Shayan Zadeh: Practical Entrepreneur

Simply put, the United States legislature was created to acknowledge the many problems and issues faced by people from all walks of life, and come up with sensible solutions that address the needs of most while preserving the rights of everybody. Then solutions are legislated into writing, in a way that’s consistent with past legislation.

If your job is to develop software, the above description may sound weirdly familiar. Software engineers build systems that address certain specific needs (problem space) within certain specific restraints, like the confines of our operating system (or various other restrictions), while maintaining backward compatibility with older solutions (most of the time).

These days, we hear a lot about the dysfunctional state of the legislative system, the burden of over-regulation on American businesses, and the problems of money influence in the political system.Perhaps the software industry can teach the American legislature a few valuable lessons on addressing some of these challenges.

As a thought experiment, try to imagine what software would look like if it followed the legislative system’s example. Following this analogy, our congressmen would be played by product managers, developers would play the role of bureaucrats who implement the laws, and the general public would be played by end users. Think about it. You’ll see how quickly everything would unravel.

If software was built like laws:

1. You’d fix bugs only by writing new code
2. You’d never break backward compatibility
3. You’d never, ever start over
4. You wouldn’t go back and refactor code in face of learning better ways to do things
5. Your code would have an extreme case of coupling
6. Your release cycle would be measured in years
7. Product managers would claim they could come up with any and all possible future scenarios
8. Product managers would come up with implementation details
9. Product managers would put plenty of special case conditions for their “friends” in requirements
10. Product managers would favor complexity over simplicity
11. And on and on

Every person involved with software at this point will be tearing their hair out! This description can actually be applied to almost every software project that has failed. And yes, lawmaking is mishandled in a manner even worst than this.

The inefficacy of lawmaking isn’t limited by Republican or Democrat party lines; both sides are equality to blame.

Congressmen – like product managers – should come up with high-level requirements and directions, leaving the implementation details to the regulators. This would lower the cost of implementing laws, and as regulators learn more about the system at the implementation level, it’d result in more agile adjustment of details. Plus, it’d even encourage analytical thinking and accountability, with the right metrics and KPIs defined.

Congress should also allow for resetting and refactoring every so often. Maybe automatically expire laws after a decade so we don’t end up with irrelevant legacy laws that make zero sense in the present universe. Believe me, we have the concept of “tech debt” in our legislature as well.

Lastly, when implementation details aren’t in the hands of product managers/legislators, it’s a lot harder to hide favors for special interests in between the million lines of code.

So, here’s to leaner legislation. Maybe we should start talking about minimum viable law!

A few months ago, Jeff (one of our rockstar engineers at Zoosk) had an idea: that our learnings from building Zoosk Messenger could be applied to chat on Facebook. After examining the capabilities and limitations of existing off-browser Facebook chat experiences, Jeff and I believed we could build a better, improved user experience. And that’s what we did. Jeff applied his magic to this new domain and, in a matter of a few weeks, he and his team debuted a new application that incorporated our original ideas. Socialbox was born.

Not knowing what to expect, we rolled out Socialbox and threw some traffic at it. Boy, did it stick. Unsurprisingly, consumers really wanted easier ways to chat with their Facebook friends. In only 6 weeks, Socialbox grew to approximately 2M MAU (see chat below courtesy of AppData).

Seeing this burst of interest, we took it one step further and added video chat to the product. We beat Facebook to video chat functionality on it’s own chat platform by a couple of weeks, which was really cool to achieve. It pays to be agile!

Video Chat for Facebook on Socialbox

In short, Socialbox is still a side project at Zoosk — but as long as consumers show love for the product we will continue to do awesome things with it. Socialbox is an example of how we encourage our team to approach their jobs at Zoosk. We reward creative thinking and we give our people the resources and support to turn their ideas into reality. We need more people who embody this commitment to innovation.

Most of modern science is based on a reductionist framework and it’s no surprise that we, the students of this body of science, resort to reductionism in trying to explain almost anything. Reductionism can be defined as

Reductionism can either mean (a) an approach to understanding the nature of complex things by reducing them to the interactions of their parts, or to simpler or more fundamental things or (b) a philosophical position that a complex system is nothing but the sum of its parts, and that an account of it can be reduced to accounts of individual constituents.

Then, it’s no surprise that when first confronted with the question of ‘Why was product X successful or a failure’, we naturally put our reductionist hat on and and try to reduce the product to its basic parts (usually features or marketing strategy) and assign values to them in their contribution to the success or failure of the product.
This approach works very well for products that are more or less independent of other products’ or peoples’ actions. For example, we can apply this model easily to find out why a Dyson vacuum cleaner dominated the market at a particular time period. Or, why the Motorola RAZR exploded in the the cellphone market back in 2003.
Equipped with this great framework of analysis, we naturally use it to answer the same question about networked web products. Hence questions like ‘What feature(s) of Facebook helped it beat MySpace?‘, ‘Did MySpace lose because of the Google Ad deal?‘, or ‘Which feature of Zoosk makes it so much better than Match.com?’.
The problem with applying Reductionist framework to understanding social web products is that we can not capture complex interactions between different users of the same product in any simplistic way. And that’s where we usually turn to emergence theory. As wikipedia puts it

In philosophy, systems theory, science, and art, emergence is the way complex systems and patterns arise out of a multiplicity of relatively simple interactions. Emergence is central to the theories of integrative levels and of complex systems.

What we need to remember is that not all phenomenons can be explained using a reductionist framework. This is specially true for networked web products which social products are a great example of. So, maybe next time we want to analyze the rise or fall of the next social web product, we should not go straight for breaking it down into individual features. It might save us from drawing simplistic conclusions that can lead to bad decisions.

A while back, i did a talk at Freemium Summit 2010 about building a business on someone else’s platform. It looks like it has made its way onto YouTube . Below I have embedded the first part of the talk. you can find parts 2 and 3 on YouTube easily.

Based on my experience at Zoosk, I have submitted a panel to SXSWi conference for next year titled “Throwing Your Shoestring At Being Platform Agnostic“. I am planning to talk about what it really means to have a platform agnostic product and how can a startup afford to build their product for multiple platforms.

The selection committee has liked the idea enough to put it to community vote. Now, I need your help and votes to make sure it gets accepted. Please visit the panel picker and give it a thumbs up! Make sure you are logged-in or the vote won’t count.

See you all in Austin in 2010

For the past few years online advertising has been the de-facto monetization mechanism for online businesses that do not sell physical goods. As a result very few companies have actually focused on receiving payments directly from consumers. This lack of attention really shows when you try to figure out a holistic approach to online payments for your business. The short news: global online payment systems are a mess especially if you are not a 10 year old S&P 500 company!

As this economic downturn puts pressure on more and more web startups to skip the dwindling online advertising revenue stream and go directly to the consumers more entrepreneurs will get exposed to this mess. Since I have been dealing with this problem for almost a year now, I thought I should just summarize what is wrong with this ecosystem. Hopefully this post provides some guidance to fellow web entrepreneurs that are/will go through the same and save them some time.

I am also going to propose some basic requirements for anybody who wants to solve this problem. My hope is that somebody in the payment business will take a look at this list and hopefully provide a product that satisfies these requirements. I really believe that there is a huge opportunity here and whoever cracks this problem properly will benefit immensely. If I was not busy building Zoosk, this would be on the top of my list of opportunities to tackle.

Read the rest of this entry »

I am testing blogging from the wordpress application on iphone 2.0. So far it’s pretty smooth. I wonder if I can post a photo snapped by the phone directly into a post. That’s the best feature of the Facebook app if you ask me.

So, let’s see how this test turns out

Aha! I just figured out the photo stuff! Pretty neat and intuitive. Attaching a photo I snapped yesterday of first Zoosk roof get together. Now let’s see how it posts…

Recently vice president of product marketing at Facebook, Chamath Palihapitiya, spoke about the Facebook developer platform at the TieCon conference. Some of the figures what were quotes by the attending press caught my attention.

And about 33 percent of Facebook application makers reported profits of up to $500,000 a month. Finally, at least one-quarter of the applications running on Facebook have 100,000 active daily users.

Looking at adonomics or other Facebook application trackers, you can see that only 50 Facebook applications boost 100,000 and above daily active users. By Chamath’s math, Facebook only has 200 applications! I guess MySpace platform is not doing too bad in retrospect.

In reality, Facebook has some 25,000 applications on it’s platform. and only 0.2% of those have more than 100K daily visitors. I am just hoping that Chamath was miss-quoted, and that he and his team don’t really think every application on their platform is thriving.

And imagine if 33% of Facebook applications were making half a million in profits every month. That would put the profit of just this 1/3 of applications at around $50 million dollars a year! That should put their revenue in the $150 million range, exactly how much Facebook made in 2007!
Simply AMAZING

If you are following social media development, then you probably have seen or heard of the O’Reilly Research analysis published here and analyzed here (among many other places). Ben Lorica’s post on Radar provides a very interesting glimpse into the Facebook ecosystem and highlights some industries that are doing well through an in-house categorization of applications and comparing the corresponding number of active users.

As this chart from the research highlights, the biggest category on Facebook is applications that are ‘just for fun’. Obviously the assessment that the most successful applications on Facebook are those that have marginal value and are time wasters resulted in extensive coverage of the report in the blogosphere and ended in headlines such as: On Facebook, Girls And Boys Just Want To Have Fun.

I think the categorization of the applications by the research team has a lot to do with this assessment. I have been looking at the same question from a different angle for a while which might be useful in structuring a discussion around the topic.

Facebook was a social network way before it became a platform. It had millions of users that were active on the site way before any of these applications existed. When the platform was introduced last year, from a user’s perspective, all of a sudden new functionalities became available. Some of the functionalities were improved versions of the existing features (think SuperWall) and some opened completely new opportunities (such as iLike for music).

So what happens if you look at the Facebook universe in terms of activities that were facilitated before and after the platform launch. Below i have tried to summarize the main activities of a typical Facebook user (please comment if I have missed major areas). I have also included (a very subjective) assessment of how Facebook was doing in facilitating each category before the platform launched by using different font sizes.

In the right most column I have included some of the applications that have emerged to improve the user experience for each activity plus the total daily active users of all such applications in the top 20 facebook applications (in terms of daily actives). Please note that the active numbers are based on back-of-envelope calculation on a random day and only for top 20 applications. In the O’Reilly report, they summed all applications up which takes the tail into account as well.
I think this perspective highlights a couple of trends/questions

• What’s up with Photo Sharing? Is Facebook Photo application so perfect that nobody can provide value beyond it? This seems like an excellent opportunity for a brand new application to come out of no where and take over.
• Improving existing functionality/purpose has the least barrier to user adoption. This is especially true in case of sub-perfect existing experiences. Facebook wall was boring and text only when FunWall and SuperWall exploded in popularity. People discovery still sucks on Facebook which is a great opportunity for Zoosk and others innovating in this area.
• Media consumption has always been a part of most successful social networks except for Facebook. To this day I don’t understand why Facebook is so bad when it comes to media discovry and sharing. But this was a known winner and iLike and Flixter capitalized on the opportunity very well.
• Social gaming seems to be the only truly new activity category when it comes to successful applications. It was basically ignored by (or unknown to) social networks until Scrabulous and Friends for Sale came into the picture on Facebook. This category now is rightly so getting plenty of attention from the venture community and entrepreneurs.

What does this mean for applications not in these categories? Is this the extent of utility that application developers can provide on social networks? I don’t think so. Social gaming didn’t really realize it’s potential until early February this year, full 9 months after the platform opened up. What is the next big emerging activity category? That being said, I believe it is much more natural to grow on social networks if you are enhancing or extending existing user bahaviors.

I should also highlight that this analysis is only looking at very large applications. If your application does not benefit from network effects then you don’t need to be huge in order to be able to provide utility. One more thing to keep in mind is that monetization potential is completely left out of this discussion. Even though having a large active user base helps with monetization opportunities, it does not guarantee the way to profit.

Is this trend going to hold up on other social networks that are opening up? It’s too early to tell. But the initial indications suggest it might hold true on MySpace and Bebo as well.

PS: Obviously a large contributor to the success of applications is the viral engineering of applications. For each of the successful applications in these categories there are hundreds of tiny copies. Just because you pick a hot category does not mean you will have a huge user base. It might just make it a little bit easier to do so. Follow Dave or Andrew‘s blog if you want some viral wisdom.

One of the most mind boggling aspects of the MySpace Developer platform is their “Application Approval” process. In short, when you create a new application and want to make it live, you “publish” it which means somebody at MySpace  gets notified, installs the application, and plays with it. Then they check the application profile and home views (basically stuff that the application puts on your profile, etc) and if all is to their liking, they bless your application and normal users can see the app.

In my experience it takes between 24 to 96 hours to get somebody from MySpace to look at an application and make a judgment. On their forums I have seen horror stories about applications that have been in pending state (from the time you publish to the time they look at it) for more than a few weeks but I haven’t personally experienced it (thank god!).

This approval process is definitely costly both for the application developers and MySpace. MySpace has to have people on staff that basically just install applications, test them out and make approval decisions. Basically MySpace is helping you outsource your QA team (and them some – see below) and foots that bill.

It’s also costly for the developer because MySpace Approval Team rejects submitted applications multiple times, in each round disclosing some of the areas they are not please with. So, from the time you think your open social application is ready to go to the time it makes it to real users it could take between a few days up to weeks depending on your luck and complexity of the application.

It even gets worst. Every time you update something in your application code (say you decide to make that navigation bar purple from blue!) MySpace flags your application for re-approval and a human being looks at it again! I am really curious how much this whole thing is costing them. They are not making that much money so they better get on some cost cutting!

So you might wonder what is their approval process like. MySpace says they check to see if the application meets their Terms of Service. But in my experience they look for these categories

• Application Works
• Application References External Scripts
• Application Uses OpenSocial
• Application Impedes Browser Functionality???
• Application Makes MySpace Tester Happy?
• Application Satisfies MySpace Tester’s Expected Pointing System (for games only!)
• Application “is Valuable”?
• Application…

Basically it’s a very subjective process that really doesn’t add any value to the platform, MySpace user experience or anything close to that. It simply costs MySpace and application developers time and money! There is nothing in the approval process that market forces (i.e., MySpace users) can’t weed out. I really wonder what would have happened if Facebook tried to do something this backwards with their platform. In my opinion this just proves how different MySpace mentality is from a typical “technology company”. If you ask any tech company about moderation/approval/policing they would all say we let our users help us with it. Let’s develop some technology that users can flag “bad” behavior and we punish the common offenders! But MySpace likes the police state route better.

Just to give you an idea about how the police state if MySpace works, I am posting a few “official” rejection emails that I have receieved from them in the course of past couple of weeks. It makes for a fun reading. These are verbatim MySpace comments with identifiable pieces taken out (replaced with all caps placeholders such as LINK1)

Your application does not have any meaningfull functionality. Their is one problem with your application (when clicked on “LINK” your application has some formatting problems) Please try to add some useful functionality to your application, fix formatting problem and resubmit your application

In your application after passing the links “LINK1 and LINK2″ a link is open “LINK3″, which redirect the user to add/remove page and impede the browser functionality. Secondly, when a visitor comes to view a user’s profile who has your application installed ( visitor – user without app; user – user with app),and click on the link “LINK4″ this link redirect the user to add/remove page, which is ok but impede the browser functionality

the transition between the canvas and install page impedes browser functionality as it stalls and doesn’t allow the user to use the back button, but if you press it fast enough in repetition, then it is accomplished.  So please complete the functionality.

There are still some functional difficulties with your application. When one user votes another user, the former user gets a point added in the POINT-SYSTEM, when only the latter user should be getting the points.

The application is working but partially. The meter does not go beyond 5 even if the user wins the polls several times and neither does the comment change for the user who is winning. So please review this part of your application