This article is based on a presentation at the 2016 IPWEA conference by Jamie McPherson, chair of the REG Procurement SubGroup, and Chris Olsen, director, Chris Olsen Consulting. By ALAN TITCHALL.
A RECENT REPORT BASED on 11 road maintenance case studies among road controlling authorities (RCAs) has identified themes, key drivers and objectives, and resulted in a matrix tool for selecting the optimum contract model.
This report is the result of the Road Efficiency Group (REG) setting up a Procurement SubGroup to identify procurement opportunities and tools to help local authorities pick their road maintenance models, especially when carrying out a Local Government Act Section 17A review.
The Procurement SubGroup has since produced a report called ‘Road Maintenance Procurement: delivery model guidelines’, with findings made up of international research and 11 New Zealand RCA contract delivery model case studies.
The Road Efficiency Group (REG) was formed in 2012 on the recommendation of the Road Maintenance Task Force. It is a collaborative project between local government and the NZ Transport Agency. Its aim is to create and embed a new national funding and activity management structure for roads (the One Network Road Classification); and improve value-for-money, customer focus, consistency, collaboration, and quality in road activity management. It has the following four subgroups working on various initiatives ONRC, Data, Capability/Enabling and Procurement.
The annual budget for local authority road maintenance, operations and renewals is around $1.4 billion, so even small ‘improvements’ can add up to substantial dollar savings in expenditure.
Delivery models can impact ‘value for money’ significantly as some models are better suited to different market conditions and client objectives.
“This has resulted in significant sector discussion over the past 10 years or so with no consensus being formed over which models suit which conditions,” says the guidelines.
Back in 2012 the Road Maintenance Task Force found that around 80 percent of local authority road maintenance contracts by number were ‘Traditional’, where the council retains control over the programme of work, which is delivered on a measure and value basis.
In early 2015, a survey by NZTA’s Planning and Investment Group found this percentage had reduced to 70 percent through greater use of newer contract models.
The ‘Road Maintenance Procurement: Delivery Model guidelines explores reasons for this change and suggests ways for RCAs to determine which delivery models suit different circumstances and drivers.
The guidelines define contract types as: Traditional; Performance-based, aggregated and bundled and/or PSMCs; Alliances, including cost reimbursable and fixed fee for overheads and profit; and Framework/Panel.
These are the model names commonly used for maintenance, it says, but are not always the same as those used in the Transport Agency’s Procurement Manual.
The Traditional model will overcome a lack of contractor competition and leaves the client in control of the programme of work. Traditional models are not suitable when aspects of the scope, work programme or risk is uncertain so an accurate schedule can’t be developed.
Performance-based models let the client specify performance requirements and measures, while holding the contractor accountable but letting them innovate and decide the work programme. They are not suitable when there is a lack of data for the contractor to determine the optimal way to deliver the performance specified, or when the client cannot determine suitable, robust KPIs. Also Performance-based PSMC are likely to be problematic if funding or service levels change over the contract term.
The Alliance model sees the client work collaboratively with the contractor in understanding cost structures (for identifying potential cost savings) and targeting continuous improvement. It also provides ‘flexibility and risk sharing’ in a structured ‘one team’ approach with the contractor. Alliance models are not suitable when the client does not want to risk share, or there is a lack of ‘maturity’ or collaborative working practices within the organisation.
The Framework/Panel model works best when the client wants a group of specialist local suppliers on tap to carry out work as required and directed. They are not suitable when the client does not have the resources to handle multiple suppliers; doesn’t need specialist skills; or when the industry needs ‘certainty’ in regards to the forward work programme.
Another key finding is that in every case study covered by the report, the delivery model had been chosen because the previous model had not been meeting the client’s objectives.
For example, the Ruapehu District Council moved from a Performance-based contract to a Traditional model to ‘unbundle’ and use local contractors due to the lack of interest from large contractors. The Tasman District Council also moved from a Performance-based to a Traditional model to bring back in-house control of the work programme, because it required ‘flexibility’ on its network maintenance.
The Kaikoura District Council used a Performance-based contract to enforce its contract and ensure it got what it asked for (while minimising its time commitment to the programme). The contract also had strong subcontractor requirements due to local market considerations.
The Waikato District Council chose an Alliance contract because of its flexibility and sustainable market pricing after being burnt with unsustainable bids. Hamilton City also chose the Alliance model (over the Traditional model) for its flexibility around service and funding (after 20 percent funding cuts) to better understand costs structures and target the opportunities.
The NZTA Highways and Operations Auckland Motorways Alliance uses an Alliance model due to the complexity of the network, while the Southland District Council uses this model because of the flexibility around levels of service/funding and its ‘flat line’ expenditure budget.
There was only one Framework contract and that was Kaikoura as a temporary measure while it got its Performance contract up and had specialist contractors on hand as required.
Two step guidance matrix tool
The guidelines has developed a matrix table where the RCA can score (rate) its ‘key drivers’ and the preferred model is the one with the highest score. The model is then tested with a secondary matrix table with another set of ‘secondary’ drivers.
The guidelines note that client drivers are not fixed, and change with the organisation and the environment. But to use the matrix tool, RCAs need to totally understand what their current key drivers are. Also, although the report defines the characteristics of different performance models, it acknowledges overlap in their application.
“Case studies have shown that the reality is that many delivery models are a mix of the Traditional, Performance and Alliance models. In recognition of this, a local authority may wish to custom build their delivery model by determining the delivery model characteristics they seek and constructing a hybrid delivery model from first principles. If this approach is taken it is suggested to develop three options and run them through the matrices to test them and find the optimum model.”
The authors say their ‘matrix approach’ has been validated for both rural and urban situations by testing it against the delivery models of the Ruapehu District Council and the Hamilton City Council. “The proposed delivery model selection matrix tool also aligns well with the UK Highway Maintenance Efficiency Programme’s (HMEP) Toolkit and Guidance document ‘Procurement Route Choices for Highway Maintenance Services’.” This toolkit is very similar to the matrix tool the Procurement SubGroup came up with.
Case studies also showed that the best way to apply the matrices is through a ‘workshop environment/process’ with an independent facilitator. This team should be made up of key local authority staff and one or two independent industry experts. Their task is to identify key and secondary drivers, the desired outcomes, critical success factors, and preferred ‘contract culture’. This process should only take around four to six hours, say the authors.
- Click here for the model guidelines. Included is a delivery model toolkit – a shortcut to the selection process that can be used as a discussion/workshop resource.