Thousands of applications have been created on Facebook’s open platform since the social network opened up to developers earlier this year. But launching one, and generating interest, is not necessarily easy.
‘Attitudinal matching’ company Synature is one UK firm that has already taken the plunge, launching a version of its qubox software that allows Facebook members to search for like-minded people on the site, as well as potentially becoming a platform for targeted advertising. Here, John Woods, its CEO, talks about the challenges the company faced, and why he thinks the API can be a significant opportunity for brands.
Why did you decide to launch qubox as a Facebook app?
Synature has always intended to build qubox into existing online communities – it’s part of our business strategy. We want our approach to become pervasive. We piloted this with LunarStorm, the social network, which was very popular with users. We wanted to build on that by allowing end users to be able to embed qubox in their existing social networks.
With that in mind, we began developing an embeddable widget form of qubox earlier this year. We’ve had a certain amount of success with that, but the capabilities of a widget like this are really limited because, in general, third-party sites offer only very limited support for embedded functionality.
To be specific; if you take our widget and embed it on your blog or in your profile for a site like MySpace or Bebo, you’ll be able to show each individual visitor to your profile how like-minded they are with you. So it’s a one-to-one comparison between the owner of the page and the visitor. That’s useful, but it’s only a small part of what qubox can do for you.
When the Facebook API became available, offering much more powerful access to the Facebook platform, that was clearly a natural environment for qubox. At the time, we were working on the launch of an official beta for our own qubox.com website, but it was too good an opportunity to miss, so we cut down our plans for the first qubox.com beta and concentrated our efforts on Facebook.
The application is far more sophisticated than a widget. It allows you to complete qubox like-mindedness quizzes within your Facebook page, compare yourself to all of your Facebook friends and see who is most and least like-minded, and search the whole of Facebook for like-minded people.
You can also invite new friends to try the qubox so they can compare themselves with you and link your quboxes to an account on qubox.com, so you can use the same like-mindedness information elsewhere.
The Facebook API will allow us to offer even more functionality in the future. For example, we have a ‘blogfeed’ feature on qubox.com that allows you to monitor blogs written by like-minded authors. We could embed that as an option within our Facebook app.
How has it gone so far?
The real measure of success will be whether we get strong viral growth in user numbers. If we do a good job with qubox, people will use it, enjoy it and invite their friends to install the application.
At the moment it’s a bit too early to say. The growth rate today [30th June] was 6%, and if we keep that up I’ll be very happy. That’s a daily growth rate of 6%, around 600% per month. But we need another couple of weeks before we can assess how robust that growth rate is. We’re on a steep learning curve and we accept that we may need to tweak things before we see the user numbers really take off.
How challenging has the technical side of things been?
At a technical level we’re really pleased with the application. A lot of existing Facebook apps are very lightweight – just a little messaging gimmick that you can embed in your profile page, for example.
With qubox we’ve built something that is technically complex – far more sophisticated than the vast majority of current Facebook apps. If it takes off you will be able to use qubox to search for like-minded people potentially across millions of Facebook users, so we had to be sure that the implementation would scale well for very large numbers of users.
We’ve had a few frustrations and learned a few lessons that might be of interest to others planning a Facebook app. Although the Facebook API is great, and Facebook offers very good support to developers in a variety of ways, there are still some rough edges.
We had a really frustrating experience immediately after our launch when Facebook made a change to their API which broke the qubox application. The timing couldn’t have been worse, and the symptoms were really frustrating for our users. The application appeared to lock up once the qubox was complete, so a user would go through the whole process and get no results at all.
We had several people posting frustrated messages on the discussion forum along the lines of “this app sucks!” when the problem was entirely down to Facebook’s change.
What advice would you give other developers about the Facebook API?
Because the Facebook platform is not fully mature, things will break for reasons beyond your control, and users will blame you.
There’s more risk of this the more complex or rich your application will be. This same problem applies when embedding widgets in other social networks; the social network can always change the technical rules, either deliberately or by accident, and that might break your app. Facebook is much better in this respect because it has an official API and really works hard to create a stable and well-defined platform, but things can still change.
You need to have damage control strategies in place and to monitor at high frequency. In our case, we were able to reduce the impact by disabling our application so users got an explanatory message instead of hitting the bug. This bought us time while we quickly worked out a fix.
It’s not easy to test a Facebook app in a controlled way, especially one like qubox where the functionality is about interactions between multiple users. There is a development environment in Facebook, which works really well, but it’s hard to have more than a handful of people test the app at one time.
This doesn’t matter too much for simple applications where testing for a single user might be enough, but for something like qubox which is all about interactions between multiple users, it’s hard to test everything without making it live. As soon as you do that, people you don’t expect will start using it – this is the downside of working in an environment that is so primed for viral spread! In our case, one of my personal friends dived in to the qubox app within about five minutes of the switch-on. We were still officially testing!
It’s really useful to have a dialogue with your user community; however clued up you think you are about Facebook, people will react in surprising ways to your application!
You need to plan on an evolutionary approach where new features are driven by what you learn from your users. Facebook automatically provides a Discussion Board for your application when you launch it, which is great, but it’s also useful to get feedback at the pre-launch stage so that when you go public you are confident your app really is ready for launch.
We created a qubox group on Facebook some time ago and recruited a few hundred people who were interested in the application. We used this to test our ideas pre-launch and to help kickstart the initial takeup of the application.
The other big challenge is the obvious one of end user relevance. I can imagine a lot of brand owners are thinking “I really need to be on Facebook now!” Smart thought – Facebook is going places and is a fantastic marketing platform.
But remember that Facebook apps are totally opt-in and can be uninstalled in seconds so unless you are offering something of genuine interest and value to the user, your app won’t succeed. On the flip side, if you get it right, you can get a huge audience very quickly and cheaply.
Will you be using your app to collect data for targeted advertising?
We are not doing this at the moment, but it’s a really interesting possibility for the future. We know from other projects that using our attitudinal information for target marketing purposes can give a huge lift in response rates, and obviously that’s good for both advertisers, who get more results, and consumers, who see more relevant content.
Anything we did in this area in the future would be transparent for the user and would comply with the very substantial privacy protections that Facebook’s API and terms of service offer. Ideally we’d try to package any targeted content in a form that is fun and engaging for the user.
Our strategy is to provide ‘attitudinal search’ functionality that will give compelling benefits to end users and spread virally, thus giving us a large user base. We look to layer target marketing functions on top of that. As an analogy, think of the way Google began with natural search and then layered keyword-driven advertising on top of that.
What are your thoughts on Facebook’s future and the significance of its open source strategy?
First off we think Facebook is a very exciting place to be. It’s fast-growing and successful, with 30M+ users. Although I’m sure other social networks will continue to succeed alongside Facebook, it seems inevitable Facebook is going to be one of the winners.
The Facebook API is a fantastic thing. It’s technically accomplished, allowing some very sophisticated styles of integration which most existing Facebook apps haven’t even scratched the surface of. It also creates a very natural and safe experience for end users. The rich API and the ease of access to the Facebook audience will continue to attract developers in huge numbers. There are already nearly 2000 apps.
So Facebook users are going to have access to huge amounts of exciting functionality. There will be a problem of ‘app spam’ – how many ‘new and improved Poke’-type apps do we need? But Darwinism will soon sort out the winners!
There will be other irritations for Facebook arising from the flood of third party apps. But I think it will be able to manage these problems, and ultimately the benefits for end users will far outweigh the negatives.
What are the costs?
To give you a rough idea, we spent about three person-months building our app – very quick given how sophisticated it is. That was because we already had a really solid and scaleable technology platform – when we first started developing qubox, nearly two years ago, we were already thinking about how to build the technology so it could be embedded in lots of different places. That forethought and planning really paid off big time for the Facebook app!
How do you expect other social networks to respond? Do you expect them to open up in the same way?
Other social networks are going to have to respond to the Facebook platform. It will be really interesting to see what emerges to woo app developers to other networks. Most likely we are going to see a raft of incompatible APIs appearing. Developers will have to look carefully at the cost-benefit of developing a custom app for each platform. Again, the more complex your app, the harder this problem becomes.
As a philosophical diversion, I’m not sure it’s all that helpful to think of Facebook as a direct competitor for MySpace, Bebo and the like. I think it’s a different beast and the differences will become more marked in the future.
We can already integrate into many social networks using our widget approach, but that offers limited functionality compared to what we can do with the Facebook API. I am sure we’ll see other networks opening up their platforms in the coming months in response to Facebook, and we’ll certainly look at integrating qubox more closely as the APIs become available. Of course we’re also open to licensing the technology to networks that want to keep control or get access to our data for targeting.
Q&A: Synature’s John Woods on attitudinal matching