Personalisation and Resource Discovery (Or, Can the Archives Hub Learn a Few Lessons From Amazon?)

As the team thinks carefully about the future of the Hub, we are pressed to examine current trends and developments surrounding the UK (and broader) Information Environment, and to make sure our long-term strategic aims are in line with those trends. In other words, we need to predict which technologies and user-expectations are going to take a hold, and make sure the Hub is in step with that future. Understanding those trends and these possible futures is no simple matter, and the trend of ‘Personalisation’ is a perfect example of a seriously complex area that we must examine with real scrutiny.

JISC (our funders) are investing a lot of time and revenue into personalisation — funding several studies in the area, including this one — Developing Personalisation for the Information Environment and encouraging its services to consider ways in which users might have personalised experiences when accessing and using content. The first of these studies has specifically looked at the relation of JISC services to social-networking and collaborative environments, surveying all ‘web 2.0’ implementations currently in effect within JISC services. The second study, still underway, continues this scoping work, but aims to look specifically at ‘opportunities to personalise sites adaptively in a way that is transparent to the user’ (see page 1 of their interim report).

What is ‘adaptive personalisation’ and what might it mean for services like the Hub? Many JISC services offer some sort of personalisation where users can customise their experiences — for instance Zetoc’s RSS alerts, Copac’s search RSS, or Intute’s bookmarking tools — but adaptive personalisation is different in that the system uses information it knows about a user to ‘push’ content. This technique is already used to great effect by commercial organisations, the most obvious being Amazon and eBay, who collect usage data (what you searched, what you clicked on, what you bought) to suggest or ‘recommend’ items to you.

This is a rather clever marketing technique, but of course from a resource discovery standpoint, there is a great deal of potential — notwithstanding the fact that you need a vast amount of usage data to make this form of personalisation meaningful. In my days slogging through the Ph.D., I often found Amazon a useful research tool for discovering books that my library searches had not uncovered — I would search for a book that I already had, and scavenged the ‘people who bought this, also bought this’ lists. (I suppose this might be cheating, but I prefer to call it ‘enterprising’!) In the interdisciplinary field I was researching (history of technology) this was a highly productive method of surfacing relevant records, as the library metadata might not necessarily reflect the subject matter.

More interestingly, however, is the fact that not only was I finding content, I was also — if on a very peripheral level — engaging with a community of peers. People ‘like me’ who were also interested in the same research questions (or, in more mercenary terms, I knew what the competition was up to).

So what will the Archives Hub of the future look like? More to the point, what will be the experience of Archives Hub users? These are questions that form the focus of a think-tank meeting we are holding next week here in Manchester. Our Steering Committee, along with some other stakeholders, will be joining us to think collaboratively about our future, and we’re very much looking forward to it. Will personalisation (in its many forms) or social networking have a role here? And if so, in what ways? Will the Hub users of the future find records ‘recommended’ to them? Will they be able to share, comment, or annotate records (will they want to?) All of these questions, of course, get at the very heart of what it is we do as a profession (archivists, information professionals, researchers) and in some ways begin to undermine some of our traditional practices or assumptions about cataloguing and standards — what it actually means to describe something. Who gets to describe (and who doesn’t)? For what purposes?

It’s tricky territory, for sure, but exciting and challenging nonetheless. We’d be curious to hear your thoughts about these issues, and especially in terms of the Hub’s future. In turn, we’ll look forward to sharing with you what we learn from our day.

It’s a matter of trust

I attended the Eduserv 2008 Symposium recently. The theme was ‘What do current Web trends tell us about the future of ICT provision for learners and researchers?’. The day provided a good mix of speakers, and for me one in particular stood out, Geoff Bilder from CrossRef. His talk was intriguingly entitled ‘Sausages, coffee, chickens and the web: Establishing new trust metrics for scholarly communication’.

Whilst writing this blog I visited Geoff’s blog Louche Cannon (very entertaining it is too) and there he refers to his feelings about the thorny issue of trust:

“It may sound incredibly un-hip and reactionary, but to hell with the wisdom of crowds. Watching the crowd might be entertaining, but when I need to work, I can get far better results if I constrain that crowd to a few people whose opinions I have reason to respect.”

Geoff’s main point was that we are really continuing to underestimate the importance of trust. It is often implicit but rarely explicit. He referred to what he called the ‘Internet Trust Anti-Pattern’ whereby a system is set up by a core of self-selecting high-trust technologists. Then the masses, for want of a better expression, start to use the system and as it becomes more successful the risk grows that a strain will be placed on it by untrustworthy users. Think of spam, viruses, phishing, and generally dodgy content – we are all well aware of these things and find them a real nuisance on a daily basis. There is undoubtedly a trust problem and users are often uneasy, though generally we have not reached the point where systems are widely declared to be ‘untrustworthy’.

When we think about how we establish trust, it may be through personal acquaintance, perhaps we trust someone because someone we know trusts them. It may alternatively be through a proxy, where trust is effectively extended to strangers. Looking at trust from a different perspective, we can think in terms of trust among equals, where coercion cannot really be applied, as opposed to trust where coercion can enforce behaviour. The traditional scholarly publishing domain works largely through personal trust and there is the possibility of the use of coercion. The internet works largely in the sphere of strangers and there are few means to know whether something is trustworthy or to enforce behaviour.

Geoff argued that the success of eBay, Amazon and Google is partly about their understanding of the importance of trust. Right from the outset eBay thought about how to ensure that people would trust the mechanisms that they had for buying items online – they built in a trust metric. Amazon implemented uncensored reviews, lowering the risks of buying something online. Within Google the page rank is an implicit trust metric that works extremely well.

Web 2.0 is very much about trust. We can pretty much subscribe to someone else