Dropbox – The Lightweight, heavyweight.

Tim O’Reilly’s 8th web pattern is about “Lightweight Models & Cost-Effective Scalability”. This week will I will examine what it means to incorporate cost effective scalability into a successful web 2.0 endevour and examine the best practices involved in doing so. I highly recommend the following video, I think its perfect to tie the entire semester together, yes it’s long at 55 mins but it’s Tim O’reilly. He’s essentially the reason this subject exists and not only does he talk about the 8th pattern, he provides an excellent summary of how all the patterns work together.

If the model for Web 1.0 companies was “get big fast”, now it’s “small is the new big” why? For a start consider the up front costs involved in the antiquated thinking associated with web 1.0. A company could invest untold amounts of money, construct a website that they thought was great and hoped would be successful, and then watch it fail very quickly. This outdated thinking was to burst onto the scene, to take the internet by storm and be the new big thing without thinking too much further than that. The burst of Dot.Com bubble in May of 2000 is testament to this type of thinking just not working. 
There has been a paradigm shift, companies now realise that they need a defined strategy of approach, that re-use and leveraging from other sources is important, that they don’t need to do everything themselves. It is now about working the smartest and not necessarily about spending the most amount of money.
Dropbox adheres to these best practices in an examplary fashion. I think much of it has to do with it being a text book example of what a business needs to get right in order to become a true lightweight heavyweight.
Dropbox had humble beggings, essentially developed by Drew Houston as a program he could utilise himself after becomming frustrated with forgetting / losing USB sticks. Dropbox was essentially a one man development team responsible for a piece of software, that did everything right on a small scale. Let’s examine exactly how Dropbox grew to having over 100 million users, being responsible for .29% of the entire bandwidth on the internet by adhering to the best practices for this, the 8th O’reilly pattern.
Scale with demand
Scaling with demand is Dropbox’s forte. From it’s humble beginnings as a single user piece of software, it quickly established right away that it was indeed a force to be reckoned with (on a very small scale at least). As Dropbox gained clientele through strategic marketing it was simply a matter of acquiring more server space and keeping the service exactly the same fundamentally (of course there will always be feature add ons), and leveraging the long tail of the internet to reach more users.
Syndicate business models
Dropbox is has gone to great lengths to ensure that it is integrated into most online storage offerings. There are a plethora of addons and applications available that offer seamless integration with Dropbox. One click to Dropbox from many applications,  smart TV’s, smartphones, and its almost “part of the original package” appearance in the windows 7 operating system ensures it is exposed to a host of mutually beneficial revenue generating opportunities. In 2012 the Dropbox integration with Facebook groups was yet another display of innovation that has become synonymous with the name “Dropbox”
Outsource whenever practical and possible
Dropbox utilises amazon s3 data hosting services adhering to the “rent-a-server” model. This strategic move allows them to concentrate on what they do best, seamless integration with other programs without suffering the expense involved with establishing a physical data bank.
Provide outsource infrastructure, function, and expertise

Dropbox is a popular choice for infrastructure outsourcing, and its easy integration over almost any platform is particularly enticing to business ensuring adoption costs are low. Real time file synchronisation for all business team members, excellent reliability and backup offerings, as well as mobile access to files at all times makes it a perfect fit for many organisations.

Scale your pricing and revenue models

Dropbox caters for almost any type of user. They offer anyone a basic 2GB account, but offer extra storage space to those who refer the service to their friends, or purchase certain products or services. Recently, connecting the latest Samsung Galaxy to Dropbox via their application resulted in 50GB of bonus space. The high end of the market is also catered for offering businesses and large organisations a range of services for greater amounts of data with flexible pricing options.

Market Virally

There is incentive to use Dropbox almost everywhere and again this is in many ways as a result of its seamless integration with the products and services that matter. Offering users free space through third party programs is a win-win situation. Display a small Dropbox advertisement or integrate a plugin and gain the ability to offer your users storage space through you. 

Further Reading;


Leveraging the long tail with eBay

Leveraging the long tail is about exactly that, using the extreme popularity of generic products available on the internet that sell in high volume and levering that strength to expose more specific niche products that do not sell nearly as well with only limited popularity. The beauty of the “long tail” is that there are  a lot of these niche products – in fact they make up the majority of shop fronts on the internet.

Leveraging the long tail

Chris Anderson’s economic model of the Long Tail discusses the now diminished need “to lump products and consumers into one-size-fits-all containers”. With now infinite shelf space, low cost of materials and production courtesy of the internet, businesses are realising it’s simply a matter of exposure. So, how best do you leverage off the popularity of the best selling products on the internet? How are businesses able to sell their niche products that suit only very limited clientele and stay afloat? Remain profitable?

Tim O’Reilly outlines the best practices for leveraging the long tail in his seventh web 2.0 pattern, and this week I will examine eBay, the most popular online retailer in Australia an excellent example of these in effect.


Build on the driving forces of the Long Tail

eBay allows its users to list almost anything for sale relatively cheaply and with a minimum of fuss. It caters for almost any item allowing listings limited only by the imagination of its users facilitating an almost innumerable product base. Customers are drawn to eBay by cheap prices, the chance to win at auction or buy outright and possibly, most enticingly,  by the possibility of finding that item that they just cannot find anywhere else. eBay delivers hits to niches by first drawing in clientele with extremely competitive pricing for popular, generic products then offering then facilitates niche sales by custom tailoring searches for maximum exposure.

Use algorithmic data management to match supply and demand

Whenever an item is bought, whether is is via an auction, fixed price or classified sale, eBay displays similar items for the buyers perusal. When a buyer is searching for an item that is possibly generic in nature and arrives at a specific sellers page eBay displays all other items available by that seller. A user could be lead down a very niche path from seller to similar items (and repeat) until the ideal item is found. It is possible for a user to search for an item only once, then 10 sellers and different stores later – eBay delivers the final desired result specifically tailoring the list of item results to the user at every stage.


Use an architecture of participation to match supply and demand

eBay offers a feedback system where sellers can be ranked based on their past performance and quality of their items which allows buyers to always make informed decisions and purchase with confidence. When a search is carried out items are listed in an order of relevance to the search term by default so ideally the user receives the best matching item as the first result. In addition the results may be further sorted by the user selecting a relevant filter such as (lowest price first). If eBay does cannot return the exact item the user searches for it returns similar items in order of relevance.

Leverage customer self-service to cost effectively reach the entire web

eBay affords the user complete control of their accounts. User are able to list an item for sale on eBay in a matter of minutes (or seconds with the right app) through not only the website it self but also through many third party applications and the mobile apps. Users are able to purchase items, pay for items, and record feedback all from their very own account centre with the knowledge that eBay guarantees safety through PayPal where money can be refunded if a seller is fraudulent. The live chat support feature, extensive support articles, customisable user portals, and readily available API for developers contribute to reduced costs of operation.


Leverage the low-cost advantages of being online

The filtering, aggregation and search features offered by eBay for its 112.3 million users in conjunction with the self service model it is based on ensures that it reaches to the very edges of the spectrum of user interests and is coded appropriately to be effectively integrated into external searches. Indeed, it is so good that an item typed into a Google search box can yield a relevant result from eBay. Further, eBays involvement with extensive partner organisations ensures that advertising and marketing costs are kept to a minimum.

Further Reading:

Debating the long tail

Long Tail 101

Rethinking the Long Tail Theory: How to define ‘Hits’ and ‘Niches’

What does niche marketing mean?

Perpetual Beta – Development 2.0

This week I examine Tim O’Reilly’s 6th Web 2.0 pattern: Perpetual Beta and it’s corresponding best practices as harnessed by the social networking giant Facebook.

Software being in a state of perpetual beta means simply that it is is constantly evolving, constantly being shaped and crafted right in front of the eyes of its users. Software must be forever changing to remain a competitive commodity in the current era. Users no longer see software being “under construction” as something they have to wait out; an inconvenience. Rather, it is instant exposure to new developments, ideas and features for which they can provide instant feedback and rest assured that the developers are listening.

Perpetual BETA

Development 2.0 allows users to shape software with developers permanently implementing new concepts that rate highly  and those that rate poorly quickly disappearing from view.  Users must be treated as co-developers  the software release cycle is no longer about versions, it has become versionless – does anyone really know what version of Facebook they are using?  New features are integrated seamlessly and service remains uninterrupted. This is possible due to software moving away from an install once, wait, download update, install again cycle to an always on, always connected, Software as a Service.  O’Reilly states that “so fundamental is the shift from software as artifact to software as service that the software will cease to perform unless it is maintained on a daily basis.”  There is no down time, if you are not using your favorite web app someone else is and it is the opinions of these very users that dictates what is developed and when.


Facebook, like Google, exist among the pioneers of perpetual beta and adopt the best practices in textbook like fashion.

(1) Release early and release often

Facebook are masters in this realm, the introduction of timeline, chat and different layout format are classic examples of releasing early so that user feedback can be obtained and utilised to gauge interest in, and reception of, the new features it plans to introduce. If a feature is popular it is often quickly integrated and tailored around user feedback.

(2) Engage users as co-developers and real-time testers

Another advantage of releasing early and allowing users to trial new features well before a planned integration date is the fact that any bugs, whether they be logical or technical in nature, are likely to be uncovered by early adopters. This also allows Facebook to monitor user behavior and make well informed decisions regarding the way it rolls out the new features and what user demographic receives them first. “Timeline” is a classic feature that was made available to users as an optional update on September 29 2011, and after an overwhelming positive response eventually became a non-negotiable part of Facebook for all users.

(3) Instrument your product

It is absolutely essential to establish a framework for measuring success and implement metrics that provide an accurate snapshot based on this framework. Facebook features are meticulously monitored throughout different stages of implementation. Data gathered from user participation in varying areas is referenced against business objectives to further aid in successful feature implementation.

(4) Incrementally create new products

Facebook is constantly releasing new products and additions and manages strategic decisions through this, moving with technology in response to competition and to flat out user demand . Releasing a stand alone chat application for mobile devices competing with the likes of MSN and simultaneously recognizing the high demand for its chat feature is testament to such strategy.

(5) Make operations a core competency

Facebook realises that the people they employ are paramount to its continued success. They capture this perfectly by stating “We don’t have rules, we have values, focus on impact, move fast, be bold, be open, build social value”. Second to none systems monitoring, built in redundancies, state of the art hardware,  simplified fault tolerance, and razor sharp execution of all is the technology through which it empowers.

(6) Use dynamic tools and languages

Through utilising cutting edge agile development models Facebook is able to offer industry leading technology at all levels  including:

Front end: LAMP (Linux, Apache, MySQL and PHP); Back end: C++, Python, Java, Erlang

In addition to this Facebook offers an impressive array of open source developments.


Further Reading:

Version-Less Development: What is it and Why it’s Important

How Does Facebook Work? The Nuts and Bolts [Technology Explained]

Will product innovation win it for Facebook?

Careers at Facebook