Cloud business analytics full of promise but hitting barriers

Cloud analytics is expected to grow more rapidly than on-premises BI. As yet it confronts technical, performance and security barriers, but users like Cipal and Mindshare prove the concept.

As businesses struggle to process ever-larger volumes of data, the uptake of cloud-based business intelligence (BI) is expected to grow rapidly.

Figures from market research firm IDC suggest that cloud analytics and BI deployments, though still a small part of the business intelligence market as a whole, will grow at three times the rate of conventional, on-premises BI systems through 2013.

IDC's most recent survey of the market found that half of businesses were considering BI in the cloud -- a far higher level of potential take-up than for cloud computing technologies overall.

Interest in cloud analytics is being driven by two main factors. First, companies are gathering and storing more data and want to make that data -- and BI reports -- available to more people in their organisations.

Cloud analytics extends access to staff working on mobile devices, and potentially to suppliers and customers. It is also more "elastic" than conventional IT systems and can more readily offer more, or less, storage and processing capacity as required. That helps businesses deal not only with larger volumes of data overall, but also with data analytics tasks that have significant peaks in demand.

The cloud analytics cost advantage
The second factor is the way that cloud-based analytics is reducing the cost of using business intelligence systems overall -- and thus bringing BI within reach of smaller businesses and startups.

Using BI on demand, or through a cloud-based Software as a Service offering, reduces the up-front cost and possibly the total cost of ownership for an analytics system. Cloud systems also make financial sense for one-off analytics projects, as there is no ongoing cost of hardware or software once the work is complete.

As a result, conventional BI and enterprise software vendors, as well as a number of startups, are offering business intelligence in the cloud. In addition, some technological developments, including Google's MapReduce framework for distributed data processing and the open source Apache Hadoop technology, are making it easier for vendors to build large-scale data analytics engines that take advantage of the scalability of the cloud. Systems such as Hadoop also make it easier to process and analyse unstructured data, such as information from documents or multimedia content; such content accounts for anywhere up to 80% of typical enterprise data, according to analyst estimates.

The cloud BI challenge
IT leaders, however, have to balance the advantages and potential of cloud-based analytics with the relative immaturity of the marketplace and some of the practical barriers. Given the right skills, a business starting with a blank slate could well build analytics capabilities in the cloud. For a business looking to drive greater efficiencies and effectiveness from an existing installation of analytics tools, transferring workloads to the cloud will be less straightforward.

"One immediate barrier is the complexity of the analytical models," said Greg Todd, an executive director in Accenture’s analytics practice. Although cloud-based BI tools might have comparable capabilities to on-premises software, there will be sufficient differences at the technical level to require models to be rewritten, he suggested.

Nor can the cloud always match the performance of an on-premises system: businesses are increasingly using data warehouse and analytics appliances to help boost throughput, but few cloud services currently offer these. Businesses also need to consider data security and privacy and whether there is sufficient bandwidth and time to transfer significant quantities of data to the cloud provider for processing.

That will depend as much on the workload, and the urgency of the analytical reporting, as the technology itself -- something chief information officerss and BI managers will have to assess in the context of their own businesses.

Nonetheless, a "maturation of the market is underway," Todd said. "We are starting with basic information management and moving to predictive analytics and full-blown decision processes."

Cloud analytics for social welfare provision
Public sector IT provider Cipal -- which is run for a consortium of local authorities and government agencies in the Flemish portion of Belgium -- is using cloud-based analytics to help officials analyse and benchmark social welfare provision. The analytics service runs from a central data warehouse and is part of Cipal’s strategy to offer all of its applications in the cloud.

For data analytics, using the cloud is a practical solution for public sector organisations that lack the scale, expertise or budgets to deploy relatively complex technology on their own, according to Sven Meermans, Cipal's business development manager for business intelligence.

"Many of our customers, especially the smaller municipalities and social welfare bodies, do not have the know-how, the budget or the infrastructure to set up their own BI solution," Meermans said. "We set up a centralised data warehouse at Cipal that enables all of our customers, not only the big organisations, to benefit from a complete BI suite."

Deploying BI through a private cloud overcame both the skills and resources problems for government agencies, as well as making it easier for their users to access BI data and reporting through a Web browser, he added. Cipal also chose open source BI technology, to allow for more flexible future development.

A further benefit from using a cloud-based infrastructure is that Cipal can help local authorities to benchmark their services. By holding information in a central data warehouse and using standardised key performance indicators for services, it enables users to compare their performance with other municipalities.

Local authorities, though, have historically not been heavy users of business intelligence, not least because of the cost. With Cipal's cloud-based system, Meermans expects that to change.

"We realise that BI for most local government, for the moment, is a ‘nice to have,’ but we are convinced this will evolve to be a ‘need to have,’ and we want to be ready," he said.

Private cloud for data analytics
Mindshare UK, a media planning and buying subsidiary of advertising and marketing services firm WPP, also operates data analytics using a private cloud. The company takes a range of information, including planning data such as an advertising outlet's pricing and reach, as well as audience evaluation and emerging social media feeds, and makes them available to clients through a Web browser.

Using a general-purpose BI tool -- as well as a cloud delivery model -- has a number of advantages for both Mindshare and its clients, said Andrew Corroll, head of data integration at the company. For example, it removes the need for clients to purchase and install specific media monitoring software packages, and it simplifies the security issues around allowing third-party access to data.

"We can communicate [our data] with people without them having to download software to a machine, and they can access our reports from anywhere, via the Web," Corroll said.

Planned upgrades include bolstering the in-built analytics capability of the BI and adding shared third-party data to the system, so clients can compare their advertising campaigns with industry-standard metrics, without the need to leave the Mindshare application.

Read more on Business intelligence and analytics