Five New Zealand enterprises share their lessons on the deployment and effective use of business intelligence tools.
Meridian Energy: Guarding against data silos
While other IT projects start and end around parameters related to fixed deliverables, BI and data warehousing needs to be an ongoing process and will tend to operate on a lifecycle, notes Peter Simcock, business intelligence manager markets and production for Meridian Energy. He says the reason BI projects and strategies cycle is partly because the people in the business change. As new people question current approaches, if IT hasn’t developed a methodology that connects the BI tools to business needs, IT will disassociate the data warehouse from users again by ceasing to be an enabler.
“The structure can be all good, but after a while users will find holes in it. So when new data sources turn up, there can be a big delay in moving that data to the data warehouse so it is accessible to users via BI and reporting tools. The risk of having to repeat the cycle occurs at about that stage and the business starts to build data silos again. To avoid the cycle, the IT function has to be more agile,” says Simcock.
He says BI initiatives are more likely to fail cyclically when IT is asked to make BI and data warehousing a set deliverable. Instead, data warehousing should be run by the business because that’s where the demand, use and value exist.
Meridian Energy has traditionally used a mix of Business Objects 5.12, Excel, Access Databases and TOAD over Oracle for data warehousing and BI. Together with Eagle Technology, Meridian recently upgraded to the Business Objects XI platform, which Simcock says streamlines the modeling output distribution, data management, and prototyping and report development for a wide range of functions including retail, operations and outage planning. Around 100 Meridian employees, including modelers, data information and business analysts and analytical decision makers now use the system to monitor up-to-date sales activities, conduct customer analysis and pricing functions and produce KPI reports.
Simcock says the all-important internal champions contributed to a smooth upgrade path and early benefits for Meridian, many of which are intangible.
Champions occur at different levels of the business, he says, and some people can be so far removed from a business problem there’s no point in asking them to champion the solution. At the same time, champions need to be senior enough to have authority or to be able to act as an advocate for the solution.
“I would say champions tend to come from Level 3 and 4 managers, depending on the size of the business. Champions set the vision while advocates underwrite it. Champions need to have a grasp on what’s needed at a management level, but can’t lose sight of what’s needed at grassroots level,” says Simcock.
Technically, all BI products seek to place a layer between the database and the user; whether that’s a Business View, a Universe, a EUL or a simple cube, says Simcock. In the past five years there has been a move to a graphic user interface and web delivery; tools have become more process and object orientated and the language or interaction more sophisticated behind the scenes. But BI tools also need to serve the power user, who still needs access to the data, as well as offer a front-end drag and drop for the business-focused analyst.
“Businesses are now crunching a lot more data and complex data so BI tools have evolved to reflect that. My belief is that you shouldn’t give everyone access to the data source, because not everyone needs a lot of data to do their job; they just need information and the two are not the same thing.”
Fletcher Aluminum: Multiple sources, one truth
Faced with a move deeper into the global export market, aging business intelligence tools, 14 different data systems and one main ERP system, Fletcher Aluminum sought a BI system that would incorporate new reporting features and interrogate multiple sources for one instance of a report. The goal: To create reports that ensured delivery of customer orders in full, on time, and according to specification.
“We were also looking to retain investment and knowledge in our previous Crystal, AS/400 Query and Cognos systems, and for a reporting tool that would embed as far as possible into our ERP for one set of user access and password authorities. The system also needed to deploy across intranet and internet across Citrix, so our people can log in from wherever they are around the world and run the reports they need,” says IT manager Trevor Jones.
Fletcher Aluminum uses a Lawson M3 ERP system and Lawson M3 Business Performance Warehouse tools. After evaluating other products on the market, the company decided on Lawson Business Intelligence (LBI) tools that contained the Crystal Reporting system developed by Business Objects. Jones says implementation was fairly rapid, with the first report being produced in two weeks.
“We covered off a lot of the basic reports and bloodied our hands on those and built up a raft of questions,” says Jones.
Graeme Gallagher, senior business analyst for Fletcher Aluminum, says Fletcher business managers are committed to ensuring data comes from one place, so that there is ‘one truth’ and one particular path to follow.
“A lot of KPIs and key reports have been completed with a data analyst sitting next to a key member of the executive management team,” says Gallagher.
Despite this, Jones says initially IT management had a “nice fluffy idea” that IT would not keep a stranglehold on reporting because it didn’t want to.
“It’s more work for us. But we found that having a model where anyone can bang up their own reports wasn’t getting us to where we needed to be. That’s not because LBI and Crystal are difficult tools to use, it’s because people [lacked] a good knowledge of where to go to find the information,” he says.
Like Meridian Energy, Jones says Fletcher Aluminum has observed a cyclical nature to BI strategies and tools.
“That is absolutely what happened to us; we had AS/400 Query going back 10 years and that developed all sorts of limitations. Then Crystal had some limitations on what you could do with data and we started using Cognos. But you then have to maintain knowledge for all three systems and pay licences. If we use different tools to interrogate the same databases, we get different results and it becomes a case of what is the right version of the truth here? Because LBI can look at all the main versions of the databases I think we will get a little more longevity out of this; it will be okay for the next five years,” says Jones.
Gallagher says Fletcher Aluminum takes a balanced approach to engaging with BI and data warehousing vendors, requiring a full product presentation using Fletcher Aluminum’s own data “to prove there is no smoke and mirrors”.
“We did a lot of interrogation and [asked] hard questions right up front and it took us a long time to validate that this was going to be the right solution. Once we had made the decision though, we were very open; we said you know your database and what your system can do. Here is a set of common reports we will need; let’s go through and develop these together.”
Jones says while the usual intangible BI business benefits are present, the IT team has also been able to show the business the importance of quick and accurate information to avoid different versions of the truth — margins on products and orders being just one example.
“The system change allows us to know the volume base of our business; we get 2000 work orders a day through our factory; and a lot [of information] could get lost in the haste. So our BI tools need to be looking for exceptions in orders — for example failure to meet the required volume,” says Jones.
Auckland Regional Council: Critical engagements with key staff
Data important to the Auckland Regional Council (ARC) resides within a wide SAP processing environment, making pulling accurate information together for decision making a challenge, particularly when it has to be done manually, says John Holley, group manager IS for the ARC.
“It’s one thing to present information at the right time and another to have it automated. We have more automated project management solutions on the way, but unlike commercial organisations our outcomes are driven by the public so our business intelligence can never be a static solution,” says Holley.
The ARC recently implemented SAP’s SEM CPM suite — BI and corporate process management tools that deliver balanced, scorecard-corporate performance reporting, collection and management of manually entered data, 86 measure definitions, a strategy map to be used by all scorecards and end-user training and rollout support.
Holley says while ARC previously had an established corporate performance reporting process and standard output, compilation of the monthly performance report was manual, time-consuming and the resulting output was paper-based. Opportunities identified to improve the process included the need for business unit managers to be more responsible for data entry and data management, as well as narrative development, and for the decommissioning of spreadsheets. The organisation sought implementation of a true balanced scorecard, reduced reporting cycle times based on ability to parallel process and reduced waste through elimination of paper.
Holley says many organisations have hundreds of reports within their IT systems; some used only once or twice. Examining business intelligence requirements includes an understanding of how the business warehouse talks to all the other systems in the business — and this can get complicated; around 50 per cent of the ARC’s information is held outside of SAP systems in separate systems. Service Oriented Architecture (SOA) is needed to pull all systems including portal environments together, says Holley, and ‘Corolla’ solutions built to SOA and portal and database standards can be better than ‘Rolls Royce’ solutions.
“Work to standards so you can easily connect up; this might mean [slightly] compromising on what you need, but define what the must-haves are. [New] BI solutions will then have to support SOA standards and plug into the infrastructure and connect to [existing BI] systems.”
In approaching its BI system, the ARC engaged key staff from business areas including finance, with Holley saying he can’t understate the importance of ensuring business stakeholders are heavily involved in a BI project — internal champions may need to be released from their everyday tasks for several months, meaning the business benefits of the project have to be spelt out up-front.
“BI projects run adrift when key stakeholders are not released. If this happens you often get a 50 to 60 per cent solution that never quite meets people’s expectations and fulfils compliance requirements only,” says Holley.
He agrees with his peers that to some extent data silos within a business are inevitable, because the selected system can never meet everyone’s needs.
“We can’t do it all for them, there is just too much. And we don’t want to be big brother and say all information has to be centralised. It’s one of the reasons we are putting in a straightforward web application server, so that our people can create SQL databases on the fly,” says Holley.
As to measurement, Holley says there’s a need to establish KPIs around BI processes and that measure will be people’s confidence about the information and how well they add to it and use it as a planning tool. He says user-satisfaction surveys and ICT governance surveys can also help establish the success of a BI initiative.
ASB BANK: Treat it as an ongoing project
Following establishment of its first data warehouse system in 1993, the ASB Bank has traditionally had strong business capability in business intelligence and has observed the need for the business and technology teams to work together. Early on, it sought to ensure all new information was entered into its data warehouse, but struggled to find the right tools to ensure information could be extracted to meet business strategies — although business analysts accessing data and taking the information to the wider organisation wasn’t a concept. In 2005 the ASB developed a large group of data analysts to make sense of business data; however this resulted in the development of data silos for separate business areas and duplication of information. In 2007 the bank revised its BI strategy, adapting and adopting user-familiar Microsoft tools to deliver organisation-wide access to business data and information.
Peter Muggleston, acting head of technology operations for the ASB, says for a number of years BI vendors have attempted to move away from providing tools for the analyst and executive, to address the issue of disseminating information across all levels of an organisation. The latter reduces the training overhead and ties with a trend towards the use of business intelligence in an operational and performance measurement sense, rather than a strategic or analytical sense alone.
“As this drive continues, internal processes will need to recognise the value of information as a corporate asset, and drive change through managing common data definitions and data quality, and embedding data ownership and stewardship as an ongoing practise,” says Muggelston.
He agrees the concept of removing all data silos isn’t realistic, and says removing all data repositories should not be the end goal in itself. Instead, organisations should aim to ensure information has an organisation-wide context and is defined and treated consistently. There may be little value in consolidating data exclusively used by a single line of business into ‘one view’, says Muggleston, but where data and its definitions need to be shared, reused, or combined with other data there’s value to organising it centrally.
Echoing a familiar mode of thinking among other IT executives, Muggleston says it needs to be clear at the start that BI will be an ongoing programme rather than a single project.
“The best value for an organisation with a limited amount to spend is not to necessarily invest it all in software tools. There is more value in understanding, shaping and growing the information the organisation owns. To a large extent this is always a work in progress, but there is a critical mass that needs to be reached before the tools used to present and deliver that information can realise any true value,” he says.
He says business champions need to be sought at all levels of the business, while at the highest levels champions influence the behaviours and principles the organisation wants to embody in its use of BI. At an operational level champions help realise BI benefits by ensuring adoption of tools and processes is consistent and comprehensive.
Muggleston is in good company in saying it has traditionally been difficult to quantify the benefits of BI programmes.
“If a BI programme is about a set of behaviours or principles, then benefits should be able to be quantified for each project or specific deliverable. These may differ, but will typically be based around increased accuracy, visibility, and timeliness of information and may also include operational efficiencies,” he says.
As for vendor selection, Muggleston says as BI is such a broad and expansive area, it would be rare to find a single vendor able to meet the BI challenges of each organisation.
“Each [BI] vendor has its particular strengths — find out what those are and use them appropriately.”
Life pharmacy: Boosting customer service
Retailer Life Pharmacy decided to outsource management of the customer information held by its 21 stores and which contributed to business intelligence. OgilvyOne Worldwide was selected to collect and manage customer data on behalf of Life Pharmacy, a process that involved the consolidation and organisation of information from 500 separate customer databases into a single, New Zealand-wide database.
Paul Hickey, strategic planner for OgilvyOne Worldwide, says the New Zealand arm of OgilvyOne Worldwide uses SAS Marketing Automation software suite. He says these tools, which include a data warehouse with data integration software and processes, analytics, and an online reporting tool, were instrumental in cleaning, ‘de-duplicating’ and formatting Life Pharmacy customer data so that it complied with New Zealand Post’s name and address rules and delivered intelligence back to the business.
“We created a single customer view and data that comes back to a single transactional point. This has given head office a view of their customers that they didn’t have before and prevented issues like the same customer receiving three identical letters from three different stores at the same time,” says Hickey.
He says the results have been relatively straightforward to measure: 50,000 of the 100,000 customers on Life Pharmacy’s databases have now become ‘active members’, partly though the advantage of customers being able to access their own purchasing history at any store.
“The customer will say ‘I can’t remember what colour lipstick I bought or from what store’ and the salesperson will be able to tell them. Each store has online access to a bunch of reports that update themselves every night and that meet the broad KPIs of the programme. We have campaign reports and ongoing reports; reports that automatically update each day and allow the business to calculate ROI on any campaign,” says Hickey.
Grant Febery, marketing manager for Life Pharmacy, says the business had a certain amount of business intelligence before engaging OlgilvyOne Worldwide, but didn’t know as much as it needed to about who its customers were, what they were buying and from where.
He says Life Pharmacy realised it needed to move faster than the time it would take to internally make decisions around the right retail technology platforms, which is why it chose to outsource to OgilvyOne. That choice wasn’t without its expensive side, says Febery, but the cost benefit analysis equation will improve the more the business uses the data.
“We are really just diving into it now. The costs to launch came in on budget and we are still looking at tweaking little things and improving engineering from a cost point of view. But the important thing is that we continue to evolve our customer [systems] and how we talk to customers.”
Key BI trends you need to know
There is so much data, but too little insight. Here’s one big reason BI is on companies’ radars: The volume and velocity of information. More data translates to a greater need to manage it and make it actionable. “Everything we use and everything we buy is becoming an information source and companies must be able to figure out how to harness that,” says Bill Hostmann, Gartner research analyst.
The problem? “Organisations are recognising they don’t have the information they need to manage the business,” says Hostmann. The data is there, but it’s trapped in different silos and its accuracy can’t be trusted. For example, how information is entered can vary widely from how it needs to be used to make organisational decisions and, all too often, definitions vary from silo to silo. For example, finance and marketing could define gross margin using different parameters, which influences how and what numbers are reported.
Market consolidation means fewer choices for users. Independent BI companies may soon be a memory. Recent takeovers include Hyperion by Oracle, Business Objects by SAP and Cognos by IBM. On the one hand, there may be some advantages to having large vendors in the BI arena. Traditionally, BI is not a top-of-mind investment for companies, but rather a sort of afterthought once the major application decisions were made, says John Hagerty, AMR Research analyst. As large vendors integrate BI capabilities, BI will be easier to integrate into that vendor’s applications. Furthermore, investment in BI may become more of a core decision, he says.
One the other hand, as independent BI companies are acquired by the giants, many practitioners and experts worry that “the software provider you’d like to use will end up in the hands of a vendor you don’t want to do business with”, says Hagerty. Moreover, the company that does the acquiring may not fit into your current architecture; competing technical stacks may become an issue, he says.
BI expands from the board room to the front lines. Increasingly, BI tools will be available at all levels of the corporation. Operational BI, which brings BI to employees on the front line, is growing especially fast, industry watchers say.
Boris Evelson, a Forrester research analyst, says operational BI will include offerings that integrate data and process dashboards, and event-driven systems that initiate a business process based on certain data conditions. For example, integrating BI into operational processes could allow companies to react faster to changing business conditions, for example, alerting a call center worker to offer a particular promotion or to potential credit card fraud.
The convergence of structured and unstructured data will create better BI. Email, memos, voicemail messages and other sources of unstructured data are rich sources of information, with companies and developers responding by looking for ways to blend structured and unstructured data for better decision making. For example, retailers could add comments and complaints from email and call centres into a BI application to enhance their market segmentation analysis, says Evelson.
Hagerty agrees: “Getting information that’s not in the typical rows and columns is the next logical frontier.” He adds that this trend holds real promise. “This gives people more information to make the right decision,” says Hagerty. People will no longer have to rely on “gut”, but will instead have more information with which to make informed decisions.
Applications will provide new views of BI data. The next generation of applications is moving beyond the pie charts and bar charts into more visual depictions of data and trends. Witness products like JMP (from SAS), Spotfire (from Tibco), Tableau, Thinkmap and others provide visualisation for complex data. Alternative ways of displaying complex data — to increase interaction and usefulness — is an area that will continue growing in the coming years, say Evelson and Hagerty. Diann Daniel
Tips for better BI
Lack of BI and IT skill sets continue to plague companies interested in taking BI to the next level, says David Hatch, research director at Aberdeen Group. Part of the answer lies in providing more user-friendly BI tools. He offers four tips for achieving that goal.
Explore new user-friendly tools. New ways of delivering BI can help in your quest to extend business intelligence throughout the enterprise. One new method to consider is BI accessed through a third party, for example, software as a service BI or on-demand BI. In addition, look into the availability of BI as an embedded capability within enterprise applications such as ERP and CRM. Since users are already familiar with your enterprise applications, they may find BI products offered through these providers easier to learn.
Find ways to integrate Web 2.0 information into BI. Web 2.0 data sources and other unstructured data do not obviate the need for traditional structured data, says Hatch, but they can be used to boost BI efforts. Amassing large sets of historical data reveals trends, performance metrics and specific business calculations. These are the foundation of most BI efforts. But the ability to enhance that historical data with relevant and timely information found in blogs, comments, and competitors’ websites is becoming more important for delivering actionable information throughout the enterprise.
Give users tools that they can be trained to use autonomously. Employees are more likely to use and embrace business intelligence tools that they can use independently. To create an environment of business intelligence self-sufficiency, establish a group composed of both business users and IT representatives to collaborate on prioritising user needs and choosing or developing BI tools. Hatch also advises being attuned to inflated vendor claims and involving vendors in proof of concept and pilot projects.
Consider operational BI. New offerings that automate data collection, assembly and delivery processes are one of the most promising areas of business intelligence. To figure out if they’re right for you, look for data generated by business processes that lend themselves to automated analysis and even actions taken on the basis of that analysis. Diann Daniel
Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.