Let’s look at Box’s tech stack and stats (at one point in time)

• Mostly PHP (1M+ lines), some Ruby, some Scala
• Mostly MySQL, some NoSQL (HBase)
• MySQL sharded based on user ID and files and folders
• Over 1B+ queries a day
• Millions of users
• Tens of millions of folders
• Hundreds of millions of files

The talk that I most enjoyed was given by the Director of Engineering, Kimber Lockhart, on ways to become an engineering leader.  She offered some tips that anyone could take away whether engineering or not.

My favorite takeaways

• Set a target or goal no matter what.  Kimber puts pictures of her 2012 goals in frames in her living room.  Her boyfriend puts his goals on his desktop image background.  It’s one thing to have goals and another to have it in your face everyday.  I have to try this.
• Read the top 20 books in your field.  I completely agree.  I’d like to extend this and say either read books or do hands-on learning like coding or writing business cases for your field and/or job type.
• Identify blockers and attack.  If there’s something that you really want, find out why you aren’t getting it.  In my opinion, no one is going to hand you things just because.
• Do the work of the job that you want.  Key piece of information, don’t forget to outperform at your real job too.  I tell others that doing stuff outside of your core job is just icing on the cake.  No matter how good the icing is, the cake still has to taste good.  Dominate your core job while building skills for the job that you want.

Alright.  So, I generally like hearing companies talk about mistakes that they’ve made in the past when it comes to mobile, starting up, launching, etc.  It helps me make a mental note of what not to do.  Spencer Rascoff, CEO of Zillow, a real estate search engine, took the stage at Mobilize to talk about “Getting to #1 through Mobility”.  What I like about Zillow is that I can take any address and find out what the Zillow market rate is.  Also, the mobile app is great especially when I’m in a desired neighborhood and quickly find out that the homes are out of my price range.  Thanks Zillow for reminding me that I’m dirt poor in the Valley (in comparison to other Valley ballers).

Mobile is the “it” topic at every company.  I keep hearing time and time again, we must think of mobile, we must think of mobile.  Yes, I agree, but how are we going to think about mobile differently from others.  Mobile doesn’t equate to just designing for a smaller screen.  It needs to be injected in all aspects of a company.  At Zillow, they think about mobile in a way that often gets overlooked by others.  For example, all the projectors in the meeting rooms have mobile dongles.  I’ve never seen that before.  Imagine whipping out a phone and showing what the native app or mobile web looks like on the projector.  At my work, I’m happy to find Mac dongles in every conference room.  Baby steps for me.  Another thing that Zillow does is make an effort to have their mobile-friendly emails sent to their users.  I can’t tell you how annoyed I get when I receive email from Yahoo! groups and LinkedIn on my iPhone.  I either can’t read the text because it was too small or the indentation is so bad that there are skinny columns of text.

With a lot of mobile to do, how does Zillow do it all?  When it comes to talent, Zillow realized that it was difficult to find great mobile developers.  So they shifted recruitment to finding great developers who can learn mobile development.  Nice.

Mobile monetization is on every exec’s mind.  People spend a lot of time using their phones so how can companies cash in.  Zillow makes money by connecting users to realtors (e.g., one click call to real estate agent) while providing realtors a subscription advertising platform and premiere real estate agent tools.  However, other companies struggle to find the right balance of good monetization and good user experience.  Companies like Twitter and Facebook, who have locked in users, will provide advertisers different ad products, but they may or may not deliver results depending on receptiveness of the users.  Eventually, advertisers may just take their money elsewhere.

Needless to say, Zillow has a solid business model and I’m excited to see what more they will deliver on mobile.

Zill Stats

• started in 2006
• 600 employees
• 1/2 usage is on mobile
• 30M users per month
• 13 native mobile apps – iPhone, Android, Blackberry, Kindle, Windows, etc
• Users are 3 times more likely to contact a real estate agent than using other sites

During a break, InterDigital sponsored a 7-man panel on “Re-imaging Wireless”.  I was most interested in the discussion around the convergence of enterprise and consumer mobile market.  In case you didn’t know, most often the money is with the enterprise market.  It’s a more solid stream of revenue than consumers although (IMHO) the consumer market is sexier and easier for grandma to understand if you hit it big.  Nowadays, consumers expect something for nothing.  Thanks to Apple for setting the bar at free or $0.99.  For every paid mobile app downloaded, about 300 free apps are downloaded.  This is why I recommend mobile app developers to consult rather than develop their own mobile apps unless developing apps is for fun.

The general consensus of the panel was that the difference between enterprise and consumer mobile apps is becoming less and less.  Enterprise requirements include privacy, security, data-oriented, and easy for IT to manage.  For consumer apps, it’s mainly social, fun, entertainment, but will soon (if not already) address many of the enterprise requirements.  One day there’s going to be a merge of the two markets.  Imagine using your iPhone to view spreadsheets with animated bright, bouncing numbers getting eaten by monsters.  I love it!

Now most companies are looking 18 to 24 months out for their mobile app needs.  In reality, these companies don’t really know what about mobile that they need to do, they just know that they need to do something mobile.

One panelist said that his consulting company is deploying 20-30K iPads to large enterprises.  He declared that companies will not continue to pay $600 to $700 a tablet (hint: this is directed to Apple) and that someone will come out with a tablet (made in China) with a $100 price tag.  If the tablet gets lost, the company can just buy a new one.  Here!  Here!  Where can I buy this tablet?  Samsung? 

Stay tuned for follow-up postings about Mobilize.  Enjoy!

My top 5 take aways:

1. Ending your opinions with questions may make you sound not confident.  Action: Don’t end with questions unless you really want other opinions.
2. Being collaborative could make you look indecisive.  Action: Set decision criteria and timelines for when decision will be made.
3. Not enough self-promotion.  Action: If someone asks how you are doing, talk about recent achievements like “oh I’m great because the team just hit this milestone…”.  Take opportunities to promote in everyday conversations.
4. Using “I feel…”, nobody cares how you feel.  Action: People care about what you know.
5. Don’t formally ask people to be your mentors because it requires time and commitment.  Action:  Just adopt mentors informally and don’t tell them.  What they don’t know can’t hurt them.

One thing that I’d like to add is to say no to meaningless (or not as meaningful) tasks.  I don’t know how many times I’ve seen women be asked to organize team events or lead up some random task force while others are constantly are passed over or say no.  I’m all for being a team player, but not at the expense of my valuable time and maybe my work reputation (Who wants to be known as the potluck queen?).  If you keep saying yes, then you’ll keep getting asked.  Say no already!  It’s okay.

1. Conduct meeting without tables and chairs aka standing up – studies show meeting times will be cut in half!
2. Money jar for late comers or no-shows – make it hurt (financially) to be late or miss a meeting, use that money to buy the team beer or donate to charity
3. Take a moment of silence after a major decision – reflect on decision, give time for last minute objections, let the commitment sink in really good
4. Put a clock on the wall – meetings should begin and definitely end on time
5. Assign a note taker – the meeting chair needs to run the meeting while someone else takes notes, hard to do both
6. Divide up a large block of meeting time (with same people) into smaller chunks – helps guide the agenda and keeps the meeting going
7. Enforce a “no meeting” time period or day – seriously, there needs to be time to do work, come on people!
8. Provide food – attendance will be good, also many will pay more attention (while eating) instead of typing on their laptops
9. Clarify participant roles in invite – sometimes it’s obvious, sometimes it’s not, provide some context, maybe it’ll help form the agenda easier
10. Bonus: If the meeting requires 6+ people or more than an hour, the CEO is notified – one company did this, the theory is that a meeting of that size or time length should really damn important, I wonder if the company is still around, does anyone know?

Hadoop is open source software for distributed computing. It consists of two main parts – Hadoop distributed file system, a distributed file system with high throughput, and Hadoop MapReduce, a software framework for processing large data sets. And then there’s HBase which is a columnar key value store on top of Hadoop.

To handle large amounts of data (e.g., BigData), many companies are using a front end cache (e.g., memcached) or having their entire database in memory. In addition, companies partition the data across many nodes for query parallelism and mitigation of resource contention. Borthakur said that with every 1K node cluster, they get about 2-3 bad nodes everyday.

For Facebook messages the requirements were:

• Massive data set, large subsets of cold data
• Elasticity and high availability
• Strong consistency with data center
• Fault isolation, quick recovery

They wanted a database that could store at least a terabyte of data. So they looked Cassandra, HBase, and sharded MySQL. After some testing, Facebook picked HBase because it gave excellent write performance and good reads. Also, Borthakur commented that partitions of different nodes (e.g, Cassandra) in a single data center are very rare.

Some Facebook messages stats:

• 6K messages a second
• 50K chat messages a second
• 300TB linear data growth per month compressed, compression is 1 to 4
• Most data is chat messages

For Facebook, HBase is good for counters like finding out how many people in certain age groups are logged on or finding how many people clicked “Like” on a particular brand. They have more than 1 million counters using 100 node HBase with reads mostly served from memory. Currently, they do not use tiered storage for messages but maybe looking at it for photos with Hadoop.

Frequency of Facebook items accessed (pre-timeline):

• Photos and videos get a lot of reads for first few days then after 60 days not many
• Last 30 days of comments are in the cache
• Most people interested in looking at last 10 days of messages, all 10 days worth are already loaded when you log in

As for the future of Hadoop at Facebook, Borthakur is looking at possibly migrating all user data and graph data from MySQL to HBase.

One thing to note, if you need a database with more reads than writes, HBase is probably not the database for you.

From Hadoop website:

The Apache Hadoop software library is a framework that allows for the distributed processing of large data sets across clusters of computers using a simple programming model. It is designed to scale up from single servers to thousands of machines, each offering local computation and storage. Rather than rely on hardware to deliver high-avaiability, the library itself is designed to detect and handle failures at the application layer, so delivering a highly-availabile service on top of a cluster of computers, each of which may be prone to failures.

About Dhruba Borthakur:

Dhruba Borthakur is the Project Lead for the Open Source Apache Hadoop Distributed File System. He has been associated with Hadoop almost since its inception while working for Yahoo. He currently works for Facebook and is instrumental in scaling Facebook’s hadoop cluster to multiples of petabytes. Dhruba also is a contributor to the open source Apache HBase project.

Earlier, he was a Senior Lead Engineer at Veritas Software (now acquired By Symantec) and was responsible for the design and development of software for the Veritas San File System. He was the Team Lead for developing the Mendocino Continuous Data Protection Software Appliance at a startup named Mendocino Software. Prior to Mendocino Software, he was the Chief Architect at Oreceipt.com, an e-commerce startup based in Sunnyvale, California. Earlier, he was a Senior Engineer at IBM-Transarc Labs where he was responsible for the development of Andrew File System (AFS).

Dhruba has an M.S. in Computer Science from the University of Wisconsin, Madison and a
B.S. in Computer Science from the Birla Institute of Technology and Science (BITS), Pilani, India.
He has 17 issued patents. He hosts a Hadoop blog at http://hadoopblog.blogspot.com/ and can be subscribed via his profile page at https://www.facebook.com/dhruba

Read more about Hadoop at Facebook:

• Realtime Hadoop usage at Facebook: The Complete Story
• Realtime Hadoop usage at Facebook: Part 1
• Realtime Hadoop usage at Facebook: Part 2

First up is LivingSocial. Overall, I like LivingSocial’s app more than Groupon’s app. I really like the way LivingSocial presents the deal details page (third screen). It is very clean with a lot of white space around the text. The price is big with the text color orange. I’ve always liked the color orange as a standout color (see my app Eventabulous). The best design element of the details page is the “buy now!” button. The button remains as a fixed footer when you scroll up and down. Also, the button is at the bottom where I can just use my thumb to tap the button. Lastly, I like that there are headings for each deal, short and sweet.

I do not like that the first page (first screen) doesn’t tell me any information about today’s deals. It’s basically a menu of items which I understand because it saves screen space on later screens. But I have to do a button click in order to see anything about deals when I first bring up the app. Don’t make me work to buy things. On the second screen, it’s not obvious to me that the city name “san jose” can be clicked on to switch to other cities. It can be easy to miss the small triangle on the right hand side indicating a drop-down box.

I recommend that LivingSocial assumes that users want to see “today’s deals” first and show “today’s deals” as the first screen. If users want “instant”, “escapes”, etc, they can hit the “home” back button (see second screen) to see the menu. Also, Living Social should consider having a “buy now!” button on the list of “today’s deals”. Why does a user need to see the details to buy the deal?

Next up is Groupon. You know that Groupon is great at selling because you can’t ignore the “Buy!” button everywhere. On the home page (first screen), it goes straight into the list of daily deals and from right there you can buy the first one only. I do like that Groupon attempts to highlight the savings of the deal more prominently than Living Social, but then again I assume all of the Groupon and Living Social deals are the best exclusive deals, so I don’t bother too much about the percentage saved. Maybe I should be more cautious about buying deals hahaha. I really like seeing discussions of the deals where people ask questions to the company. I didn’t see the same discussion feature in LivingSocial.

Unfortunately, Groupon has too much information in a small amount of space. It hurts my eyes looking at it. I can’t tell if it’s the amount of information or the size of the font. Another thing, I have to use my index finger to hit the “Buy!” button. Groupon has a “Buy!” button as a fixed header instead of as a fixed footer. I like one-handed navigation of apps. Lastly, I like the thought of the menu tabs on the details page (second screen), but on an iPhone, I am ok with scrolling to see all the information. Sometimes tapping disrupts the flow of information.

I suggest that Groupon leverage color as a way to highlight parts of a deal and they need to create more white space around their text. Less clutter is better.

Let’s take an example of an e-commerce website wanting to increase shopping cart conversions (e.g., completed purchases).

One of the first things to look at is where customers are entering and leaving the website.  This will give you a general idea of the day-to-day activity on the website.  In GA, you can check out Top Landing Pages (under Content), Top Exit Pages (under Content), and Bounce Rates (under Visitors->Visitors Trending->Bounce Rate).

What is a bounce rate?  From Google:

Bounce rate is the percentage of single-page visits or visits in which the person left your site from the entrance (landing) page. Use this metric to measure visit quality – a high bounce rate generally indicates that site entrance pages aren’t relevant to your visitors.

For pages with high bounce rates, there could be one of two things happening, 1) the pages plain out do not have what the user wants or 2) the pages have poor usability.

It’s important to pay attention to the entrance keywords and sources for these pages with high bounce rates (under Content->Overview look at right lower side of page, click on Entrance Keywords or Entrance Sources, then select a specific page by using drop-down box next to the word “Content”).  Search keywords should align with the purpose of your site.  If they do not, stop bidding on these keywords (if you are stupid enough hahaha).   Entrance traffic sources related to campaigns (e.g., Facebook ads, email newsletters, etc) should help yield high quality leads (e.g., low bounce rates, not high bounce rates).  If not, re-evaluate the effectiveness of those campaigns.  If entrance traffic sources are not campaign-related, investigate the reason users come to your site from these other sites.  Are there any opportunities to capitalize on this free source of traffic?

A high bounce rate could also mean that your site does not engage users in a positive way.  Is it hard to find products?  Maybe prices are too high or lack of customer reviews.  It’s always good to re-evaluate pages often for usability.  For instance, there are tools available that can generate heat maps and the typical eye path around a page.  A user study could also help identify major problems.

Next, it’s good to look at where customers are dropping off at each step of the buying/checkout process.  In GA, I can set up a goal called “Checkout” with defined funnels, one funnel for each step of checkout process.  To do this, go to the overview page of all of your sites (click on GA logo if you are lost), then click on “edit” on the far right column next to the site that you want to add goals.  See here for more info about how to add goals.

Let’s pretend that this one e-commerce website has 7 steps for checkout.  Below is a table with an example conversion funnel with fake data.  In this example, the first major drop-off is step #5 then at step # 7.  This is your big clue to investigate the reasons for those significant drop-offs.

Table: Fake funnel conversion data

I hope that this helps you get interested in learning more about Google Analytics for your site.  Happy conversions!

Facebook

What is it?  An online social network where you can add people as “friends” (must be mutually accepted by friend) and post information for others to see.  Your main page is a running feed of your friends’ most recent status updates, pics, links, and videos.
Pro:  Stay connected with people you don’t see often, re-connect with long lost friends, easy to share info about yourself to your “friends”
Con: Not so private if you have many friends, sometimes awkward to de-friend friends, companies could consider as a no-hire decision, privacy options changes often
Decision Factor: You like to stay up to date with personal friends and family and/or you like to share information with a lot of people
Tip:  Only share information that is mom and co-worker friendly

Twitter

What is it?  An online social / micro blogging network where you can subscribe to anyone with an account and post your 140-character long updates (known as tweets).  Your main page is a running feed of updates from people whom you follow and people who mention you.
Pro:  Easy to stay up to date with the latest news on any topic, great for discussions during big events
Con: Waste of time to “tweet” if no one is following you, takes time to build credibility and get many followers
Decision Factor:  You want to stay up to date with a topic and/or you are a subject matter expert in a field and want to share your knowledge with everyone
Tip: Get more followers by participating in Twitter hash tag discussions (e.g., “#” followed by a string of words)

LinkedIn

What is it?  An online professional network where you can add co-workers, acquaintances, etc. as “connections”.  Your main page is a running feed of updates from your connections
Pro:  Recruiters love finding candidates here, opportunities to show off co-worker recommendations
Con: It’s hard to publicize that you are job hunting if you have co-workers as your connections, you can’t directly contact unknown connections (as a basic account member)
Decision Factor:  You may one day want to actively or passively seek a new job
Tip: Build up your LinkedIn connections today, don’t wait until you need a new job

There you have it!  F-T-L power!  It’s not for everyone, but surely one, two, or all is a fit for you.