How do you grow an online business? Take one small orange…

Single small orange

We have all heard of the amazing success of some online business. One common theme is their ability to apparently do more with less and avoid the large size and cost of the ‘bricks and mortar’ business’.

One stellar success is Wikipedia, the product of Wikimedia. The success story goes something like this: In 2007, Wikimedia was based in a shopping centre with fewer than 10 employess, raised under $3 million annually, and had total assets of over $5 million. Today, Wikimedia has roughly 160 paid employees, raised $25 million from 1.2 million donors, and has total assets of over $49 million. So what did they do right?

Maybe we should be asking, what did they do wrong? Apparently, their first version was called Nupedia and users had to fax copies of their credentials AND pass a seven-stage review to prove they were qualified to contribute. Unsurprisingly, it was a total flop – but provided the lessons required to make Wikipedia the success it is today.

The success of Wikipedia and other online business can be explained by one of the Web 2.0 patterns proposed by O’Reilly of ‘lightweight models and cost-effective scalability’. Although O’Reilly was focused on programming, this can also be applied to business models, development models and technology solutions. The main idea is to do more with less and focus on small low-cost startup, scale in a responsive and cost-effective way, and fail fast and correct quickly.

Since we all know Wikipedia has got it sorted, I’d like to take a closer look at another small startup and see if they are headed in the right direction.

A Small Orange is a small web-hosting business that is getting good recommendations in the web-hosting space but have only 30 employees who serve tens of thousands of customers around the world with 24/7 support.

They started 10 years ago with 2 employees and have grown to be named in the top of the web-hosting options  and have gained a reputation for having over $1 million in revenue, are profitable and didn’t take venture capital. So how do they do it?

A Small Orange has spent the last 10 years working towards the best-practices of the pattern:

  • They scale their pricing and revenue models with a small up-front and sliding costs for more reliable servers.
  • They outsource their servers and recommend other services (eg HootSuite at so customers can manage and track their accounts.
  • They scaled their business model with the recent introduction of the ‘affiliates program’ to reward customers for recommendations.
  • They market virally, with their chief executive creating quirky and commercial YouTube videos.
  • They ‘fail fast and scale fast‘ when they were suffering from poor servers and bad customer reviews in 2008 leading to a server upgrade in 2011 and 2012 and improved customer reviews.

A Small Orange highlights some of the benefits of the pattern with their faster time to market and faster return-on-investment due to the lower initial costs and avoiding the need for venture capital.

But they have also experienced some of the problems. Their low staffing has resulted in some slow customer service and their outsourced servers left them exposed when the servers were not proving adequate.

They could also capitalise on some other ideas like outsourcing some customer support by using ‘crowd sourcing’ to let other customers provide solutions or suggestions to reduce the drain on customer support staff.

Overall, they have proven success from this pattern and look like they are headed for more success in the future.


An example of this light-weight and scalable pattern being implemented as a business initiative within a large organisation is this example of innovation, crowd-sourcing and the concept of ‘fail fast, fail cheaply’ within AMP, Australia:  


UPDATED 24 May 2013: How does the Tumblr acquisition by Yahoo fit the ‘scalability’ approach?

UPDATED 25 May 2013: Tumblr users are fleeing


Is LinkedIn a site worth recommending?

(Image: Nan Palmero)

(Image: Nan Palmero)

They say that great Web 2.0 organisations understand how Data is the next ‘Intel Inside’.  On that basis, is LinkedIn a Web 2.0 site worth recommending?

Tim O’Reilly identified 8 core Web 2.0 patterns in his article “What is Web 2.0: Design Patterns and Business Models for the Next Generation of Software”. This week, I am looking at LinkedIn in relation to O’Reilly’s pattern “Data is the next Intel Inside” . This concept is based on the idea that “Web 2.0 serves as a platform for users to accumulate content that everyone shares“.

LinkedIn has developed their application as a professional networking and recruiting platform.  Individuals and companies can add their own information which is then available to other individuals or companies. It is based on the 6-degress of separation concept which highlights the value of your network.  I have stepped through some Best Practices to see how LinkedIn applies them. So lets take a look inside.

Own a unique, hard to recreate source of data. 

The LinkedIn dataset is truly unique since it was built by the users themselves and is not available anywhere else. It could be recreated by a compilation of resumes from users and recruiters but nowhere is it found all in one place. LinkedIn has also established significant ownership driven from their default privacy settings.

Enhance the core data.

LinkedIn has made their core data better by enhancing it. Explicitly, they enhance through user recommendations and skills endorsements. Implicitly, they track and report usage (e.g. how often a user has come up in searches, how often a user is viewed).

Allow users to control their own data.

LinkedIn does allow users to control their own data and their are plenty of sites with advice on how to do this. They also allow users to view and export their data. Despite how counter intuitive this seems, it gives the user more control and more confidence in the site.

Make some rights reserved, but not all.

LinkedIn has walked a fine line between giving users all data rights and giving users no data rights. There is plenty of advice from LinkedIn on what you own and what they own – if you can wade through the User Agreement.

Linkedin imageDefine a data strategy.

I believe LinkedIn is very aware of the value of their data – and has grown a very successful business from it. They got negative feedback from their user data changes in 2011 and had a more user-friendly attempt in 2012.

Own the index, namespace or format.

This Best Practice is worth considering if you can’t or don’t own the underlying data, but this is not a big issue for LinkedIn since they significant data ownership.

Design data for reuse.

O’Reilly says not to overlook the value in operational or organisational data. One great example is Coke using their stock control data . There seems to be no public information about this in relation to LinkedIn.

Outsource or supply data access management.

Again, there is not much public information about this in relation to LinkedIn.

You would think that any organisation that follows Best Practices, would have solved all of their problems.

Unfortunately there are some data issues and debates that every organisation, including LinkedIn, needs to consider.

I think LinkedIn has clear data ownership based on their strong User Agreements, refined over their 10 years in business, so this is not a problem for them.

LinkedIn has now overcome the start-up challenge of getting enough data to get started.  With over 200 million users world-wide, they have plenty of users to provide user content.

The Open Data Movement is something LinkedIn should watch since their business model relies on data ownership. However, since they have enhanced their core data with new features (events, groups etc) they are less susceptible due to this data diversification.

The Concerns with Copyright is less of a issue for LinkedIn since most of the content is user created and they have User Agreements clarifying ownership. 

So, is LinkedIn a Web 2.0 site worth recommending? 

I believe LinkedIn has clearly focused on this pattern for successful Web 2.0 applications.

Their data is core to their business and they seem to have addressed most areas of Best Practice quite strongly. They have not avoided issues around user data privacy, but based on the large number of users and small amount of bad press they seem to be managing this quite well.

My advice to a business looking to succeed like LinkedIn in the Web 2.0 arena is – take a good look at how LinkedIn has implemented “Data is the next Intel Inside”.

My advice to users of the LinkedIn application is – take some professional advice from LinkedIn’s founder… and be aware of your LinkedIn privacy settings.

If you want a quick personalised tour, take a look at this video to see how your content is of value to you and your network based on the ‘6 degrees of separation‘ concept.


Related sites:

“I can’t get this data out of my mind” about LinkedIn – blog by Matt Low

“Harnessing Collective Intelligence” with LinkedIn – blog by Eman Alsheheri

How does TripAdvisor rate in social media?

ImageThe TripAdvisor website is a popular site that allows users to read, raise questions, and provide comments on travel experiences. It was an early adopter of user-generated content and produces ratings on travel destinations and providers based on user content. They are successfully ‘managing the human cloud‘ and utilising a ‘virtual workforce’ – but how are they doing it and why is it successful? 

Companies have increasing opportunities to tap into a virtual, on-demand workforce.

To give some framework, I’m using the pattern  “Harnessing Collective Intelligence” from ‘What is Web 2.0: Design Patterns and Business Models for the Next Generation of Software’. This concept of ‘harnessing the crowd’ means that others outside your organisation provide information to your business that can become a ‘powerful asset’. So, in the spirit of user reviews, this post explores “how does it rate?“.

1. Reward the user first: TripAdvisor allows user to achieve their objectives by either accessing other people’s reviews and comments or  posting their own questions or comments. Their users seem happy, based on their recent milestone of 100 million reviews and opinions. (My rating: 5/5)

2. Set network effects by default: Network effects are critical for TripAdvisor who need a good supply of content and reviews to attract and retain their users. The network effects start to show when TripAdvisor ratings are quoted on other travel booking sites. (My rating: 4/5)

3. Involve users, explicitly and implicitly:  TripAdvisor clearly has explicit user participation covered through their user-created content, and guessing by their advertising they would not be losing the implicit user information around user actions and preferences.  The site has attracted a strong user base and there are good indicators of this increasing. (My rating: 4/5)

4. Provide a meaningful context for creation: The TripAdvisor context of travel is an easy attractor, but I think the loose structure  means posts can end up in the wrong location and “findability” is currently not great. Some users, may be happy to ‘wander a while’ as they plan their trip, but it seems TripAdvisor has identified some improvements with new search functionality  in the pipeline.   (My rating: 3/5)

5. Trust your users:  A true Web 2.0 application must trust the users to provide content and to share control. TripAdvisor allows users to  provide comments, start new threads, and provide advice to other users on where their thread should go. The users also seem to trust each other since the majority of users won’t book a hotel that has no reviews.  (My rating: 5/5)

6. Software design that improves with more users : The goal here is to ensure more users produces more value and benefits not more mess and chaos. Their large user base also gives TripAdvisor credibility in the competitive landscape and means their ratings are actually quoted on other travel sites. (My rating: 5/5)

7. Application that facilitates emergence : This is all about giving the application enough flexibility to change to suit the emerging needs of the users. It means not presuming all features up-front, but instead adding extras in as required. TripAdvisor has some new innovations with stats and infographics for businesses,  advice for hoteliers to increase user engagement and  support for providers under attack. (My rating: 4/5)

So, according to O’Reilly’s pattern of “Harnessing Collective Intelligence”, TripAdvisor should be on your “must see” list of social media sites. But is this the whole story?

There are some issues around the “Harnessing Collective Intelligence” and TripAdvisor has experienced a few of them.

Privacy and liability for individuals: The anonymous posting allowed under TripAdvisor leaves few risks to the posting individuals. However, there is still a chance that individuals associated with the provider could be at risk based on negative reviews or comments.

Privacy and liability for providers: This seems to be causing the most challenges for TripAdvisor. Their policy of accepting all anonymous posts has led to legal battles in Australia and the UK from providers who believe they are the target of untrue, malicious, or adverse commercial attacks. 

Quality, not just quantity, matters: At the end of the day, the site content needs to be credible and reliable. Some comments suggest that if TripAdvisor develops strategies for verifying posts and checking the credentials of posters, then the quality of this site will strengthen.

So, how does TripAdvisor rate as a social media site? According to the Web 2.0 pattern of “Harnessing Collective Intelligence” it gets top-marks.

Their founder journey is a great story about ‘Harnessing Collective Intelligence’ and also explains how they had to fight to retain the value they had built up against big players like Google.

Disclaimer: Bronwyn is not associated with TripAdvisor but she does have a love of travel and is a frequent visitor to their site.


Related Posts:

How TripAdvisor is using ‘Innovation in Assembly’ – blog by Matt J Low

How TripAdvisor is ‘Harnessing Collective Intelligence’ – blog by Monique Alvis


Related Articles:

Using the crowd as an innovation partner –


Updated 12/4/2013 – added Related Article on ‘using the crowd’