Inside intelligence: approaches to business intelligence

Analytics are an increasingly important tool for business.

Mining data and using analytic tools to derive business insights have become key components of the decision making process, and companies like Google and Microsoft use them to define products and plan future developments. With users driving the demand for more and more information, BI has a distinct place in modern business architectures.

As BI's role and importance grows, you're going to need to consider how your BI strategy will evolve over the next few years. Things are changing in the datacentre, with a shift from on-premises to cloud-hosted operations — and the resulting change in how we think about, and how we manage security. User needs are changing too, with more and more demand for mobile access to information and services.

Information-centric security

With the shift from on-premises datacentres to using cloud services, there's a need to change from server-centric to information-centric security models. Often referred to as "re-perimiterisation", it's an approach that ring-fences your data and information, rather than trying to protect entire networks.

With the shift from on-premises datacentres to using cloud services, there's a need to change from server-centric to information-centric security models.

With information-centric security there's no need for expensive (and often slow) data-loss prevention systems, as use of role-based security at a record level means that data can be controlled and audited effectively.

Modern BI tools and services can take advantage of this deeply integrated security model to control just how users have access to raw data, and how the resulting information and reports are distributed.

The same approach can also help you manage working with distributed data sources across a heterogeneous computing environment — an approach that is likely to help in any transition from on-premises to cloud-hosted BI.

External data

BI needs data, the more the better.

Much of the information you'll need to get the most from your data is in external sources and needs to be imported into your systems. You can use the demographic data in tools like Microsoft MapPoint as part of your analytics framework.

Much of the information you'll need is in external sources and needs to be imported into your systems.

New open protocols like OData mean it's easier to bring in external data, and you'll be able to work with large-scale third-party information sources like those provided by the Azure DataMarket. The arrival of services like this gives you access to a mix of commercial and government data that can be used to add value to your business analytics.

Adding Dun and Bradstreet data to a sales BI application will give users additional insight, while the European Environment Agency's Carbon Emission data could add benchmarking features to a corporate sustainability report.

Integrating cloud services

Integrating cloud services like those provided by Azure DataMarket is going to be increasingly important, and needs to be considered carefully when implementing your BI security model. One option is to use federated identity services to ensure user identities and roles are shared with information partners, maintaining an audit trail even across application boundaries.

Integrating cloud services like those provided by Azure DataMarket is going to be increasingly important.

This approach also means that you'll be able to ensure that security contexts can be maintained and managed between desktop, datacentre and cloud provider — something you'll need to stay in compliance with business regulations. Securing your information this way also simplifies moving BI services from your datacentre to cloud services like Azure, without worrying about changes in security models.

Cloud-hosted BI will also allow your users to...

... work with more complex models and analytic functions, with on-demand compute resources able to scale to handle large analyses.

In the near term, BI software will add more complex features, including the ability to work with unstructured information sources, and these functions will require more compute power. Using in-memory processes in cloud servers will allow these services to scale and still deliver the performance users expect.

Mobile users

Having an information-centric model also simplifies delivering BI to mobile users.

Web-based reporting and data-analysis tools are increasingly commonplace, and with the latest generation of smartphones finally giving users reasonable quality web-browsing experiences, users are going to want to work with BI tools while on the road.

Users are going to want to work with BI tools while on the road.

Larger form factor tablets are also likely to increase demand from mobile BI. Gartner is currently predicting that 33 percent of BI functionality will be consumed by mobile users with smartphones and tablets by 2013 — just two years away.

You'll need to consider carefully how to offer this access, whether using a native application on supported devices, or providing an access-controlled secure website (or alternatively configuring mobile devices for secure access to intranet resources using a VPN). Mobile access to BI resources is likely to work well with cloud-hosted BI solutions, as you will already be using mobile-friendly authentication methods.

BI culture

Like any IT function BI is as much about people as it is process. You can implement all the BI tools and services you want, but without user uptake they'll just be so much 'shelfware'.

You can implement all the BI tools and services you want, but without user uptake they'll just be so much 'shelfware'.

Training your users is only part of building a business-intelligence culture, and without an organisation developing that culture any deployment is likely to stagnate after an initial surge of interest.

Treating BI as a decision support service is one approach that can help, bringing together BI with familiar desktop tools (perhaps using the analytic features of Excel 2010) and social software.

Services like Salesforce.com's Chatter, which lets documents and reports be active members of a conversation, can help push results around a business, encouraging increased use of services and further sharing of BI-led insights.

Culture doesn't just permeate from the bottom up, it needs examples and leadership to grow. You'll need executive-level sponsorship to drive the cultural changes that BI needs, especially if you're planning to use it in conjunction with collaboration tools, delivering what Gartner calls "fact-based, transparent decision making".

Sponsors need to be ready to drive the changes that are needed, and be able to support BI as part of an alignment of IT and business strategies. Starting with specific projects will help encourage users, and at the same time demonstrate the value that BI can add to business processes and planning.