Last year we ran an evening conference call UGTV ‘06. It seemed to create some worthwhile conversations. (At the very least, we enjoyed it and met some interesting people.)
Since then, this whole area has developed rapidly. There is growing understanding of the the types of project that are likely to work and those that aren’t.
On a Mint Digital level, we’ve recently opened a New York office to better service the needs of two big American clients we are creating social applications for.
To celebrate, this year we are running events in New York and London. We’ve got a great line-up of speakers and good food, so they promise to be enjoyable and informative evenings. To request an invite, please go to the relevant page. Hope to see you there!
A few weeks ago we launched Joseph Choir Search, a new BloomBox site for the BBC. Running in conjunction with Any Dream Will Do, the website is a nationwide search for a school choir to support Joseph on the TV finale and at a West End charity gala performance.
A fortnight ago, traffic took off like a falsetto. As the deadline for submissions approached, more and more choirs uploaded their entries. For every choir that entered, there’s an army of pupils, parents and teachers behind them. It was not uncommon for local radio and newspapers to get involved, further fuelling the traffic. Grassroots marketing was driving traffic to a degree we simply hadn’t imagined possible.
Then we got linked off the BBC homepage, one of the most trafficked pages in the UK…
The Joseph BloomBox installation took a battering.
But it kept alive.
Last week, we had to disable voting to decrease the load on the database (we’d had 250,000 votes in 48 hours). Since then we’ve added more hardware, improved caching and made a heap of other optimisations. Since last Monday, we’ve resumed voting and Joseph Choir Search is performing smoothly amidst the voting frenzy.
The energy that’s gone into the competition is amazing. It is well worth checking some entries.
Joseph Choir Search is a standard BloomBox installation. There are a couple of innovations worth mentioning:
Audition room
We were keen to ensure that the best choirs rise to the top. We were worried that large schools would be able to mobilise more parents and teachers to vote, distorting the rating process. To counteract this potential problem, we implemented an audition room concept.
The audition room means that you can’t follow a link to vote for a choir. Instead, if you want to vote you enter the audition room, where you are shown a random stream of choirs (including the choir you initially came to see). This decreases the incentives for ballot stuffers and increases the value of the contribution from dispassionate amateur judges.
Looking at it now, the audition room concept does add a certain burden to the usability of the site. However it also seems to be helping ensure it is the best choirs that rise to the top.
Using JavaScript to create personalised, cached pages
One of the technical challenges of building many-to-many publishing sites, like Joseph Choir Search, is that every page is dynamic. This means there are more hits on the database and the sites are harder to scale.
With BloomBox, we’ve developed a technique whereby we cache versions of each page and then use JavaScript to figure out the dynamic part. This has been a key reason in why we’ve been able to serve many millions of dynamic pages, off a relatively small hardware installation.
Finally, I’d like to say thanks to Jo, Ted, Dan and Kate at the BBC. Firstly, for commissioning this project. The more we work on it, the more it seems genuinely innovative in the way it reaches out to and connects with its audience. And secondly, for supporting us as we’ve develop the site. It hasn’t been the smoothest process (we’ve made a couple of big mistakes along the way) but there’s been a genuine sense of teamwork as we have all strived to deliver a project that has turned out far bigger than any of us had initially anticipated.
