Under the Canopy
- 12 June, 2005 22:00
Carter Holt Harvey’s Forests business has a secret. It’s not just about the breadth of its new tree-to-customer computer implementation. It’s much more than that. What’s immediately evident as soon as you meet the project team is, well, its synchronicity. Put these people together and you can immediately see how Project Canopy has come together as a world-leading implementation that encompasses planning and managing forestry operations, harvesting and distribution across the business’s entire supply chain. This easy sense of camaraderie, this synchronicity, has enabled the project team under Forests chief executive Jeremy Fleming and business planning manager Roger Kay to fulfil an implementation that is already making an impact on the way the organisation manages its business and meets its targets. Not only that, and perhaps of less importance, is the fact that it expects to show a return on its hefty investment within just two years.
I met several of the key players at a round-table discussion at CHH’s Manukau headquarters, an oasis of green, calm and placid water right beside the Southern Motorway.
The first person to shake my hand is Kay, the quietly spoken programme manager. We find a small conference room where we can talk and we are soon joined by other senior players on the implementation team – Conrad Keast, Paul Sipos, Aaron Bolton and Doug McCutcheon. It’s a confusing situation for a journalist, having to keep track of who is speaking, but it soon become apparent that this is how things work here. This is all about teamwork. Kay is the man who stands at the head of the activity but it is the team that has the biggest story to tell. It is Kay, though, who must set the ball rolling…
In the beginning
“Before Canopy, Forests had an inhouse software system to manage spatial activities and to capture forest information,” says Kay. “Documentation for the sale of logs and associated distribution activities were handled via a piece of paper we called a log delivery docket. We would capture data off the log delivery docket and use it to invoice customers and create invoices for suppliers such as harvesting contractors, distributors.”
The system as it existed then had been in use for about 10 years and had been through a variety of ad hoc changes to accommodate changes within the organisational structure of Carter Holt Harvey. Those changes included a change from an integrated business to an operation that was split into various components, including Forest Resources, charged with growing and managing the trees, and Fibre Solutions, which was engaged in sales and marketing.
“The old system was nearing the end of its life and needed significant re-working to update it and to accommodate some of the new business processes and opportunities within the Carter Holt Harvey business,” says Kay.
At the same time Carter Holt Harvey had made a significant investment in standardising on SAP’s ERP software. Out of this activity was to grow Oxygen Business Systems, originally headed by Jeremy Fleming, who came from the forestry side of the business.
Around 1999-2000, the big question for Forests was whether to spend a significant amount on upgrading its old system or whether to adopt a new business model centred around CHH’s investment in SAP. The decision to go with SAP was made in 2002 and the project kicked off in 2003.
At the same time as SAP considerations were going on, Forests was facing another big challenge – how to integrate its spatial information with its GIS system, and how to integrate all of this information usefully into the supply chain. “We had made a couple of starts with supply chain planning in forests but we hadn’t been able to get all aspects of it together to make it functional,” says Kay. “Handling the amount of data we received had always been a big problem. We had six or seven regions across the country, a couple of hundred CHH employees and a contractor base of 3000-4000. All the data they generated proved difficult to get to a level where it was appropriate to use a software solution to help us out with our supply chain plan.”
The opportunity was there, nevertheless, and if it could be achieved it would add a lot more sophistication to CHH’s supply chain considerations. One thing that did hold promise was SAP’s APO (Advanced Planner and Optimiser). The promise was tantalising. The outcome was still some distance off.
Ironically, there was no complete solution available from overseas or even from within CHH’s 51% owner, International Paper. That’s because each forestry operation around the world tends to operate a little differently. In the US, home of International Paper, the forestry operators typically do not own the forests, which means they do not manage them. “They purchase their fibre at the mill gate,” says Kay. “They tend not to manage the forest, the harvesting contractor and the distribution in the way that has developed in New Zealand. Some of the Scandinavian countries have similar business models, though, as does Chile, which has been heavily influenced by New Zealand.”
When Forests decided to go ahead with Canopy the team knew that what they were doing was highly ambitious. The implementation would be larger than normal for CHH because of the wide range of activities it covered. It made sense to split it into two major components – the SAP implementation with the supply chain activity surrounding it, and the spatial activity with its own attendant software components.
With Kay as overseer, it was determined that Doug McCutcheon would head the SAP component and Conrad Keast would head the spatial component. Keast’s domain would include land information, yield information and forest valuation. Kay’s credentials included his training as a forester and his work with CHH’s IT before Oxygen, as it was to be called, became a separate entity. “When I was at CHH IT, one of the discussions I was involved in was about our strategic implementation of SAP. Canopy provided an opportunity to demonstrate its appropriateness to the whole organisation, including its forest planning business,” says Kay.
Project Canopy kicked off in January 2003 and ran through to July 2004, a total of 18 months. Keast continues the story …
Out in the forest
“The land database encompasses a number of things, including the business applications we use to manage our operations in the forests,” says Keast. “They include a spatial definition of where we are planning to or have completed operations. These can cover harvesting, planting, thinning and pruning – whatever is needed. Integration with SAP was very important: SAP was keeping the financial transactions, about the cost of doing those operations, and our part was to maintain the spatial information of where the operations are happenings and costs are being incurred. You can imagine over time this allows us to understand the full financial history of our forests in a spatial context”
The GIS area is an ESRI environment centred on products such as ARCMap. Functions around these products range – among other things -- from planning, creating and managing operations through to standard map production.
“The other part that I looked after was the bit we call our Estate Description systems which rely heavily on the data managed in the Land Database. These systems handle our medium and long-term strategic estate modeling – we left the short term supply/demand modeling to the SAP/APO Team. Another important function of the Estate Description systems includes forest yield modeling and storing the output of these models in the Land Databasse so this information is then available for subsequent processes, including SAP/APO – these models tell us how much wood we currently have in our estate and how much we expect to have in the future. Other important functions of the Estate Description systems include the allocation of current and future prices and costs to each bit of our estate, and management of land tenure. While these systems are independent of the SAP side of the project, the operation model is the key integration point between them. All of the Estate Description systems were developed in Java, and the ArcMap sstems were all developed in Visual Basic.”
Here lies an important point. That part that had handled forest valuations, woodflow modelling and yield calculations had been developed by technically astute forestry people, starting in the late early 1990s. The result had been a mix of programming environments, depending on the talents of the people employed. “A new person might be familiar with Perl, so we would use that,” says Keast. “We ended up with a programming environment that was getting quite difficult to maintain. But the most important thing, from my point of view, was that we ended up with our Resource Foresters having to spend a lot of their time writing and fixing code. So we made a decision that when we moved to Canopy we would standardise on Java.”
The intention, he says, was to make a clean break from the past. From that point on the our Resource Foresters would spend their time analysing the wealth of data we have on our forest resource, not getting stuck in the bowels of SAS or Perl. That decision, says Keast, has already brought massive benefits.
“In the eight months since go-live I have seen a significant improvement in the quality and quantity of resource analysis we are doing. There are some other things driving that, but there is no doubt in my mind that these guys now have the time to get on with the job they are qualified and skilled to do. We are seeing a lot better decision-making.”
As an example of this achievement, Keast says CHH has have recently completed a complex analysis of various interacting factors affecting future woodflows , culminating in reducing harvest levels. The analysts could have done the same job previously but the quality of the analysis would have been inhibited or prevented by a lack of time as they worked on software development or fixes. “We are seeing a lot better decision-making.”
Keast points to another example of success. “Every weekend we generate two integrated datasets we call the General Purpose Overlay and Standard Forest Description - think of these as a massive spatially enabled data warehouse that takes something like 36 hours to process. The process involves spatially intersecting a number of GIS layers and then for each of the resulting slivers calculating everthing you could imagine a forestry company would be interested in – from future costs and revenues to altitude to previous, current and future crop information. This is the most important resource that our Team have to conduct their day-to-day work. Prior to rebuilding this process under Canopy it was not uncommon for this process to fail every 3-4 weeks effectively leaving our Team without up to date data to use in their analysis. Since the implementation of Project Canopy this process has failed only twice and both of these failures occurred soon after go-live. This has obvious benefits for us in achieving often tight reporting and project deadlines.”
Keast says the move to Java has brought a lot more discipline in the programming process, which includes formal development, testing, documentation and release procedures. This was unheard of with our previous systems. And it was all been done by a compact team consisting of key CHH Forests staff, technology solutions provider Gen-i and a number of independent consultants.
By most standards the scale and detail of the GIS operation is huge. The forests cover 440,000 hectares, broken down into 2.5 million polygons, each representing an activity. Think of forestry in New Zealand as a form of farming, with environmental issues, with crop rotation encompassing about 28 years. Each plot, or polygon, must be assessed for maximum crop performance and this is what the Estate Description systems and Land Database is all about.
Keast is clearly in his element as he moves to a whiteboard to outline how the operations work. “Imagine this is a forest,” he says, drawing a circle. “On day one we come along and map out the areas we are going to plant, plus those bits we are not going to plant. This might be a bit of native forest, for example. Then we might say, right we need to do some land prepping. From this work we might create, say, operation A. Against that we have to plan our start date, our finish date, plus the actual start and finish dates. The next thing we might do is plant the area …”
What it comes down to is a form of micro management. Each polygon might be split into smaller plots for different operations, and those pods, or pieces of dirt might multiply into smaller sub-pods as new activities are started.
“Basically, what we end up with is this thing we call an OpStack, an operations stack,” says Keast. “If we look at, say, pod four, our system will tell us that it had some land treatment in 1982, was planted in 1983 and thinned in 1994, whatever, and would be clear-felled later this century. So, for us, this piece of dirt is homogeneous in terms of its history, its current state and its future state. All of this information is integrated into the SAP system, where we start to assign dollar values to it. Now you can ask questions such as who planted an area, how much it cost per hectare, or how many kilograms of chemicals were sprayed.”
Keast’s team generates yield forecasts from the consolidated information. Without it, the business would have a hard time forecasting its future growth and income. With its enormously long product cycles, the forestry business needs the information to assess the value of its activities as the product slowly matures. Felling and planting can be manipulated to meet forecast market demands.
It’s a far cry from the beginnings of forestry in this country. Nowadays the forecasts stretch far beyond a single crop cycle. Long-term planning encompasses the business’s needs and requirements as far as 50 to 60 years into the future. This information is anchored into the five-year plan as well. Now, when an adjustment is made at any point in the plan, the system is able to assess its impact far into the future. That’s something that didn’t happen before Canopy.
Kay suggests we should think of a tree as being made up of a series of salable components. Some trees need be nothing more than a pruned log, others can be used as structural saw logs, some can be saw logs for appearance products or packaging. At the end of it all is the pulp. If any single part of the business has a turndown – such as the valuable export market with its high shipping costs – Forests can end up with a piece of the tree that can be hard to sell. That’s where modelling becomes so important – deciding which part of the forest should be accessed to get the best result at any given time.
Doug McCutcheon, SAP Project Manager at Oxygen Business Solutions, continues the story…
Drawing it together
“It was an interesting project in that everybody had to deal with something that hadn’t been done before. But we managed to do it with the lowest level of customisation in any of the CHH businesses. We set out not to write customised code that would have to be maintained and we achieved this with some creative thinking in the way we could use SAP’s APO software and Project Systems. We kept to standard transaction processes when we talked about purchase orders, sales orders and the like.”
“The big concern for us was the integration of the ESRI information into SAP,” says Mcutcheon. “ While GIS and SAP had been integrated before the level of integration we were trying to achieve hadn’t been. We went to SAP and out to the rest of the world to find people who had done similar things so they could tell us the pitfalls. But we couldn’t find anybody.”
McCutcheon defined three areas to focus on in particular. One was the integration with ESRI and the other was the integration into the business of several distributors. Delivery instructions, for example, needed to be reciepted as electronic dockets when trucks arrived at a customer site. The final being the matching of demand and supply and the subsequent logistical execution using the APO/JDE solution created by Canopy’s MRPII project manager Aaron Bolton and Paul Sipos, Service Line Lead for Supply Chain Management. Once they were filled in, the dockets needed to be pushed back through the third party distribution system into SAP.
“It was an interesting project in that everybody had to deal with something that hadn’t been done before. But we managed to do it with the lowest level of customisation in any of the CHH businesses. We set out not to write customised code that would have to be maintained and we achieved that by using SAP’s APO software and project systems that provided flexibility. We kept to standard transaction processes when we talked about purchase orders, sales orders and the like.”
Naturally, CHH’s US counterparts were interested in the activities Down Under. “They came over and looked,” says McCutcheon. “They decided that what we were doing, especially on the planning side, was well in advance of what they were seeking to achieve at the moment. Also, because their business stops at the procurement side, they were not looking for total management and total cost collection. The didn’t need to know the same details – how much a piece of wood cost to produce, as opposed to buying it off the back of a truck. Our operations also differ from those of the Asian forestry companies in that we sub-contract a lot more. They own their own workforces and move them around as required..”
McCutcheon points to another feature of the SAP implementation – its service management module. The forestry business here uses this software to secure a lot of its subcontractors. However, it is not designed to handle the business’s materials management requirements. “That aspect of the application was developed to support the engineering and harvesting side of the business, particularly the engineering side.”
Canopy’s MRPII project manager Aaron Bolton continues the story …
Taming the tiger
“The discipline of using SAP has meant that we have managed to get our data sufficiently under control across the organisation, something we hadn’t been able to achieve in the past,” says Bolton. “Its discipline and data handling requirements have allowed us to train, educate and commit people to the business processes.”
Yes, SAP’s requirements can be seen as a sort of benign tyranny, but Bolton points out that the constraints result in a greatly improved organisational performance.
It’s hard to argue with the results as they are outlined by chief executive Jeremy Fleming, who confirms that Canopy is already reducing harvesting and distribution costs. Other benefits, for example from improvements in land use efficiency, are expected to result in total annual cost savings of more than 10%, he says. “Project Canopy is a major advance in our ability to make business decisions based on accurate, real-time information,” Fleming says. “Instead of being supply-driven, where we basically sell logs as they come available, we can now adopt a much sophisticated demand-driven approach. On a day-to-day basis, this means we can be efficient in filling customer orders. We can see which trees are suitable for each order, and we can plan and execute the most efficient harvesting, processing and distribution scenarios. At a strategic level, we can do sophisticated planning which takes into account likely changes in the whole tree-to-customer supply chain. We can look out five years and make informed decisions as to new plantings and harvesting. Most important, we can see the impact of those decisions on a day-to-day basis and act quickly according to changes in sales or costs. Canopy gives us real competitive advantage – not just in terms of driving down costs, but also in giving us the ability to offer reliable, consistent fibre supply to our customers, and hence improving customer satisfaction.”
Project manager Doug McCutcheon says one of the highest return-on-investment areas has been demand matching – being able to see demand in terms of orders and to see supply in terms of available trees for harvesting. This has resulted in savings by minimising external procurement.
Considering the size and ambition of the project, it has all been achieved at a surprisingly low cost -- $7 million. More importantly, Forests expects to have recovered its investment within two years. That’s an inspiring outcome.
More important, though, it is an example of what can be achieved under a solid, strategically orientated IT leadership. In a market where the pundits are constantly telling its leaders they need to invest in their information technology, CHH Forests is providing an inspiring example of leadership.
Under the hood
SAP applications are the major solution building blocks in Project Canopy, integrated with ESRI’s ARC GIS products. Webmethods is used as the integration platform between the core internal systems and the systems of external suppliers. Key SAP applications include SAP APO, SAP Project Systems, SAP Sales and Distribution, SAP Finance, SAP Profitability Analysis, SAP Materials Management and SAP Service Management. SAP’s BW (Business Workbench) is used to bind together the complete solution, integrating the data from GIS and SAP systems, as well as data from an Oracle-based land yield database which holds all information on CHH Forests land holdings.
JD Edwards Strategic Network Optimiser (SNO) is the main component of the supply chain optimization functions.