Rackspace #PowerOfSearch: defragging the data supply chain with contextual intelligence

As you will know, journalists love nothing better than a really early start — as such, the global PR industry is fond of staging what are known as ‘power breakfasts’ now and again.

So it came to pass this week that Rackspace hosted another in its series of coffee- and croissant-fuelled get together(s).

Rackspace #PowerOfSearch

1 raxspace.jpeg

Attendees at this event included: Nigel Beighton, VP technology, Rackspace; Chris Harris, VP of international at Hortonworks; Mark Harwood, a developer from Elasticsearch; Peter Owlett from Capgemini and Tony Duffy, e-commerce manager at Oddbins.

Beighton contended that most ecommerce websites have move on very little (or not at all) since the turn of the millennium.

Silos vs. centralised architectures

The problem here (the speakers suggest) is down to fragmentation and the fact that the “data supply chain” fails to exist as one single solid stream because it has not been unified.

This issue is further compounded by the fact that elements of IT sit so separately from each other:

• transactional processing sits apart from…

• analytics which sits apart from…

• bricks and mortar IT… and so on

According to a recent survey from Rackspace, almost half (45%) of UK consumers actually prefer to shop on the high street instead of online.

“Frustrated by long winded search functions and too much choice, over a third of shoppers will give up browsing a website after just 10 minutes if they can’t find what they want. This demonstrates consumer frustrations at search capabilities and retailers inability to use big data to offer a truly useful experience online, beyond the best price,” said the company, in a press statement.

Looking routes to defragmentation

So looking at data streams now and thinking about how software application development professionals will code to a data landscape that is fragmented with too many imprecise undefined elements – where do we go next?

Rackspace’s Beighton argues that much of the challenge comes down to TRUST – and the question of where we are happier to provide access to personal information.

Contextual intelligence

Computer Weekly technology editor Cliff Saran suggested that he gets frustrated with search with regard to Amazon and the fact that the web services driving the site’s offers keep promoting products to him that he had already purchased.

Surely the solution here comes down to contextual intelligence:

• If I buy an exercise bike (a high value item that may be a once in a lifetime purchase) then I should not see ads for this item again.

• If I buy a birthday present, then the systems used should offer the option to know that it is a once a year purchase.

• If I buy a food item or other similarly regularly repeatable product, then flagged promotions are more permissible

The lesson here for CTOs is one that should make them look back at their own data supply chain and their internal approach to information share – and we could be talking about in-company data usage.

Coming back to Beighton’s point… we can not always expect all of this information to be available and it comes down to trust — and perhaps, privacy, identity and security.

This debate may have highlighted some of the issues impacting search, the trouble is — even if we do know how to make things better, privacy and trust may represent a barrier here.

The Chief Data Officer’s job will always be a tough one.