In search of retail search

Profile picture for user mbanks By Martin Banks February 4, 2015
Retailers’ exploitation of search services has barely advanced an inch for the last 15 years and that it’s time to smell some new coffee, if they can find it.


When it comes to the retail business, is search engine technology letting users down, or is there a problem with the way that either half of the equation – buyer or seller – are approaching search and the way it is used? Would it help if retailers and search service businesses talked to each other?

Such questions formed the platform for a recent panel discussion hosted by Cloud and hosted services provider, Rackspace. With an estimated 40% of end user searches now being conducted on mobile devices, a share that will only grow, there is a need to reconsider what searching is about, what consumers are looking for and why. That in turn leads to questions of how search services, and the technologies that underpin them, may need to change.

The panel assembled by Rackspace consisted of the company’s own VP of Technology, Nigel Beighton, the VP of Technology at Elasticsearch, Mark Harwood, Tony Duffy, the eCommerce Manager at Oddbins, Chris Harris, Technical Director at Hortonworks, and Peter Owlett, a data scientist working with consultants, Capgemini. It was moderated by Paul Miller from Cloud of Data.

There was a general consensus amongst them that the dominant requirement now is to provide search services where users can expect to receive responses that fit the key criterion of being `appropriate for me’, and appropriate within the context of where `me’ happens to be, what `me’ is searching for and other factors such a 'me’s’ recent purchase history.

It is certainly true that users can learn to tailor their search arguments better than they do, but that does require the development of a fair degree of skill in using particular search tools. And given the amount of data now available, it is misplaced to consider the way search services have developed so far as being the fault of users.

Same old same old

The real problem is that the retailers perceptions of how search services can advance their business prospects have not significantly changed over 15 years. As Beighton put it:

I worked for in the 2000s, and I look now at the systems, and they haven’t moved on. What I see on the mass majority of e-commerce sites is that all they’re left with is the ability to put up some product or stock and hope that you choose based on the price, having spent a lot of money of Google Ads to get people there.

He then gave an example of how he would expect searching for hotel rooms to now work:

As a family person I would desperately love to be able to book connected rooms online. I have teenagers who I wouldn’t trust in separate rooms to me, what with the TV and the minibar. Please, somebody tell me where I can book connecting rooms, because it seems to be nigh on impossible.

This, as all the panel agreed, is a role that can, and should, be filled by big data analytics. But to date this is not happening to any tangible extent. As Harris observed, most retail businesses don’t think that way yet about the information that they make available. There is still a strong tendency towards siloing data in ways that may well be useful internally to the business, but can be of little value to search services that could help expand the business.

Browsing the shelves

Yet it is through those analytical processes that the wider context of a consumer’s need – which will be a mix of the retailer’s own data plus that from third party aggregators and partner businesses – will be generated. This would stop the current trend for retailers, once knowing that a consumer has bought something expensive such as a car, camera or guitar, promptly bombards them with blandishments to buy another one, rather than something complementary.

One important factor here that will help the process in future is that, despite the generally high level of FUD used around consumers making personal data too widely available, younger consumers are now more liberal and relaxed about their personal data. This suggests that they appreciate the benefits of the `trade’ between giving their data to retailers and data aggregators and the better level of service they get in return.

Consumer perceptions have also changed in terms of their experience with using search and other retail communications services, as Duffy noted:

Five or six years ago, if I’d been sent an email by a retailer thanking me for my order, I’d have been annoyed if they’d called me ‘Tony’. Nowadays, ‘Dear Guest’ sounds worse and we feel a bit hostile towards those who don’t engage with us on first name terms. The industry has definitely moved on and we haven’t kept up to date with that.

Talk to me

Underpinning this issue is what the panel saw as a general lack of communication between the retailers and the technology companies, with retail CIOs being identified amongst the chief culprits. Beighton suggested they are missing an important trend in the way IT developments are now discussed, formulated and created.

This is the growth of communities of both users and industry representatives, where the needs of one and the capabilities of the other can be stirred in the same melting pot. This is already increasingly common in the open source applications area, with Beighton citing the existing Hadoop communities as ideal places for retail CIOs to learn what is possible with big data analytics and how they might exploit it.

It was noted by Miller that company boardrooms are changing their views on technology application, not least because CIOs are at last starting to move on to be CEOs. Even if only through their growing use of mobile systems their appreciation of what makes and good service for customers continues to grow. Through using good and bad services in their own lives they become far better equipped to assess the service provided by their own business.

There is also, it seems, a growing appreciation that the success of any retail sector as a whole has a benefit to all its players, and seeing `competitive advantage’ in too much business information can be counter-productive. According to Beighton, people are now starting to share models of how they do things such as searches.

My take

It looks as though both sides of this equation have fallen into the trap of thinking that improvement in the responsibility of the other. When I was a lad, parents used to say things like `don’t sit around, join a club’. Maybe it still holds true and they should all get out there and get stuck in to relevant communities to see what they can learn, and even contribute.