Knowledge for all

Knowledge for all

Making the most of your information assets means access to data more widely within your company. But be prepared for the challenges in shifting from an enterprise information system to ‘everyone's information system'.


  • How to ensure crucial information is within reach of the appropriate staff.
  • Why it is important to focus on ensuring broader adoption of systems.
  • What are the signs that projects may be taking the wrong direction.

IT IS A truism that knowledge confers power, and this is certainly so in business. For efficient and effective performance you need to know what the market wants, and how the business at all levels is performing.

That is perhaps why the tools and insights of business intelligence (BI) have often been jealously guarded as an organisation's great secret weapon.

However, in employing BI as a competitive tool, businesses often make the mistake of guarding the information system within the company; making it accessible to a chosen few at high levels of the organisational tree.

There has, however, been a shift in the last few years; a people-power revolution if you like, which has seen many organisations significantly "democratise" the distribution of BI throughout their workforce and released that information to workers at the coalface.

Technology change has played a major part. Web interfaces for leading BI products such as Cognos and the redesign of some products with an eye to ease of operation have made it easier and cheaper to allow more people to access BI tools. Such democratisation has been characterised as the expansion of the "enterprise information system" to become "everyone's information system".

The downside of keeping advanced information tools away from front line workers is that they tend to develop their own tools in an ad hoc fashion — typically through spreadsheets, inconsistent from one team to another and through small and disconnected silo-like databases using such applications as Microsoft Access.

This was a situation Steve Camford, IT manager at New Zealand King Salmon knew too well. The Nelson company grows, processes and sells 40 per cent of the world's Chinook King Salmon. As a salmon farmer, manufacturer, and sales and distribution business, it has distinctive business system needs, different for the various divisions but capable of information flowing between divisions and having a common repository — a single version of the truth — which the company significantly did not have in 2002 when it was selecting and installing a new ERP system.

"We had various ad hoc ways of creating reports; one of our [IT people] had Crystal Reports experience, so we used that a lot." Various departments created Access databases; these were disparate efforts born of a fire fighting perspective rather than a systematic and co-ordinated approach, he says.

Evaluation and testing of ERP software candidates showed up true business intelligence as a significant lack, says Comerford. "The standard reports provided with the ERP systems did not meet our requirements." It was clear that a specific product or suite of products was needed.

The team looking after the project-two from the IT department and management and system accountant Aaron Schroeder-identified two critical needs: "First, the operational stuff; telling people each day what orders are expected and matching it with what we've got in stock" and second, the higher level ad hoc analysis of the performance of the various business units against defined indicators.

Tools such as Crystal Reports were of some use in the former area, but did not tackle the latter at all. New reports meant a request to IT, which diverted the department's staff from crucial longer-term work.

"We needed something intuitive for the end-users," says Comerford.

Staff needed to be able to easily put together their own reports and explore the data to meet a specific point need or diagnose an apparent problem.

Movex was selected as the ERP system (now renamed M3) and there was an expectation that the Cognos suite would be implemented for business intelligence, since the two systems had particularly good links. However, the team decided a wider range of choices should be explored.

A single repository

NZ King Salmon felt there was a lack of a central repository for data. It had an SQL-based data warehouse that predated the Movex move, but it had not used this intensively or well, says Comerford.

Comerford acknowledges King Salmon's approach has been rather more reactive.

"We thought we had a reasonable idea of our needs already, and we talked to the general managers of each division and put together a report. We didn't go into great detail," he says, being conscious that unduly deep analysis would delay implementation. With an ability to define new reports easily, it was felt that users would themselves come up with whatever they felt was needed.

Different tools may well be needed for the different kinds of BI, but for a relatively small company like King Salmon this would have posed challenges of interfacing and of maintaining different bodies of expertise, says Comerford. "We thought it would be better to get everything from one supplier."

The company drew up a list of six candidate suppliers, later reduced to three final contenders. They were asked to demonstrate the capabilities of their offerings on the company's live data. "We hijacked the systems for an afternoon to let the suppliers run their demonstrations; we didn't think packaged demonstrations on dummy data would be a fair test."

Ease of use, the ability to build reports quickly, a dashboard capability and the ability to drill down in the data were considered crucial. The BI suite had to be capable of interfacing with both the live system and the data warehouse.

There was surprising variance in the capabilities of the candidate systems, he says. "Some analyses that took an hour in one product would have taken a day in another."

The chosen system was from California-based Hyperion Systems.

The software has met expectations, Comerford says; there have been few challenges, though one major customer report took a while to subdivide in the most appropriate way. New Zealand King Salmon is in the process of upgrading to the latest Hyperion implementation, Version 9. This delivers improvements particularly on the dashboard side, but the upgrade is also motivated by a desire simply to "stay current" and ensure continued support, says Comerford.

Top-level view

Whereas in the past the technology was often too complex to be easily used by the target audiences, it also suffered from a failure of process and people, and it frequently led to BI being a frustrating area of missed opportunities.

For years, technology leasing specialist Flexirent laboured with disparate databases for areas such as credit applications, contracts and repayments for its 325,000 customers in New Zealand and Australia.

Ben Taylor, Flexirent's chief marketing officer, says the business had no real macro or top-level view of its information. If a customer had two contracts with them, it was probable that no one at Flexirent would have known about it, and therefore, wouldn't have been able to put a credit application together with a repayment record.

"There was no real way of understanding our customers in a total sense, and there was little effort in communicating with customers because we didn't know what they were really interested in; we didn't know if we were selling hamburgers to vegetarians," says Taylor.

Flexirent's BI project was integral to its customer relationship management and marketing strategies because it helped deliver a single view of the customer. The company implemented a system from SAS Institute that sits on top of the various systems and distributes reports to users on a web browser.

Taylor says the system is used by everyone in the organisation and there is a link from the intranet through to SAS reporting with, multiple folders depending on an individual's role in the organisation and what they do.

Not only is the system helping Flexirent join the dots in understanding its customers and reducing the risk of bad loans; but it is delivering tangible benefits through more cost-effective direct mail and has helped the company triple its revenue from cross-selling.

"Now we are really able to profile the customer based on propensity scores, and understand whether, based on their history with us, they might be interested in other products and offers," Taylor says.

A collaborative environment

SellAgence, Auckland-based provider of market management services for global brands such as confectionery maker Ferrero and formerly Gillette, provides lessons in working with the technology provider for BI systems.

The company, the link between international clients and local retailers such as Foodstuffs and Progressive Enterprises, took the one-supplier route in 2002. It matched Oracle's Discoverer business intelligence product with other elements of Oracle's e-business suite, from supply-chain management to financials.

The reporting end of Oracle's BI was of value not to the company itself but to its clients. They required their own regular reports of business with agencies, in many cases to feed into international information summaries, says executive chairman Kerry Gleeson.

"We don't tell our global clients how to run their businesses, so we need to match their requirements of compliance and financial reporting," says Gleeson. "An important factor in winning business has been our ability to show that we have the systems in place to fit within their own reporting frameworks."

New reports can be produced readily, he says. "We were recently asked by a client to recast their last five years of trading history using a different fiscal year end. This report only took a day to produce. Before, it would have been impossible to meet this type of request."

Equally, Discoverer provided the facility for multi-dimensional analysis of SellAgence's own business. "Using Oracle Discoverer, we have really refined our decision-making," says Gleeson. "It helps us look at the business from many angles and determine where we can reduce costs or increase margins. We can even look at full customer profit and loss accounts by channel or product on a daily basis if required."

SellAgence's clients define its performance on the case fill rate, which is the number of items delivered per order on the first shipment.

"We strive to deliver the complete order, accurately and on time but, as anyone in the industry can tell you, this can be tricky as we also aim to reduce the levels of stock we hold."

In January 2003, case fill was at 75 per cent. Now it is up in the 90s, says Gleeson. "Where we once carried 85 days of inventory cover, we are now down 40 per cent to only 52 days.

Normally, if you reduce your inventory, your case fill will drop off. However, by improving the accuracy of our forecasting, we have actually reduced our inventory."

More recently, SellAgence has downsized, losing the Gillette agency and is looking to a smaller and less costly alternative. "Oracle was costing us about $300,000 a year," says finance director Barry Fallwell, though it has served the company well. SellAgence is now adopting Greentree and will again be following the advice of the supplier in the BI direction and using ClickView. This has yet to be implemented, and Fallwell cannot give an opinion on its merit in practice, "though it looks as though we should be able to do everything we've done with Discoverer", he says.

Gartner, meanwhile, cautions against too readily accepting the BI system put forward by an enterprise application vendor. This clearly has its attractions. For example, it can be assumed that the product will have an easy interface to the rest of the suite, particularly the data warehouse. However, the enterprise vendor's system should at least be compared with a market-leading speciality vendor in case valuable advantages are missed, says Gartner.

Certainly, "because it comes bundled in free" is not a good reason for adopting the enterprise vendor's BI. Gartner says this is falling into the trap of concentrating on up-front cost and ignoring total cost of ownership over the long-term.

The power of pervasion

Business intelligence was the "top priority issue" for chief information officers in 2006, says Gartner. But Stephen Bennett, the analyst firm's research director, asks whether that is because it is now solving their issues or because it's a bigger pain in the butt than anything else.

"Actually, I think it is so important now because [chief information officers] know they have to work their information assets better," he says.

"All the traditional pillars of business: Money, people, materials, have been joined by information and we are not talking about single dimensional problems, we are talking about a whole bunch of factors such as increased regulation and greater competition. If CIOs don't have effective mechanisms to understand information and communicate information, then they know they will be at a disadvantage."

Bennett says the maturation of BI has been by stealth rather than giant leaps. He believes that if you assess what was possible five years ago and what is happening now, you can see greatly enhanced capabilities being used by smart organisations which are re-examining the fundamentals of the discipline.

Before he joined Gartner, Bennett was working in Europe, implementing BI in various organisations. One of his projects in an insurance company showed BI intersecting not just with supply chain strategies, but customer relationships as well.

He recalls being brought in to do a data warehouse project, but the business had so much information, it couldn't focus on what was important. They built a BI platform but delivered it primarily to the largest corporate customers of the insurer. Because they were worried about losing customers, the internet-based analytical platform allowed those corporates to look at their portfolio of insurance policies. "They could save it in whatever format they wanted and customise the reports we delivered to them once a month," Bennett says. "So we were adding value to the customer but also giving them golden handcuffs if you like, because they knew they'd lose that service if they went elsewhere."

The result, he says, was BI helping drive customer retention through better service. It created an environment of demand for this new information, and gave customers new abilities based on better knowledge.

The main lesson from all this is that the best implemented BI no longer exists in a silo of isolation.

Integration projects are linking it with customer relationship tools and in some organisations; it is also dovetailing into the whole area of enterprise performance management. Elsewhere, there's a link between BI and the information-sharing aspects of the best supply-chain management projects. The BI catch-cry of the moment is "make it pervasive".

Unfortunately a silo attitude to BI, a failure to distribute its benefits is still all too common, says Gartner. An organisation invests a large amount of money in BI infrastructure and products, but still achieves meagre use and adoption in the business community. Gartner describes this as "information apathy". Relevance to needs, accuracy, consistency and timeliness (the "RACT" test) is essential to ensure broader adoption, says Gartner. "Build it and they will come" is not a productive attitude when it comes to BI implementation.

Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.

Join the newsletter!


Sign up to gain exclusive access to email subscriptions, event invitations, competitions, giveaways, and much more.

Membership is free, and your security and privacy remain protected. View our privacy policy before signing up.

Error: Please check your email address.
Show Comments