Like most chinwag events (in the main attended by commercial media/advertising types), this event followed suit with a strong commercial edge. The panel included a good mix of talent from both sides of the fence. Walid Al Saqqaf (Trusted Places), Lisa Ditlefsen (Base One), David Maher-Roberts (The Filter), Luke Errington (Reevoo) and Jon Myers (Media Vest)
Pitched in a ‘what’s better?’ format, the recommendation chaps all conceded that a high percentage of site traffic arrives via search. So no surprises there then, recommendation sites are dependent on search to ‘get found.’ Where the panel felt recommendation services offer mileage, is in the richness of information, credibility of voice and social value offered. So, even though people use search (aka Google) to find stuff they want, recommendation services can often offer users more relevant and helpful results. Particularly when it comes to ‘product,’ queries.
For example, when searching for earphones to replace the crappy ones that come with your ipod, conducting a query in Google will return results that take you straight to purchase. As a user, what I really want is for search to point me to independent reviews about different earphone models that I can compare. Credible recommendations that enable me to make a more informed decision about the best earphones for me.
So, the independence and social credence of recommendation services has weight and value for users. But, because ‘search’ is our universal starting point, we often miss out on useful review content. There is an opportunity therefore for search to up it’s game and ‘get wise,’ to recommendations. Offer users results that blend ‘direct to product purchase’ and ‘credible independent reviews.’
It’s not all sunshine and roses in recommendation land though. While search often fails to pick up on reviews, recommendation services fall short in terms of completing the deal. They rarely link users through to the information, products and places discussed (directly to purchase/action). Other issues facing recommendation services include managing the quantity of reviews, integrating personalisation and broadening content coverage – plugging the gaps. Depending on the information I’m looking for, the number of recommendation results I’m offered can often feel overwhelming. This makes it difficult to compare, make judgment or digest the information on offer. Because recommendation services are generally underpinned by a commercial business model, there are gaps in the recommendation service market. Services rarely reference public service ‘products’ for example, across health (e.g. doctors and hospitals), arts and culture (e.g. plays and exhibitions), education (e.g. schools and colleges) and other local services and information. So, as well as helping with ‘what to buy,’ it would be great if recommendation services started to emerge around broader areas of public life.
Thinking about the future, a feature of recommendation sites applauded by the panel was the possibility to search ’emotionally,’ and discover new things. On Trusted Places for example, I can search for the ‘type of place’ that I am looking for as well as by a specific name, type of food or location. I want to go somewhere romantic, somewhere cheap, somewhere sophisticated or somewhere quirky for example. Although it’s less intelligent, I Feel London offers a similar kind of service. A user generated location based mash up, I Feel London is designed for emotional search. Where’s good to go if I feel hungover, naughty or broke? Like Lastfm in the music market, these newer services represent a next generation of recommendation website for emotional, location based and social decision making.
Following this theme, one of the more interesting representatives on the panel was David Maher-Roberts of The Filter. Currently in beta, The Filter promises an intelligent and personalised recommendation engine for entertainment (music, films and online video). With ambitions to raise the bar for recommendation, The Filter invites users to plug The Filter engine into their local media applications (i.e. itunes, lastfm and winamp) and social networks – the people that hold social currency and credibility in their networks. Sucking in this kind (and level) of data will enable The Filter to ‘get to know’ their users better and generate more intelligent and nuanced recommendations. Recommendations informed by a users media consumptions habits, taste and social world. The social element of The Filter however (the element that could really make it stand out from the crowd), still feels a little manual at the moment. Users need to invite their friends to join The Filter community in order to share and connect media consumption data. Oh joy – another social network to spam my pals with 🙂
What I’d personally like to see (and this may very well be where The Filter is heading) is a recommendation service that plugs into my distributed ‘vapour (data) trail.’ An engine that links up to my broader networked activity sucking in data about the products I choose to buy and the conversations I engage in online (e.g. via IM, facebook and twitter, the blogs I consume and contribute to, the media sharing services I use and the videos i watch on youtube). It’s not an easy task for sure, but with services like BBC Sound Index emerging, the possibility of a distributed recommendation model must be getting closer to a reality.
As the next generation of recommendation services start to emerge, I imagine that assessments of social currency and review credibility, data context and privacy will become increasingly important in the design and development of new services. Whose opinion do I want when I’m looking for new music or films to try? Is this different when I’m looking for advice about other products, for example a digital camera, a bike, a car or a home? Whose tips are most relevant when I’m visiting unfamiliar territories – when I’m on holiday or away with work? Suggestions from a friend that visited 5 years ago or local people with similar tastes and social habits to my own? A complex set of questions with implications for ‘back end’ development, IA and editorial integration. These are some of the questions that need thinking about in the development of new kinds of recommendation services…
If you’ve got this far, thank you for reading. As a gesture of thanks, the ‘sexy talk,’ bit that I promised at the top follows right now…
EXCLUSIVE FROM CHALKBOARD – Between a question from the floor (@chinwag) and a comment by Jon Myers, it emerged that ‘Google personified’ has a penis and a set of breasts. Probably more than a handful 🙂
-claire welsby –