House buyers are demanding higher quality than ever before, but fewer new homes are being built each year. John Miles predicts an imminent construction boom and describes the latest design and construction techniques that will benefit the house building industry.
Setting the scene: the need for volume
When the Government outlines its intention to encourage the building of two million new homes in the UK over the next decade or two, it is not making a party-political point. The rate of house building in this country has dropped to a new post-war low and the average age of existing stock is increasing. Those properties at the oldest end of our property spectrum are in danger of becoming too dilapidated to repair and, in any case, much of this existing stock is in the wrong place. In addition, lifestyles have changed quite markedly in the last 50 years (presaging a need for different types of residential accommodation, such as properties designed to suit middle-income, single-person occupiers) and the case for increasing our rate of build and rejuvenating our national housing stock becomes clear.
The build-rate trend diagram of Figure 1 (prepared by the Office of the Deputy Prime Minister as part of its Sustainable Communities publication) says it all. The continuous decline in housing construction is a trend that stretches back over 30 years. We need an upswing in the rate of build from a current average of some 150 000 units per year to somewhere in excess of 180 000 units per year for a sustained period if we are to redress the balance. Occasional peaks could exceed 200 000 units per annum from time to time in the next ten years. Rates of house building seem sure to accelerate; there is no other way of arresting the downhill slide which is currently occurring in terms of the quality and disposition of our national stock.
Setting the scene: the need for quality
Besides the physical need for housing, there are several other pressures present in the marketplace at this time. The first serious generation of owner-occupiers from the never-had-it-so-good era of the fifties and sixties might have been willing to put up with poor quality and reliability in new-built houses as they sought to flee from the inner cities. The second generation at the turn of the twenty-first century is inclined to be much less forgiving. Private house builders now have to subject themselves to customer surveys and league tables of performance, and the negative comparisons between quality standards in housing and car manufacturing are rife. Customers are demanding better design; better build quality; better functional reliability.
Customer power isn’t the only force driving residential providers to raise the quality and performance of their products. Legislation, in the form of new editions of the Building Regulations (Parts L and R particularly), is driving up standards in terms of noise attenuation and energy consumption in use. Health and safety requirements are demanding ever more stringent practices on site to protect the safety and well-being of the labour force. This combination of factors makes it much more difficult to produce houses now than it was in the previous generation.
Setting the scene: the skills shortage
There is yet another issue waiting in the wings to knee-cap the industry: the availability of skilled labour. Using conventional methods, it takes a man-year to build the average privately owned house. A sizeable fraction of this total comes from skilled labour (plumbers, electricians, plasterers, etc.). This labour intensive approach to production is old-fashioned and often criticised, but it has served the industry well for many generations. Amongst other things, it is cheap; it allows a huge variety of product types to be built; and is flexible in terms of where it produces houses each year. One of its most attractive features is that it is scaleable, although this is only true if there is a sufficient sub-set of skilled labour available. Draft in more site labour and the build rate increases proportionately. Current trends highlight the fact that the core skilled labour pool in the house building sector has dwindled to an alarmingly low level. Worse, the average age is pushing upwards through the 50-year-old threshold. The implication of this is that, within the next decade, an entire generation of skilled tradesmen could be lost because there will be no-one to mentor the next generation in on-the-job training – even if the next generation decides that a life in the building trade is preferable to one in a service industry, which current trends suggest is unlikely.
Setting the scene: the challenge
The consequence of the triple pressures of need, consumerism and skills shortage is that we must prepare ourselves for a decline in the skilled-trades base within the industry just as the demand for build rate and quality are set to rise. We must either train or import more skilled labour, or significantly increase the productivity of the projected skilled labour pool, without compromising build quality. Training a new generation of skilled tradesmen has an inherent time-lag built into it – maybe six or eight years. Importing skilled trades from abroad has logistical difficulties, not to mention language barriers, and consequent build quality and safety-on-site issues. Increasing the productivity of site-based labour is left as a possible option but, given that productivity levels need to be lifted by a factor of two or more, how might this be achieved?
Clearly, marginal improvements to conventional processes stand little chance of delivering productivity increases in excess of 100%. What we need is a complete re-think of how housing may be delivered in this country over the next 20 years.
Modern methods of construction
The kernel issue is that we reduce the number of hours on-site that are required to build a house, which can be achieved in a variety of ways. The most obvious is the accelerated introduction of easy-to-fix products, such as plasterboard and snap-fit pipe couplings. There are undoubtedly reductions in site-hours to be achieved through these types of development but, essentially, such products only streamline the conventional process. They do not fundamentally change it and, therefore, the extent of those savings is limited.
More dramatic reductions can be achieved by moving towards a higher degree of pre-assembly. This leads towards more integrated systems and, ultimately, to ‘volumetric systems’ in which everything is pre-fitted and pre-finished in a factory before being shipped to site in large, pre-assembled modules on the back of a lorry. On paper, this is a most attractive proposition, yielding a dramatic reduction in site labour requirements, coupled with a simultaneous increase in build quality stemming from the quality-controlled manufacturing environment in the factory. Such products have the potential to reduce the time on-site required to build a house from around one man-year to around 30–40 mandays. This is exactly what we are looking for. If products of this type could achieve around 30% market penetration, the productivity of the sector would be increased by a factor of just over two.
As usual, however, there are two sides to this story. The most obvious disadvantages of volumetric production are:
Volumetric products are perceived to be inflexible in a design sense, leading to ‘cookie-cutter’ housing.
Transporting large boxes of fresh air is perceived to be uneconomic. This is true if the box is of low value; however, if the value of the box is high (i.e. it is fully fitted out and finished), then the economics of transport are resolved.
The design of good products and the setting-up of factory production lines require capital investments to be made. This is money at risk in an industry which is highly risk-averse. As a consequence, the development of some initial ‘business momentum’ is difficult. First movers take a big risk.
If this is so obviously the right thing to do, why has it not been done before? We have tried (vide the factory-built systems introduced in the UK in the immediate post-war period and, again, during the housing boom of the late sixties and early seventies), but we have not, in the past, succeeded.
The following sections attempt to answer this question, and deal with the other apparent disadvantages of large scale pre-assembly, making occasional reference to the Meteor system (a volumetric concept which Arup has worked on over the past several years. This is but one of several new systems currently under development in the house building industry, all of which fall under the general heading of ‘modern methods of construction’).
Lean production and mass customisation
The dream of factory-built housing stretches back a long way. Architects have been fascinated by the ‘machine for living in’ (Le Corbusier) for most of the twentieth century and there is plenty of witness to this interest in the literature and architectural folklore. But it has never been delivered within the UK at a competitive price and, for this simple reason, experiments with factory-built systems have had short lives.
There are two reasons why the obstacle of price might, at last, be overcome in our generation:
For the first time, it is possible to design and operate flexible manufacturing systems at a relatively low cost. The enablers for this are the powerful IT systems which are widely available these days. These systems have the ability to store all the information necessary to define a particular house type and then control the individual operations on the shop-floor so that it is built precisely to specification. The result is an ability to build a variety of different designs in one production facility, a capability that is essential if repetitive. Unimaginative housing forms are to be avoided (see the ‘Flexible manufacture’ box).
Unlike previous attempts, this particular wave of interest is led by manufacturers and house builders who are interested in construction products, rather than by architects who are interested in the mystique of factories. This is an important reversal, because the skills of the product planner and the production engineer are essential to delivering a desirable product at the right price, yet they have never been properly harnessed in previous attempts to establish factory-built housing in the UK.
The search for low-cost, flexible production systems did not start with housing. The development of lean thinking and mass customisation for the consumer-product industries is well documented but the lessons of these developments may be readily adapted to the development of production facilities for housing systems. The consequence is that the delivery of truly cost-competitive housing products could now become a reality.
However, the outcome remains unproven. The cost base of the construction industry (particularly the housing sector) is extremely low. The construction cost of private housing, excluding the surrounding landscaping and site infrastructure costs, prelims and fees, runs in the order of £400 m–2. This is barely twice the raw material cost – a very low ratio by the standards of any sector within the manufacturing industry, despite the fact that construction involves low levels of automation. Anyone designing a cost-competitive factory-based production system for housing therefore has an enormous challenge on their hands.
This challenge is underlined by the current off-site manufacturing capability in the UK. There already exists a considerable body of production capability that has supplied preassembled buildings for many years. The level of design expertise and product development is high, with products having originated in the temporary buildings market (Portakabins and Terrapin Units, for example) and then migrated to more permanent buildings such as budget hotels and fast-food restaurants. These products, however, were not designed originally to suit the housing market and none of them get near to the low construction cost targets set by the private house builders. Some of these products have been adapted to produce residential units, and some fine examples of off-site manufacture for housing have resulted (see the ‘Generation 1’ box). But, unless the cost can be reduced, they will become isolated examples of a fond aspiration rather than the cutting edge of a new genre. They might all be termed ‘generation 1’ products – products designed originally for a different market and adapted, imperfectly, for housing.
What is needed is a new generation of products, designed from the outset to meet the special needs of the house building sector (which, incidentally, is far larger than the sectors that spawned the generation 1 products). We might term these ‘generation 2’ products.
‘Generation 2’ products
Because of the driving need to meet the low price targets set by the housing sector, the design of economic and flexible production facilities will be the foundation stone of any successful ‘generation 2’ products. The ideal factory for the production of housing within the UK must exhibit three key characteristics if it is to have a reasonable chance of success:
It must have a low capital requirement. The target housing production volumes are quite low in manufacturing terms, with the entire UK requirement peaking at around 200 000 dwellings per annum. Therefore, the systems cannot amortise excessive up-front investments, and the capital cost of a fully equipped factory must run in the order of £10m, rather than the £100m (or multiples thereof) which characterise the manufacturing facilities of the motor industry and other sophisticated consumer-product industries.
It must allow product flexibility. Although the criticism is that ‘all private housing developments look the same’, all of the major UK house builders use a huge range of designs and design variations. One national house builder employs over 90 different designs in the Midlands region alone! This, added to the fact that orders for houses will come in small batches (most sites sell at a maximum rate of one or two units per month), means that the production facility must have an ability to continually produce different designs at very short notice (i.e. build to order).
It must have a high degree of output elasticity. The demand for product is variable, with sales forces attuned to the drivers of monthly, quarterly and annual sales targets. Overlaid on this is a seasonal trend, with certain periods of the year being recognised as ‘house buying seasons’. For a factory to be viable, it must be able to run economically at well below its peak capacity during those periods of the year when demand is low, yet it must have the ability to respond with increased volumes when demand is high.
These three characteristics, which must be delivered simultaneously, pose a real challenge to the production designer. The facility must be a masterpiece of clear, simple thinking without unnecessary complication and expense. Yet it must embrace the best of modern manufacturing concepts with control systems that will prevent assembly errors by semi-skilled workers who are effectively tasked with building a different product every time.
There are no obvious precedents for the development of these facilities. Whilst the technology risk is quite low (the degree and complexity of automation in the plant is likely to be low and the nature of the product being assembled is unremarkable), the development of new products and their associated production facilities nevertheless represents a considerable leap of faith for those who are called upon to make the investments.
The view of the product designer/manufacturer
The leap of faith required of the product designer/manufacturer is the investment of money sufficient to develop and produce a new generation of products. Some essential attributes of these products, illustrating why the task is not simple, are outlined below.
Flexibility of appearance is the first attribute required of generation 2 products once the price problem has been cracked. ‘Kerb appeal’ is key to selling houses and that, plus the constraints of the planners, mean that a pre-assembled system needs to have an external façade that allows a variety of appearances to be generated with ease. Ideally, such claddings would be attached directly to the pre-assembled products at the factory, or panellised, making them high quality and swift to install on site. However, most generation 2 products under development at the moment leave the outer skin to traditional brick cladding and roof tiling and accept the downside of the looming bricklayer/roofer skill shortages, arguing that great advances have been made for all other skilled trades by installing all the interior value in the factory. The argument assumes that cost-effective panellised brick work and roofing products will evolve naturally as the pressures of bricklayer/roofer skill shortages intensify.
Flexibility of layout is the other necessary attribute. Ideally, the product should allow the external perimeter to be of arbitrary shape and the constraints on internal layout should be minimal. The system should be able to accommodate a range of needs from complete open-plan to any arbitrary arrangement of interior rooms on each level. A ‘clear-span system’, in which the floors span across the entire envelope requiring no intermediate support, is the solution to this need. An example of this approach is illustrated by the Meteor system. This system allows the modules to be laid side-by-side, with the interior long walls removed, thus accommodating any arbitrary interior layout (including complete open-plan). Particular internal layouts may be generated using non-load-bearing walls built using dry lining systems. The demonstrator house shown in Figure 2 was deliberately chosen to illustrate the potential of this system to deliver a typical design of the type built speculatively by traditional house builders. It had all the idiosyncrasies that real designs exhibit: a stepped frontage; different windows at each location; a bay window; porch; integral garage, etc. And it was built adjacent to other properties of the same design which were constructed traditionally. The objective was to deliver the house builder's intended layout and appearance, without compromise, using a volumetric system. This was achieved and subsequent on-site tests showed that the resulting product not only met all these criteria, but also displayed acoustic characteristics and robustness that were superior to those of its traditional stable-mates. Nevertheless, from the outside, it is indistinguishable from the rest of the houses on the development.
The point about acoustic performance should not be overlooked. As mentioned previously, the racheting-up of Building Regulations is demanding increasing performance in the area of acoustic performance and energy-in-use. It is becoming more difficult to meet these demands using traditional on-site construction techniques and the discipline of attention to detail design (which factory manufacture requires), plus the ability to control the quality of assembly, give factory-built products an advantage in meeting these demands.
The view of the house builder
The leap of faith required by the house builders is a commitment to changing the entire procurement and on-site processes by which their products are produced. This is clearly a high-risk (and therefore unwelcome) strategy, particularly when their minds are on other things.
From the house builder's point of view the issues are very clear. There is a continuous pressure from shareholders and the City to increase market share and improve margins. Yet the industry is ‘throttled’ by the planning system. The number of sites with planning permissions/approvals is the single biggest factor in regulating the number of houses built per annum and, if asked, this is always where the house builders point their fingers. For this reason, the problem of an impending skills shortage is not as high a priority to private house builders as might be expected. Many house builders argue that if the authorities simply release the land, more houses will get built. The capacity to expand production has always been latent in the construction industry. In the eyes of the majority, if modern methods of manufacture are to become a reality, they must exhibit characteristics which match their immediate (not their future) commercial needs. Namely, that they are priced competitively; improve their return on capital; reduce their cost of operation; and please their customers. This is not an unreasonable point of view.
The first commercial characteristic (cost) has already been highlighted as a make-or-break issue; the positions with the other three may be briefly summarised as follows:
Return on capital: Given that high levels of pre-assembly can reduce construction times on site to an average of 3–4 weeks per house (compared with a conventional average of 12–16 weeks), the profile of capital expenditure for the house builder using these systems is quite different to that experienced in the conventional business. In particular, it is likely that the house builder will make an initial stage payment when the order is first identified (maybe around 2–3 months upstream of handover to customer). This will be followed by a second payment when the order is irrevocably confirmed (maybe four weeks upstream) and a third payment when the house has been completed on site and the keys handed over. This change of payment profile should make a significant improvement to the house builder's return on capital.
Reduce cost of operation: All house builders incur material and labour wastage during construction, plus further costs at the back end of their business cycle when snagging and (sometimes) major defects need to be dealt with. Often, these costs are not directly attributed or accounted for, but everyone in the industry acknowledges them to be significant. Large-scale pre-assembly introduces factory-based practices which reduce all forms of wastage and allow much of the system testing (electrical, plumbing, etc.) to be done in the factory, with the consequence that snagging and remedial works at site are much reduced, if not eliminated. This is the experience of the hotel and fast-food industries, which have first-hand knowledge of these products. It offers a considerable improvement in operational costs (and customer relations) for the house builders.
On top of this, large-scale pre-assembly reduces the house builders’ need to coordinate supply chains and manage the work on site. In the extreme, the large-scale pre-assembler represents to the house builder a supply-chain of one with all of the administrative efficiencies which that implies.
Please their customers: With improved ROC and reduced cost of operation, the house builders can be freed up to concentrate on their core business, designing and delivering houses which delight their customers. With an influx of new generation 2 products, which lend themselves to flexible interior layouts and a variety of exterior appearances, there should be ample opportunity for the house builders to play to the strengths of these systems and develop attractive, cost-effective, functionally reliable residential properties.
There may be many arguments in favour of modern methods of construction, but do they offer a solution which is sustainable in the wider, long-term, national context? Many of the most attractive solutions involve large-scale pre-assembly at purpose-built factories supported by a transport network which feeds the sites. Initial reactions to this are often negative.
However, on closer inspection, the proposition looks much more promising. Off-site manufacture offers a range of possibilities at the generic and site-specific levels (as itemised in the ‘Sustainability’ box). Each housing development, of course, will have its own site-specific issues, and these need to be considered carefully. However, the positive profile associated with the generic case suggests that, for many real developments, off-site manufacture will compare positively with the traditional solution when all factors are taken into account.
What, then, is the prognosis for this new generation of construction products aimed at the house building industry? The theory is fine, but without tangible investment and development (the leap of faith on behalf of the producers and house builders) nothing will change. We will be left in a few years’ time with an industry struggling to match capacity with demand. It will solve these problems, in the absence of generation 2 products, by the simple economic laws of supply and demand. The price of skilled labour will spiral upwards, but the ‘lag’ associated with re-skilling the industry will constrain the supply of new workers. It is too easy to imagine an industry stretched thin, with a spiralling cost base, trying to build faster with a fixed labour force by cutting corners. It is not a pretty sight.
But it need not be this way. The lessons coming through from experiments with generation 1 products are very positive. Residential developments with a high degree of architectural quality and excellent tenant satisfaction ratings have been delivered. Performance standards which meet and exceed the demands of the revised Building Regulations have been demonstrated. And, most important of all, a family of generation 2 products, designed to compete on price with conventional building processes, is quietly in the pipeline. Whilst most of these remain ‘commercial-in-confidence’, a number of the biggest national house builders are known to be nurturing products which will soon be introduced to the market through joint ventures and shared ownerships.
Whilst there is many a slip between aspiration and reality, there is a discernible wind of change blowing through the housing business. The industry seems to be genuine in its approach to the development of modern methods of construction. The Re-Thinking Construction initiatives (particularly the Housing Forum) which followed in the wake of the Egan Report have done good work. The Rogers Report and, subsequently, CABE, have awakened the nation to our need for good urban design and architecture. Many of the key elements of the jigsaw are coming into place in a manner that happens perhaps only once in each generation. Indeed, it represents the opportunity of a lifetime.
Flexible manufacturing systems have evolved rapidly within the manufacturing industries over the past 20 years. Based on computer-controlled machines and processes they have enabled manufacturers to introduce a variety of different products, all of which can be built, without error, on a common production line. Computer-controlled information systems enable components to be ordered from suppliers, and scheduled to the correct locations on the line, so that assembly workers are fed the right part at the right time with clear instructions.
The effective integration of computer-based design, supply chain management, component purchasing, and assembly process management is fundamental to success. The IT systems for the Meteor Product illustrates the power of such systems. The database generation for the house at the point of order (reflecting customer options selected from a pattern book set of designs) embodies every detail down to the pipe couplings and cable clips.
Each house is individually specified, based on customer choice, from a pattern book of basic designs/options at the point of order
The CAD system enables the house to be sub-divided into modules for manufacture
The level of detail in the database goes right down to the pipe couplings and cable clips …
… allowing every single assembly activity on the production line to be instructed and scheduled.
Generation 1 products: demonstrating potential
‘Generation 1’ products are defined in this article as volumetric systems developed, historically, for other purposes which have been subsequently adapted for use in the housing sector. Such products have demonstrated the potential for off-site manufacturers to deliver residential buildings of high architectural and functional quality. Projects such as Murray Grove, CASPAR 2 and Sixth Avenue have set the pace for producers and developers alike.
The challenge now is to develop a second generation of products which is inherently less expensive to produce and, as a consequence, can compete directly with the price of conventional construction.
Sustainability: does off-site manufacture stand up?
The arguments in favour of large-scale off-site pre-assembly may be broadly divided into two camps:
Significantly reduced material wastage (due to better control of processes in the factory).
More secure, factory-based jobs for the workers.
Better opportunities for continuous product improvement and employee learning (because of more stable workforce and more formalised design/production processes).
Reduced traffic pollution (smaller number of larger loads produces a better ‘footprint’ in terms of congestion, air/noise quality, and road safety than large numbers of white vans).
Much quicker delivery of finished product (typical build cycle reduced to 3–4 weeks on site per house) leads to less disruption for adjacent tenants.
Improved safety for workers (fewer people on site for shorter times).
Improved safety for tenants (fewer vehicle movements; shorter exposure to general site dangers).
Reduced pollution (less debris and site-generated wastage).
Improved quality of build and reduced post-handover defects.
On top of this, factory-built products can be engineered to embody attributes such as good acoustic attenuation, robustness and low energy-in-use, resulting in sustainable designs for the community. The design concept built around the Meteor system (shown in Figure 3), in which environmental responsibility and sustainable thinking dominated the design brief, illustrates the potential.
John Miles FREng
Director of Arup Group
John Miles is a mechanical engineer with a professional background in the design and development of high-performance structures. His recent experience has been in the field of design for manufacture and he has been responsible for the developments of novel factory-built systems for the construction of housing and other residential buildings. John is a Main Board Director of Arup Group, and has recently been Director of the Housing Forum and a CABE Commissioner.