Resourcing New Product Development
Peter Hoyland explores how successful Lean Manufacturing practices can be applied to address the challenges of resourcing New Product Development projects.
According to the 2004 APQC Benchmarking Study on NPD performance, 49% of new products are launched late. Not withstanding the additional finding that 48% of businesses believe they have “Inadequate NPD Resources”, this simple failure to deliver reliably must have its roots in the way projects are being managed.
Furthermore we don’t seem to be improving. The same study showed that by the classic metric of NPD performance, the percentage of revenue that comes from new products, we have slipped from averaging 32.6% in 1990 down to 28% in 2004. Something is definitely going wrong!
What can we learn from Lean Manufacturing?
To understand the problem a little better we can turn to the indisputable success of Lean Manufacturing. It took 50 years or so, but during the latter half of the 20th Century, manufacturing organisations first adopted and gradually abandoned Manufacturing Resource Planning (MRP) systems as the best way to plan and schedule shop floor activity. Lean Manufacturing methods are now universally acknowledged as a better solution.
So what was wrong with MRP, and where is the parallel with New Product Development today?
First let’s understand MRP. MRP systems draw upon a detailed Bill of Materials (BOM) for each product and a manufacturing routing that defines the tasks and resources that are required by each component and sub assembly to take it to the next stage of production. Variously sophisticated algorithms are then used to produce a manufacturing plan based on the principles of Batch and Queue. The plan itself generally has a short shelf life, with a new plan being issued regularly (often weekly) to ‘re-optimise’ on account of unexpected production problems and new customer demand.
The tragedy for most manufacturing organisations was that this logic was robust. In theory, it worked! Unfortunately the appeal of local efficiencies that could come from running large batches always outweighed the less obvious inefficiencies that resulted from even larger queues. Production would gridlock, inventory levels would sky-rocket, physical movement and storage became an art form, and production leadtimes went from days to weeks to months. All in the name of efficiency.
Lean manufacturing, on the other hand, takes a completely different approach. The ideal batch size is one unit. The focus is on flowing not queuing; the instruction to make a component comes from the downstream process (Pull Scheduling) not the MRP planner (Push Scheduling). The result is that factories re-organise themselves into product based cells that facilitate the continuous production flow and self regulation of pull scheduling. In fact task level planning no longer exists as this is handled locally by the cell members, who can flex their resources to meet simple local objectives. Inventories and leadtimes plummet and customer responsiveness is back on the agenda.
The success of manufacturing cells has driven more advanced manufacturers to modularise their products to facilitate this model. Each cell produces a complete module that is fully testable and can be easily assembled into a larger whole. The ‘pull scheduling’ principles then apply equally well between cells as within a cell. The effect of this structure is manufacturing stability. All cells work at the same rate and work to simple and stable instructions.
Applying lessons learned to NPD
In new product development, all too often we mirror the mistakes of the MRP era. We organise by functional expertise and produce a highly detailed action plan containing every task we can think of. We constantly add more projects into the pipe so that these can “get started” thereby allowing project inventory to grow and grow. And every time something doesn’t happen precisely to plan, we generate a new version which, because we have included so many interdependencies, will dramatically re-prioritise everyone’s work. And in NPD we don’t even have the MRP tools. It’s like trying to schedule the work of a traditional manufacturing ‘job shop’ using Microsoft Project.
So what’s the solution? Are there any key principles we should adopt?
- First we must recognise that, whilst every new product will have some unique elements, we must manage them as a continuous flow process not a slow moving inventory of projects. For example, by organising as co-located, cross functional project teams (Cells) we simplify everyone’s life because questions of changing priority don’t arise within the team. The result will be fewer parallel projects delivered faster, but not necessarily fewer projects overall.
- Second, Project Managers, (Planners) need to stop planning task level detail. This is unlikely to be wholly accurate, and is the quickest way to ensure that the team have little ownership of the outcome. This is even more important in NPD projects where there is inevitable uncertainty and a high reliance on flexible, ingenious people making the project a success. In its place, a clear Phase Gate delivery plan with clearly defined gate requirements is more likely to motivate the team to deliver.
- Third, think about how to design the project to be more stable and more robust. Most basically this can be achieved by minimising risks and eliminating interdependencies. For example the need for a modular product in manufacturing should be paralleled with a modular development project. Phase gates help here too because they are rock steady check points that force all the functions involved to synchronise their activity.
- Finally use rough cut capacity planning to decide how much resource will be required for a particular portfolio of projects. Building up resource requirements from detailed task level plans creates a false sense of control and inevitably leads to the need for continuous updates. Remember that a plan that is updated every day is not a plan; it’s simply today’s view of the status. It’s only a plan if people are following it rather than correcting it.
The benefits of Lean Manufacturing are quickly recognised because it's far easier to see what works when you can physically see parts flowing. When things are not flowing in NPD, it is perhaps historically instinctive to blame the project manager. In fact if the process is inherently flawed, no amount of manager rotation, training courses or KPIs will lead to success.
By following the simple guidelines above, NPD managers should avoid taking that wrong turn that led to so many manufacturing organisations choking with inventory, whilst unable to deliver on time and battling for survival. Project management software tools can, and should, still be used but within the context of well designed projects, with minimum interdependencies and rough cut resource planning. When it comes to resource management less really is more!
Peter Hoyland. October 2006
