you can stop listening immediately, because they’ve already shared every bit of information that they care about sharing.
  • Most new Microsoft Stores have clauses in their contract that prohibit the landlord from allowing an Apple Store within a certain perimeter of the Microsoft Store. If an Apple Store moves within that range they have to pay 50% of the Microsoft store’s rent, we were told.
  • Mauvais had a good insight about the Mac vs. PC debates. For Microsoft, it didn’t make much sense to spend a whole lot of money fighting the Apple campaign, because if Apple increases its market share from 6% to 12% Microsoft’s revenues are not hurt very much. However, the campaign allowed Apple to define Windows in a negative light. Microsoft was not able to define Windows in a positive light.
  • Microsoft is trying very hard to move people away from keyboards and mice. I have seen the future, and it is a touch interface. Touch interface is more precise and allows for multi-touch and intuitive gestures. Microsoft had some cool demo touch screen interfaces, and a Windows Surface and they were pretty cool. It’s clear that we are just scratching the surface as to the best ways to interact and operate a touch screen computer. Unfortunately only two companies (Microsoft and Apple) are working on improving this interaction. We will see when Apple’s tablet comes out but I bet it will do very well.
  • Microsoft has bigger fish to fry than personal software; it will continue to lose the public relations debate to Apple, because Apple’s primary focus is on products for personal use. Microsoft has the enterprise market pretty much cornered (and still has an unbelievable edge in desktop computers). It will lose the PR battle but earn lots of money. This was also a constant theme in Mauvais’s responses. “We’re too busy making money,” etc.
  • Microsoft epitomizes the feature creep problem. When you have half a billion users or so, every single one of the features in their products is used by someone, who will be angry when that feature changes or is removed. Fifty percent of “new-feature” requests for Microsoft Office were for features that were already a part of the product.

Meebo

  • Meebo is probably the exact opposite of Microsoft: only 60 employees and the whole company is located on two floors. Our host was Robert Leon ’04, whose appearance was also carefully calculated. Robert pitches Meebo to other companies.
  • In the old days (and by “old days” I mean several years ago), most people found content by punching in queries to Google and entering sites through Google. Hence companies spent a lot of money on search engine optimization. Now, users are increasingly being driven to content through friends, via Facebook or Twitter or RSS. Thus sharing, and tools for sharing are extremely important. According to Meebo no one (fewer than 0.3% of users) clicks on the “Share This” links at the bottom of blog posts or on most web pages.
  • The most important lesson from Meebo was listen to your customers. Meebo started out as a chat client that allowed users on various platforms (AIM, Facebook, MSN Messenger etc) to talk to each other. That’s only a small part of their business now; the most profitable part of their business is that they figured out a way to allow people to share links and videos, really easily. Here is a video demo of the sharing software.
  • So far, the most effective innovation for sharing Web content has been YouTube’s putting the video URL immediately next to the video you are watching. I never thought about that before.
  • Meebo has over 100 million users (“reach,” in the sales community) so they are a valuable source for businesses like, for example, movie studios, who need to push awareness of their movie and have a big opening weekend.
  • According to COO Martin Green the whole web will be using Javascript floating bars at the bottom of the web browser, like Facebook’s, within the next year or so. I’m not sure if there’s a moneymaking opportunity here.
  • Robert made the excellent observation that when you are fresh out of college, you do not know anything. So in job interviews, you need to act extremely interested in the company at hand, and also act like someone who people would enjoy hanging out with. Especially at a startup, it’s important to be able to work well with colleagues, and be enthusiastic enough about the product to put in long hours. Robert pointed out that you don’t learn much in a liberal arts college except “how to think.” I would not even argue that much. Robert was, however, the social chair at CMC. Being a social chair is excellent preparation for a career in sales, and for making people feel comfortable around you, probably much more useful at the margin than trying for an excellent GPA.
  • Meebo subjects all new hires to an extremely extensive interview process. Not only does this show the applicant that Meebo cares but it’s an effective way to vet applicants. They also make every applicant go through a simulation of job tasks, so the sales people have to pitch the product to the hiring committee, or an administrative assistant has to explain what they’d do if the boss’s pager went working. This is excellent practice, as work sample tests are the most effective predictor of whether someone will be good at their job. In a small firm the costs of a bad hire are tremendous; it’s very important to get the position right. In general they like to promote people from within the firm, but if you need a lot of experience quickly they’ll go outside.
  • Meebo had this cool chart of how people communicate – you have Private and Public on one axis, and Real Time and Asynchronous on the other axis. So this is the breakdown:
    • Private, Real Time is SMS and Instant Messenger;
    • Private, Asynchronous is email;
    • Public, Real Time is Twitter;
    • Public, Asynchronous is like Facebook walls.

    We use all of these technologies. Everything in the industry is moving towards Real Time for everything. I handle SMS and email in the same program. The line between Private and Public is strong. But it’s a reminder that how you choose to communicate with someone is as important as the content. Marshall McLuhan lives!

  • “Startups either have customers or they have a business plan. Very few have both.” I don’t know enough about startups to say whether or not that’s true.
  • Static clients like AIM are dead. Everything is moving within other applications like Gmail chat or Facebook chat. That’s why Meebo had to adapt and move into other people’s sites, rather than staying with Meebo.com. Only a small percentage of their traffic is still using Meebo.com.

At dinner I sat next to a very successful executive who sells smart energy and renewable energy products. He also has three houses, one of which is one of about forty properties on a man-made lake in Palm Springs. His wife works at a solar energy company and agreed that lots of the gain people get from generating their own energy is canceled out by increased energy use; it’s not clear whether utility companies are actually substituting out of coal and into renewable energy, or just adding more renewable energy to their ‘portfolio.’ Solar technology is rapidly becoming cheaper.

This is turning into an essay, but I’ll close by saying that I need to work on being less critical. It is way too easy to be critical, especially because a lot of academia demands it; not many teachers ask you to write a complimentary essay. As Mr. Leon pointed out, a lot of your job qualification at this point is just being someone who people enjoy being around. I need to make my business more about making the people around me feel good.

Tomorrow: Electronic Arts, Atlassian and a dinner with entrepreneurs, including the CEO of Scribd. I’ll try to refrain from asking the VP of Marketing when Sim City 5 is going to come out.