The changing face of business intelligence

As business intelligence tools are becoming more popular, vendors are offering them in different forms, on different devices and through different models.

Business intelligence (BI) is becoming more pervasive, said Yellowfin CIO Glen Rabie, noting that BI features are showing up in a widening range of packaged software and that organisations are becoming more aware of the value of data. And where BI has traditionally been used internally, changes in licensing models mean it is increasingly showing up in customer or supplier-facing dashboards - “a huge shift”, according to Rabie. For example, one of Yellowfin’s clients uses the product to provide dashboard-style report visualisations to its customers.

There is also growing awareness of the value of using data to predict what may happen and to decide which levers to pull in order to get the desired outcome, said James Foster, head of marketing at SAS Australia & New Zealand. He pointed to examples such as a Charles Sturt University project using sensor data to predict the best time to pick wine grapes, a logistics company re-using data needed for reporting purposes as an input to its quoting and route planning engines, and the Hong Kong government identifying patterns in its call centre data in order to proactively improve services.

Trevor Legg, APAC program manager for BI at Oracle, added the example of Oracle customer Ambulance Victoria, which is using the company’s BI tools to perform real-time analytics to route patients to the most appropriate hospital.

Self-service

The desire to provide self-service environments is widespread among IT departments, Legg observed, and this applies to BI where users want to be able to build their own reports with little training rather than assigning the tasks to specialist analysts. Oracle’s BI tools are role-based to suit various jobs, and the data available is described in everyday business terms (eg, ‘revenue’) rather than database-specific terminology. Consequently, self-service accounts for up to 90% of BI use by the company’s customers, freeing up analysts to explore the data - to “uncover gems within the data” - rather than answering users’ questions.

QlikTech’s QlikView is used by Inner East Melbourne Medicare Local (IEMML) to allow clinicians and other non-IT staff to explore data without needing specialist training, said James Belsey, lead solution architect at QlikTech ANZ.

Jason Ferriggi, data manager at IEMML, said, “Since deploying QlikView, we’ve seen major improvements in how we’re able to serve the local community. Through the software’s user-friendliness, our staff are saving between 20-30 hours per week on reporting and we are able to quickly pull valuable insights from community generated data, something that wasn’t possible previously.

“As a result, we’ve been able to improve the way we use government funding to support the health of over 680,000 residents of the inner east Melbourne region. We are also providing two health services and 170 general practices in the area with real-time data to assist them in identifying potential gaps in their delivery of healthcare services.”

David Norman, national BI practice manager at NEC, also observed this trend to enabling non-technical staff to do their own BI tasks, noting that it avoids the long lead times and expense of big deployment projects and allows a more agile approach to BI.

Conrad Bates, managing partner at C3 Business Solutions, noted that current tools (which he refers to as ‘BI 2.0’) are more lightweight than their predecessors such as Cognos, MicroStrategy, BusinessObjects and OBIEE.

“The BI 2.0 tools definitely have less ‘clutter’ in the user interface but they do have some limited functionality compared to the original generation tools,” he said, predicting the emergence of analytics apps for particular purposes such as sales reporting, with user interfaces that are specifically tailored for that function and therefore more suited to self-service BI.

Mobility

“BI is going mobile,” said Paul Goepfert, marketing manager at Pronto Software. This is in line with a general shift towards mobile technology as people expect to use IT wherever they need it. Most use of BI will be on mobile devices within a few years, he said, “I’m in no doubt about that.” Consequently, Pronto has adopted a ‘mobile first’ approach to development with an emphasis on user interfaces that can be used without training or manuals, and on providing ‘task intelligence’ that draws attention to the most important information.

Suhas Uliyar, vice president mobile strategy & product management at Oracle, agreed, saying, “Mobile BI is a great way for enterprises to get onto a mobile journey,” while Foster described mobile access to BI tools as “critical”. But Alec Gardner, general manager advanced analytics at Teradata, warned that mobile BI tends to increase the volume of queries, so it is important to make sure that the system has the back-end capacity to support the additional load.

Big data

There has been a lot of talk about big data, and according to Belsey it is starting to become a reality in Australia. One example, he said, is the way Candy Crush company King.com stores massive amounts of player information in Hadoop and then analyses it with QlikView.

Teradata customers have been doing what are now called big data projects for a long time, Gardner said, with banks, telecommunications carriers, logistics firms and government agencies using the technology to better understand their customers/clients, reduce risk and fraud (eg, the ATO used Teradata systems to identify around $130 million of fraudulent claims in a single year) and to improve operational efficiency. Using BI tools in real time to improve efficiency is a continuing trend, he suggested. But “having the holistic view of a person [ie, customer] is important”, he added: a mobile phone user experiencing a single dropped call probably does not consider it a big deal, but it is a different story if three calls are dropped in the same day, especially if they all happen in the same place.

Cloud

Like so many aspects of IT, BI is starting to move into the cloud, and even where an on-premises deployment cannot be escaped it is possible to minimise entry costs by selecting a provider that offers an OPEX model, Norman said. There has never been a better choice of BI tools, and they are more affordable than ever, he added.

But Bates noted, “The challenge that remains is getting a small organisation’s data into a structure that will enable the use of these cheaper tools. This integration of their data would traditionally have been an ETL (extract, transform, load) platform in a data warehousing/business intelligence implementation.” So C3 has created a hosted data integration platform and now works with SMEs on hosted data integration from multiple applications and services, and BI reporting.

Built-in, add-on or third party?

There are three basic ways of obtaining BI tools. They may be baked into software you already use (eg, Pronto Xi or SQL Server 2012), offered as add-on modules (eg, Oracle Business Intelligence Enterprise Edition, often abbreviated to OBIEE) or sold by third-party developers (eg, Yellowfin). Naturally, each camp has its own view of the merits of the three approaches.

“Business software without BI is now dead,” asserted Goepfert (Pronto Xi includes IBM’s Cognos software, which was one of the early leaders in the BI market) as people increasingly want to understand what is happening and work intelligently rather than blindly following processes. They are accustomed to the graphical display of information on smartphones and also to being presented with significant amounts of information (eg, sports coverage on TV), and they “are thirsty for the ‘why,’ not only the ‘how’”, he said.

One advantage of Pronto’s approach is that it provides a middle layer between the ERP data and the BI tools, making the latter easier to put to work. Some vendors provide ERP and BI but not that middle layer “because it is hard to do” and provides an opportunity to sell consulting services, he said.

However, Legg said that Oracle’s BI tools incorporate a world-class data model and provide many built-in KPIs, so its customers do not face a big consulting bill for implementation.

Pronto spends around 25 to 30% of its R&D budget on BI, Goepfert said, as its predominantly mid-market customers would not accept the consulting costs associated with traditional BI implementations. With Pronto, “80% of the BI analysis people want to do is actually there and available” in the form of content, cubes, dashboards, reports and indicators that are available straight out of the box, he said, allowing those mid-market customers to gain immediate benefits.

Brent Dunning, general manager of specialist tool supplier Gasweld, said the BI tools provided by Pronto were very easy to use and configure. The company previously had to produce digests of reports for use by senior managers, but they now received automatically the precise information they required. For example, it used to take an employee half a day to produce Dunning’s weekly backorder report, but it is now generated automatically and the staff time is put to better use. And the enhanced reporting to store managers about customers that are on credit hold has improved efficiency and cash flow.

Dunning said new reports are written at Gasweld every day, cutting the white noise and giving people the information they actually need. This process is helped by the way data relating to accounts, payroll, inventory, manufacturing and other areas can be assembled into meaningful reports, he explained.

That experience aligned with Norman’s, which is that the main benefits of BI projects include replacing human-driven reporting (eg, consolidating spreadsheets) with automated processes and thereby providing the time for more productive activities, exploiting data that has not previously been combined (eg, combining HR and ERP data to reveal true performance) and providing a more accurate and less confusing view of the organisation’s data.

Rabie suggested that most ‘built-in’ BI tools were not designed-in by the companies selling them, but were instead acquired (either as products or as part of a takeover) - fine if what you get fits your requirements, but these tools may not work with all the applications in use within the organisation, he warned.

Best-of-breed BI vendors are more focused on that technology than more broadly based vendors are, and so are more likely to respond to customer requests for changes and new features, said Rabie: “You do tend to work harder for the business.”

Furthermore, there is greater pressure on such vendors to be technology agnostic and willing to connect to “anything and everything”, he added.

Foster pointed to a Forrester Research finding that ‘best of breed vs suite’ was no longer a major selection criterion for BI software. “BI has matured and standardised,” he said, claiming that SAS software is very well known, integrates easily with a wide range of other applications, and “provides richer features and functionality” than built-in BI tools.

SAS has advantages in three key areas, he said: data management (using multiple sources, whether they are internal or external and structured or unstructured), reporting (including the ability to embed advanced analytics into reports) and gaining insight (and perhaps even more importantly “making the insight actionable” by delivering it to the right person at the right time).

“Business discovery is really what we do at QlikTech,” said Belsey. “It’s all about focus on the business user” and empowering them to get answers to questions about data quickly and independently, while enforcing appropriate access controls.

ERP systems typically provide bundled reports that answer the common questions, but to get full value from their data, organisations need to combine information from multiple systems. One customer - BT Financial Group - was spending between 250 and 300 hours a month to assemble data from multiple sources, but after adopting QlikView that was reduced to less than 10 hours while gaining greater insight. These multisource reports can be constructed very quickly with QlikView, sometimes in hours or days rather than weeks or months. Belsey said that Amcor was able to combine the data from two very complex systems in just 25 days.

Legg suggested that “specialist” tools were suitable for reporting rather than BI, saying that Oracle’s strategy assumes customers are looking for high-value strategic BI aimed at changing the business. So the strategy is based on three pillars: optimise processes around organisation-level KPIs (not those relevant to an individual department or function), simplify the technology (eg, via the company’s Engineered Systems approach which provides complete stacks from application to hardware as a single product) and innovate (eg, predictive analytics, mobile BI and support for unstructured data).

Oracle’s BI product originated with Siebel, he explained, but around four years ago the company decided to make it the standard for reporting across its product range. Consequently, some functions are present in each suite, but the various Oracle BI tools bring it all together. Furthermore, integration with non-Oracle applications is not only possible; Oracle’s pre-built data model means implementation can be significantly quicker than with some other products.

Gardner pointed out that Teradata’s systems are designed specifically for decision support and analytics. There are always multiple sources of data, he said, so Teradata takes care of gathering, consolidating and then analysing it.

The company does partner with companies including Microsoft, Oracle and SAS, and their tools can be used to access Teradata stores.

Of all the people we interviewed for this article, Norman had the broadest view as NEC implements systems for its customers using products from a range of vendors. While IBM, SAP, Microsoft and others have traditionally provided “pretty complete systems”, he said, organisations are increasingly calling for a multivendor approach and focusing more on business rather than technology issues.

Image credit: ©iStockphoto.com/polygraphus