Digital DNA

The last 6 months have been a blur. Mostly due to work. We had a crazy start to the year announcing new deals with Dailymotion and NBC, and also the appointment of ex-Googler Douglas Merrill to our Board.

Following our growth, Martin (Founder/CTO) and I have added more great people to our team and I have delegated a lot of what I was doing. The aim is to give me the time I need to focus on building The Filter’s profile in the tech and media industry and that includes contributing to my blog and The Filter’s team blog.

I was drawn to the story on YouTube trying to “avoid Van Halen fatigue” by improving recommendations.
My team at The Filter have been delivering recommendations based on behavior and consumption (rather than based on key words/tags) for some of the biggest video companies on the web (US TV networks and video aggregators). So, we know that what YouTube are talking about is true – delivering keyword based recommendations creates fatigue, whereas, using behavior and consumption data delivers increases in video views per visit (by over 20%) and in repeat visits.

Just seen the latest Nielsen numbers for video consumption on Mashable. In the US alone, there were 11.2 billion video streams in July – that’s some 31% growth compared to the year before. Mashable dig further into the stats pointing out that “there were 135.9 million unique viewers of online video… that means that the average viewer watched 82.4 streams in July.”

This continued growth in video consumption in the US confirms what we know – that there is a lot of activity in the video related digital market. Traditional media companies are digitising more content and promoting their online video services, and then digital start-ups (like Boxee and Hulu) are gaining traction. This growth in new content and video services is not slowing down any time soon, making it easier for people to find specific content (because it is all available through search) but making it harder to discover new content because there is too much of it.

My team at The Filter have been working closely with some of the top 10 video providers in Europe and the US to increase user engagement by finding ways to help discovery through offering relevant video recommendations. The results have been really exciting – in all of the services we have been working on over the last 12 months, we have been able to deliver uplifts in video streams per visit of 20-50%, and users who get more relevant recommendations also visit more frequently.

I was reading NewTeeVee’s quarterly review – it highlights two main facts: (i) we are buying more and more “living room” devices that are capable of playing online video (e.g. set top boxes, games consoles) – projections are for 57 million devices to be sold in 2009 in the US, and (ii) cable operators are sensing a threat and responding with their own authenticated services – e.g. Time Warner and Comcast with TV Everywhere.

This all translates to a healthy and active sector in terms of investment and opportunity. Which is refreshing in the current climate.

I can back this trend up with what we have been experiencing at The Filter over the last few months. Demand for our video recommendations and relevance technologies is at an all time high: whether from content owners, distributors or aggregators, they all have a pressing business need to find intelligent ways of connecting more of their content to their customers. Editorial and popularity-based suggestions are better than nothing, but our experience is that it is only when companies start using consumption data and sophisticated and matured algorithms that they get seriously exciting results in terms of increases in dwell time, video streams per visit and visits.

Every week there’s a news alert about ‘ad funded music services’ Spotify or We7 reaching new milestones for user numbers. Last week Spotify announced 2 million UK users, this week it’s We7′s turn to announce 2 million unique users.

I am chuffed to bits to see two European-based music businesses growing so rapidly. These guys have clearly demonstrated that there is a demand for free (and legal) music which is effortlessly available on the web/desktop. This demand will ultimately lead to us having access to all entertainment content at any time on any device. There are still many distribution, rights and business model issues that need to be sorted before this happens, but it will happen.

So, once digital distribution has been sorted, the businesses that will thrive will be those who have also sussed out digital demand. In a world where everything is available to everybody at any time and on any device, demand is no longer based on what is available, but on what content is most relevant to us at that specific time. We will need smart filters that serve up relevant content to us based on location, time, entourage, past activity, tastes and mood.

That is why I am so excited about the work that my team is doing at The Filter. For the last 5 years we have been developing recommendation and relevance technologies for all types of digital content that use multiple inputs (consumption data, meta data, location, time zone, social data, etc…) to build models that enable us to work out the probability of what content goes with what situation – by using behaviour and mashing it up with meta data to provide a user with the most relevant content possible for them.

I am very proud that we have proven beyond any doubt that providing users with relevant choices of content leads to an increase in consumption per visit AND an increase in the number of visits because the user finds more content that they like at each visit. Over the last 12 months partners that use The Filter on their sites have seen uplifts in consumption of 15-20% based on increases in streams per visit and visits per week. And for us, it is just the start. As we get access to more data (using location on mobiles) and fine tune our engines, technologies like this will provide better content choices to individuals leading to more consumption.