Evolution of the Meat Standards Australia (MSA) beef grading system

Variable beef eating quality was a major concern to many in the Australian beef industry in the early 1990s. During this period consumers were recording their dissatisfaction with Australian beef products by decreasing consumption. Consistency of beef eating quality was also seen as a key problem for the Australian beef trade (Bindon 2001). Domestic beef consumption fell and attitudinal research indicated that consumers were hesitant regarding the product. Concerns regarding health risks had received strong publicity, knowledge of cuts and cooking was declining, product appearance failed to identify quality, consumers were time-poor and demanding convenience, competing products were performing better (Yann et al. 1993; McKinna 1995).

Two of the key strategic imperatives of the Australian meat industry strategic plan were to supply a more consistent product and to accurately describe palatability (Centre for International Economics 1996). The Meat Research Corporation advanced funding to further develop and coordinate an Eating Quality Standards (EQS) program overseen by an Industry Steering Committee. Several active eating quality and consumer testing research projects were combined under the new structure. This program became known as Meat Standards Australia (MSA) when continued under the newly formed Meat and Livestock Australia (MLA).

This paper provides an overview of the issues addressed and parameters utilised in developing consumer standards as a target for grading and the consequent development of the MSA prediction model as the grade delivery mechanism.

Delivering against the established strategic imperatives required several key issues to be addressed. Did consumers actually agree when assessing eating quality? If they did not, then any attempt to guarantee quality would fail. If they did, then how should consumer standards be set and product measured? There was considerable debate regarding the relative merits of objective testing, trained panels and untrained consumers. Did existing grading systems perform? How could they be improved?

The Australian beef industry produces from a diverse base of climatic extremes, breed and animal management systems and processing facilities. Cattle range from young calves slaughtered directly at weaning, to old animals; there are a large number of breeds and crosses, including a proportion of high Bos indicus content cattle. The climatic differences are also extreme, ranging from southern snowfields to northern tropics and from desert to irrigated pasture. Feedlots are used extensively in many areas. These combinations of different cattle type, age and production system all contribute to extreme variability in carcass quality (Bindon and Jones 2001). There is further variation between cuts within any carcass which, in turn, is affected by processing techniques, aging periods and cooking method. This background placed heavy demands on the proposed eating quality system. A simple industry blueprint, such as that adopted by the Meat and Livestock Commission in the UK (Red Meat Industry Forum 2005), was unlikely to produce consistency from such a diverse base.

A major requirement was to identify and quantify factors that could improve quality and consistency. To improve quality, one must first be able to measure it. The concept of identifying, quantifying and applying systematic control to a series of critical control points was first advanced by Morgan (1992) and described as a ‘Palatability Critical Control Point’ (PACCP) approach. A related critical objective was to accurately identify beef of equal eating quality from the diverse pool presented, in order to deliver a consistent guaranteed consumer product.

These issues have been the focus of the MSA research effort since its inception in 1996. A key principle has been that any system developed had to deliver consistent palatability to the consumer. A grading system which simply described carcasses of similar appearance was not an acceptable option, with preliminary studies concluding that it would be dangerous to introduce a branding approach to the meat industry unless tenderness could be guaranteed within reasonable limits (SMART 1994). The MSA approach was in contrast to traditional grading system objectives which aimed to facilitate trade by describing the commercially important attributes of the carcass (Price 1995).

Every piece of beef produced is ultimately judged by the consumer when eaten. No amount of product promotion can counter the direct experience at this point. It is therefore necessary to understand the perception of consumers as a group, and the relative differences between individual consumers, to develop and assess a grading system that is focussed on describing eating quality. In the past, consumer data were often regarded as too noisy to be used effectively as a research measurement tool. However, the alternatives (either trained panels and/or objective measurements) were found to lack validity, particularly over extreme ranges in quality (Hwang et al. 2003). In deciding the form of evaluation to be used, consumer data had a downside in terms of reliability, however, this was more than compensated by gains in credibility. A detailed description of the development of the consumer testing protocol and subsequent palatability score is described by Polkinghorne et al. (1999) and Watson et al. (2008c).

The MSA program has been built on the premise that the grade should reflect only the consumer-assessed result, without being prescriptive in regard to the combination of factors that might affect the result. This provides for total flexibility in production systems, with the end product benchmarked against consumer satisfaction levels.

Having agreed on a consumer testing methodology to measure palatability, the MSA research program concentrated on evaluating all possible factors, or critical control points, against the consumer test benchmarks. Factors found to be correlated with eating quality were initially combined as fixed ‘pathway’ parameters and later utilised as interactive inputs in the series of grading prediction models developed.

The differential effect and range of interactions found between grading input parameters for various muscles as related to eating quality presented a major challenge to the notion of carcass grading and led to the conclusion that if a grading system aimed to assist consumers, it had to successfully grade individual cuts within a cooking method framework. The difference in ranking and extreme differences in relative importance between various identified control points or prediction variables also challenged the possibility of using an indicator or index muscle to predict other muscles.

This is in agreement with a study of 10 muscles by Shackelford et al. (1995), who reported that systems that accurately predict the tenderness of striploin (M. longissimus lumborum) of a carcass will likely do little to predict the tenderness of other muscles. Shorthose and Harris (1990), in a study of 12 muscles, also suggested that the M. longissimus lumborum was unsuitable as an index muscle, proposing instead the M. semitendinosus as more appropriate, partly due to it being restrained under tenderstretch and Achilles hanging systems and accessible from the carcass or cut.

While initial MSA trial work addressed only grilled striploins (m. longissimus lumborum), the program was expanded to include a range of muscles prepared by a variety of different cooking techniques (roast, stir-fry, slow-cook, corn, thin-slice and yakiniku). Contributing research activity has progressed in all areas from farm to plate over time. The detailed findings of specific MSA studies are covered in associated papers (see Colditz et al. 2007; Ferguson et al. 2007a, 2007b; Warner et al. 2007; Hwang et al. 2008; Park et al. 2008; Polkinghorne et al. 2008; Thompson et al. 2008a, 2008b, 2008c; Thomson et al. 2008; Watson 2008; Watson et al. 2008a, 2008b, 2008c; R. Polkinghorne, J. Thompson and R. Watson, unpubl. data). Considerable use has also been made of published data and several international scientists have assisted generously. As knowledge has improved, the level of detail and form of calculation within different versions of the MSA Grading Model has been modified accordingly.

Preslaughter issues such as breed, feed and management effects have been at the forefront of industry interest and debate.

Early commercial benchmarking studies of grilled striploins (M. longissimus lumborum) indicated a negative relationship between eating quality and levels of B. indicus content. Similar results have been reported by others (see review by Burrow et al. 2001). No significant difference in consumer-assessed eating quality has been evident between British and continental breeds or their crosses, after adjusting for carcass weight, ossification and fatness traits that are influenced by breed type. This was in accordance with results reported by Koch et al. (1976), Adams et al. (1977), Cuthbertson (1994) and Wheeler et al. (1996, 2001) amongst others. Therefore, the grading model incorporated an adjustment for B. indicus percentage, which was applied differentially by muscle (Thompson et al. 1999). No other direct breed adjustments have been incorporated in the model at this stage.

A similar eating quality effect was found in some tropically adapted Bos taurus breeds, such as the Belmont Red which also had some visual B. indicus characteristics, including a hump. A factor relating hump height (M. rhomboideous) and carcass weight was developed as an estimator of equivalent B. indicus content and also incorporated as a cross check in the grading model. The relationship of hump as a phenotypic measure to tenderness scores was previously proposed by Sherbeck et al. (1996) in a study utilising cattle of varying Hereford and Brahman percentage.

The literature is not clear on the effect of growth path on palatability. Fishell et al. (1985) reported that faster growing animals were more tender than slower growing animals. This was in contrast to others (Calkins et al.1987; Moloney et al. 2000) who found no relationship between growth rate and a measure of tenderness. As discussed by Perry and Thompson (2005), variation in growth rate can be exhibited as both between- and within-group variation. They concluded from analysis of growth path data collected on ~7000 animals that between-group variation in growth rate was likely to be driven by intake. They also found that, as long as comparisons were not confounded by age, there was unlikely to be large differences in eating quality. However, this contrasted to within-group, where all animals had access to the same feed resource and differences in growth rate were likely to be driven by the individual’s genetic potential for growth. They found that within a group the faster growing animals were more tender (i.e. higher tenderness scores and lower shear force). This was consistent with Shackelford et al. (1994) who showed that, within a group, faster growing animals had lower calpastatin activities and their meat was more tender. Perry and Thompson (2005) concluded that the small effect of growth rate on palatability within a group could be exploited by a grading scheme such as MSA which rewarded individual carcasses with a higher palatability score.

The use of dentition, ossification and actual age in relation to carcass weight were also investigated as potential tools to assist in predicting consumer scores. Actual age and ossification aided prediction slightly and similarly, but age was seldom available for commercial cattle; dentition demonstrated much poorer correlation. Recent analysis by Park et al. (2008) showed that ossification had a small but consistent negative effect on palatability, indicating its usefulness as a predictor of palatability.

The exception to this general approach is for calves slaughtered before weaning. These animals are typically 10 months old or less, with low ossification and carcass weights from 170 to 220 kg. In southern Australia, they are often referred to as ‘milk fed vealers’ (MFV). When compared with weaned calves of equivalent weight and ossification an unexplained, predominantly positive, eating quality difference was found in various muscles (Watson et al. 2008a). The Committee overseeing the development of the MSA system suggested that a method of including this in the grading model was to create a separate category for ‘MFV’, which would segregate calves of genuinely very young age from older calves with equivalent ossification.

Several model inputs were found to be influenced by hormonal growth promotant (HGP) implant use. In most instances, HGP use was found to increase carcass weight and ossification score, and to reduce rib fat and marbling (Thompson et al. 2008a, 2008c; Watson et al. 2008b). Studies by Roeber et al. (2000) and Samber et al. (1996) have also reported increased ossification associated with some HGP regimes, while Tipton et al. (2002) found little effect. MSA data indicates a greater effect on ossification with increased days between initial implant and slaughter and with increased number of implants (Watson et al. 2008b, Thompson et al. 2008c). This trend is also seen in a study on repetitive implant use by Platter et al. (2003b). Studies by Roeber et al. (2000), Platter et al. (2003a) and Tipton et al. (2002) all report an increase in carcass weight and decrease in marbling score with various implant strategies. No significant effect on rib fat was reported in any of these studies, which was at odds with the MSA data.

Changes to carcass weight, ossification, marbling or rib fat modify the grading model prediction score, suggesting the possibility that the model would account for any consumer score differences relating to HGP use without a direct additional HGP model adjustment. However, they did not; greater accuracy was obtained by adding a separate HGP adjustment in addition to the changes in carcass weight, ossification, marbling and rib fat. The impact was found to be detrimental to eating quality, to vary by muscle and to be reduced, but not eliminated, with increased aging.

The MSA trials used to develop the grading model HGP calculation have been reported by Thompson et al. (2008a, 2008c), Watson et al. (2008a, 2008b), R. Polkinghorne, J. Thompson and R. Watson (unpubl. data). They included B. indicus cross and British breed cattle from grassland and feedlot production systems. Several cuts were tested with alternate cooking methods and aging periods applied in various combinations. A range of alternative implants and implanting strategies was also incorporated in both commercial and research cattle.

Marbling and external fat levels are also common components of industry trading specifications and grading systems. Both are affected by genetic and animal management factors. Marbling scores in AUSMEAT (Anon. 2005) and USDA (Romans et al. 1994) scales, both measured at the quartering site, were recorded for a majority of carcasses used to source consumer test samples. Fat depth was also recorded at the P8 (Rump) and 12/13th rib sites for a majority of carcasses. Analysis of the consumer data confirmed a relationship between marbling and consumer scores which varied widely among cuts. While contributing to the consumer score, marbling alone explained only some of the differences between the same cut from different carcasses and very few of the differences between cuts from the one carcass (Watson et al. 2008a).

This was consistent with results reported by Koohmaraie et al. (1995) which stated that connective tissue and marbling only accounted for 20% of the observed variation in meat tenderness. It was also consistent with the findings of Thompson (2004), who stated that marbling accounted for ~15% of the variation in beef tenderness. Miller et al. (2000) also reported similar findings but noted that, while the relationship between marbling and consumer palatability was low, it did appear to be consistent. Thompson (2004) reported a curvilinear relationship between marbling score and flavour, although this was not apparent in later analyses (Park et al. 2008). A muscle dependent linear marbling adjustment was developed as a component of the grading model.

Differences between heifers and steers (neutered males) were also evaluated across the full range of samples tested by consumers. While the direct eating quality effect of gender was found to be small, some improvement in prediction accuracy was obtained by including gender in the prediction process. The effect is applied differentially by muscle and is further adjusted according to ossification (Watson et al. 2008a). Only females and castrated males have been tested to date.

Several studies (Butchers et al. 1998; Ferguson et al. 2007a, 2007b) have evaluated the impact of stress and various management practices on eating quality. These experiments have analysed MSA eating quality scores in relation to flight speed, mixing of stock at various periods before slaughter, differing feeding and curfew practices before slaughter and application of severe stress immediately before slaughter. Significant results have been reported in several studies, with consumer evaluation often detecting differences to a greater degree than laboratory shear and compression tests. This appears to relate to the sensory experience, including a juiciness component, and is a more complex judgement than shear value alone.

Findings from the studies have been used to set guidelines for animal management and to establish minimum standards for grading eligibility. These include criteria for supply of cattle through saleyards, time off feed before slaughter and time restrictions for mixing of groups before slaughter.

The relative importance of pre- and postslaughter issues has long been a source of debate within both the commercial industry and scientific community. MSA research suggests that both are important with interaction and interdependence for delivering consistent and predictable eating quality. This highlights the importance of viewing industry segments as inter-related elements of a single production chain and the value of cooperation and clear communication between all segments.

The obvious point of interface is represented by carcass changes, reflected by pH and temperature relationships between knocking and completion of rigor mortis. Animal feeding, temperament and handling on farm, during transport and in lairage all combine to influence blood glycogen level at slaughter. Subsequent abattoir interventions including electrical inputs from restrainers, stimulation systems and rigidity probes interact to determine the rate of pH decline. Time and temperature relationships have major implications for eating quality and for eating quality after aging, due to influences on both myofibrillar shortening and enzyme activity.

The importance of understanding and controlling prerigor conditions in beef carcasses to improve tenderness has been discussed in detail by Marsh (1954) and Marsh et al. (1987) and others for a considerable period. Dransfield (1994) states that the conditions during rigor development are the most important factors controlling tenderisation and aging for most commercial meats.

Early MSA efforts towards developing approved PACCP pathways mandated electrical stimulation of all carcasses. However, an early trial produced the unexpected result of decreased eating quality with stimulation. Further analysis and follow-up trials (Hwang and Thompson 2001a, 2001b) indicated that, whilst the excessive use of stimulation eliminated the risk of cold shortening, it created a new problem of heat shortening, with an associated eating quality decline due to increased moisture loss and reduced aging potential from enzyme autolysis. Product sourced for consumer testing from a range of cattle types and suppliers and slaughtered at different abattoirs utilising high voltage, low voltage and no stimulation after slaughter also demonstrated conflicting results depending on circumstances. These effects have also been reported by Marsh et al. (1989) and Takahashi et al. (1984). Martin et al. (1983) reported a positive association between tenderness and glycolytic rate. Further work by Simmons et al. (1996) reported interactions with pH, time and temperature in relation to sarcomere length, µ-calpain and calpastatin activity and shear force, which again indicated an optimum relationship between detrimental extremes of glycolysis. More recently, Thomson et al. (2008) demonstrated that extreme temperatures at rigor caused a crossover in toughness; soon after rigor heat-shortened product was more tender but because of autolysis of the calpain system, it did not age as much as muscle that went through rigor at 15°C.

The conclusion drawn by the MSA Pathways team was that, rather than prescribe a stimulation regime which could have a positive or negative impact depending on circumstances, it was better to prescribe a defined pH–temperature relationship. This was in agreement with Tornberg et al. (2000), who stated that not only does the temperature fall, in itself, influence the pH decline, but an early attainment of low pH at elevated temperatures (>15°C) can also cause denaturation and/or autolysis of the enzymes and therefore a decreased tenderness. They concluded that this provided a plausible explanation for an optimum pH decline during rigor in relation to tenderness. Wahlgren et al. (1997) also reported that a combination of low pH and high temperature during rigor development could detrimentally affect meat tenderness measured by both shear-force and sensory techniques. In their study, they found an intermediate rate of pH fall produced more tender beef than a fast or slow rate, which is also consistent with studies reported by Marsh et al. (1987).

Thompson et al. (2006) reviewed the factors that impacted on glycolytic rate in carcasses. Studies by Daly et al. (2002) showed that higher initial glycogen concentration resulted in a faster rate of pH decline. Similarly, Daly (2005) showed that heavier carcasses exhibited a faster glycolytic rate, which was largely driven by the decreased cooling rate of larger carcasses. They suggested that the problem of very rapid pH decline often seen in long fed beef carcasses may simply reflect the decreased cooling rate of these fatter heavier carcasses and that the problem may not require a metabolic solution but rather something as simple as increasing temperature loss from the carcass during the development of rigor. Thompson et al. (2006) proposed that any effect of genotype on glycolytic rate was likely to be driven through differences in muscle fibre type, with anaerobic type IIB fibres having a faster rate of pH decline.

Based on the literature, an MSA standard (described as an ‘abattoir window’) was established, requiring temperatures to be above 12°C and below 35°C at the point that pH reached 6.0 measured in the loin. The degree of stimulation required to produce this relationship was found to vary widely, from none in the case of many heavy grain fed cattle, to 40 s for light, lean grass-fed cattle.

Carcass side suspension effects on eating quality were described by Smith et al. (1971), Hostetler et al. (1970) and Bouton and Harris (1972). The factors which impact on the tenderstretch response were reviewed by Thompson et al. (2006). Depending upon the chiller conditions, tenderstretch generally results in an improvement in palatability and also a substantial reduction in variance of palatability scores. Thompson et al. (2005) considered that tenderstretch provided insurance for palatability scores, particularly where the carcasses were likely to be exposed to extremes in processing (e.g. hot or cold rigor temperatures).

Inclusion of tenderstretch within the MLC Blueprint in the United Kingdom has encouraged widespread adoption of tenderstretch by major British retailers. The blueprint recommendation resulted from trial work that demonstrated substantial improvement in comparison to Achilles suspension (Cuthbertson 1994). While utilised widely in the UK, tenderstretch, or pelvic suspension, had not been adopted widely within the Australian industry.

Several MSA trials were conducted to quantify the potential to upgrade cuts by utilising the technique. Following encouraging initial results, the trials were broadened to include a full range of cattle types, a variety of carcass weights and all major cuts. Aging periods were also varied and, when analysed, indicated an interaction between muscle and suspension method (Watson et al. 2008a).

To counter industry concerns regarding rail space in chillers, some testing was also done on an alternative ‘Tendercut’ method described by Wang et al. (1994). Differences were also evaluated between two commercial variations of tenderstretch, suspending by the pelvic ligament v. suspension from the aitch bone (obturator foramen). The differential effect of each hanging method was incorporated into the MSA grading model with the effect varied by muscle. A specific aging estimate was also developed and implemented on a muscle by hang basis.

Development of reasonable aging estimates proved difficult, due to considerable variation in the data. This is not surprising given the strong influence of temperature and pH inter-relationships in the prerigor period, found by Hwang and Thompson (2001b). Differences between aging potential of various muscles have also been reported by Shorthose and Harris (1990).

Dransfield (1994) and Bouton and Harris (1972) reported that aging rate differences between muscles were related to muscle type, with calpain and calpastatin levels also differing between red and white fibre muscle types. Those authors indicated that this was consistent with very low aging in M. psoas major and lower aging in M. biceps femoris than in M. longissimus lumborum, which was also consistent with MSA consumer results.

Shorthose and Harris (1990) and Bouton and Harris (1972) have documented changes in toughness with animal age, due to differences in connective tissue strength which varies considerably between cuts. As proteolysis is thought to relate predominantly to changes in myofibrillar toughness, the variation in connective tissue quantity and strength between muscles, with further variation in animal age at slaughter, could be expected to influence postmortem aging.

There appear to be few studies reporting interaction between carcass suspension method and aging rate of specific muscles, although these are strongly evident in the MSA data. Bouton and Harris (1972) reported differences in aging rate between muscles and between the same muscles under different suspension treatments. They reported that the aging effect as measured by shear force was greater for muscles from Achilles hung carcasses than from tenderstretched carcasses and greater for M. longissimus than for the M. semimembranosus. They also report a decline in aging rate over time. These findings are confirmed by the MSA data and hence are part of the MSA model.

The cooking method used was also found to directly influence eating quality outcomes. The effect differed widely between muscles, both in extent and as to which cooking methods were most favourable to the cooked outcome. A pertinent finding was that, while selection of the most appropriate cooking method might improve the result for an individual muscle, it did not remove differences between muscles.

In light of these results, it was considered necessary to link the consumer grade to a cooking method. For some muscles the grade would often vary according to the cooking method chosen. Accordingly, the model output and estimation process was developed to provide a ‘muscle by cooking method’ outcome. This was described by Watson et al. (2008a). It was believed that the beef retailer could use this information to select and prepare muscles for their most appropriate use, thereby assisting the consumer obtain the best result.

Efforts to predict beef eating quality are not new. Neither is the aim of consistently meeting consumer expectations. A wide range of government and private grading schemes, industry blueprints, supplier specifications and day-to-day specifications have all been used without fully delivering on the aim. Even the USDA quality grading system, which has a major commercial impact in the USA and international markets, only accounts for 10–30% of the variability in beef tenderness (Miller et al. 1996).

Early trials which preceded the MSA program attempted to grade beef by a mix of conventional grading parameters. Hearnshaw et al. (1995) concluded that the NSW gold branding scheme had little value, based on consumer supermarket studies. It was concluded that at low-fat levels, visual carcass appraisal and slaughter floor data could not reliably categorise eating quality (Australian Meat Standards 1997). Initial MSA efforts to provide a more consistent product to consumers centred on construction of production ‘Pathways’, each pathway consisting of a set of criteria that had to be met in order for a cut to receive a grade. This approach was similar in concept to the MLC blueprint (Cuthbertson 1994) and strategic alliance project of the American NCA (1994), which linked several critical control points to reduce variation. In the MSA situation, the steps were mandatory rather than best practise recommendations. Typical criteria included maximum B. indicus content, minimum marbling levels and minimum days aging. From consumer testing, several successful pathways were formulated. Product which met the criteria also achieved the determined levels of satisfaction confirmed by consumer test.

This was a significant achievement for the industry, in that it then had a means to provide a consumer guarantee. The downside was that a large percentage of product failed to grade, by falling outside one or more of the pathway criteria. Despite failing to meet one criterion, however, the failed product also often met consumer standards because it exceeded minimum standards for other criteria. Additional criteria combinations were tested to provide multiple pathways to attain a common grade.

The number of pathways required to cover all options and then to accurately predict multiple cuts or muscles is very large, however, making management of a developed system complex. After evaluating a range of approaches including construction of decision trees, the concept of an interactive model was formulated and developed (Polkinghorne et al. 1999). Under the model approach, all factors are considered interactively, allowing full compensation (positive or negative) rather than imposing rigid threshold levels. If the data are sufficient to provide an understanding of each effect and any interactions, the model approach can provide a high level of consumer protection and satisfaction, while reducing the proportion of acceptable product rejected. A significant benefit is that each muscle can be independently estimated by utilising alternative input factors or different weightings for common factors. This removes the inherent weakness of attempting to apply a common grade to a carcass known to comprise a collection of very different consumer products.

Development of the MSA model has evolved since the release of the original 12-cut version (Polkinghorne et al. 1999). The current model is the fourth commercial version, now predicting 135 ‘cut by cook’ combination consumer outcomes for each graded carcass. The statistical processes used in developing the models have been reported by Watson et al. (2008a). Prediction accuracy and the inputs used have progressed from analysis as additional research data have been accumulated.

Research priorities have also been set or refined to rectify deficiencies in available data or to examine additional issues. Much of the research has been conducted in collaboration with several institutions, with MSA also providing consumer score data for many experiments established to examine a wide variety of production issues. A Pathways committee comprising principal Australian researchers and industry representatives has overseen the research program from the outset, with further input and review from several international meat scientists.

At its current stage of development, the model is providing commercially useful accuracy across all major carcass muscles cooked by most common methods. As such, it has gone a considerable distance towards providing the industry ideal of a rapid, automated, tamper-proof, noninvasive, accurate instrument as expressed by Koohmaraie et al. (2005).

The model estimates for individual cuts have provided a strong base from which to simplify retail description systems, while reducing the need for consumer beef knowledge and enhancing satisfaction with the cooked product. A commercial trial reported by Polkinghorne (2006) and Polkinghorne et al. (2008) elaborated on advanced practical application of these principles.

Industry interest in the program has been strong at all times. This has provided a high degree of scrutiny and, at times, challenge. The judgement delivered by consumers has, however, been well accepted and many practises have been modified across all industry sectors in response. A major program benefit has been the provision of a consumer view to most aspects of beef production.

In many cases, MSA grading has not been adopted in full due to perceptions of required change or available commercial benefit. Even in these situations there has still been strong uptake of many elements of the program. MLA commissioned survey results suggest that consumer satisfaction has substantially improved over the period of the program (Millward Brown 2007). Whilst this cannot all be attributed to MSA, it is believed to be a major contributing factor.

It is estimated that 40% of all eligible carcasses destined for the domestic market are now graded by MSA. Grading numbers have grown continually to over 700 000 carcasses in 2007. A major impact has been in training with over 20 400 training modules delivered to 8447 processing and retail industry personnel. Producers also receive training as part of the registration process and there are over 10 000 registered MSA members, the majority producers (C. Dart, pers. comm.). This has led to a dramatic increase in the level of understanding of eating quality issues at all points of the production chain. Grading numbers are increasing steadily and adoption now appears to be growing at an increasing rate, as growing capability in the model output is matched by increased industry awareness of commercial opportunities.

Despite accelerated use in the wholesale trade, visibility of MSA at retail is generally low. It is being used predominantly to support private brand initiatives or to underpin existing channel partner offers, rather than as a retail brand in its own right. Growth has been particularly marked in the food service area leading to substantial premiums for MSA graded product at wholesale and farm level (Dart et al. 2008).

The MSA program has acted as a catalyst for substantial change in all sectors of the Australian beef industry. Provision of a defined consumer target and testing protocol has served to focus the industry and research efforts on the eating quality result of production, processing and retail presentation alternatives. The program has also served to encourage and facilitate research and commercial industry cooperation, to the benefit of both sectors. Eating quality is now integral to beef industry operation and planning.

The sensory response of the Australian consumer to beef is now much better defined and understood, enhancing the ability to define and provide superior value. The industry’s ability to deliver a consistent quality product of known eating quality has dramatically improved with a commensurate opportunity to modify traditional description and pricing regimes, simplifying the retail offer and reducing the need for consumers to have any background knowledge of beef cuts and cooking relationships. The eating quality cause and effect relationships of practices, from farm to retail, and their interdependent nature are now better defined, providing a base for development of accurate value based pricing systems for each segment.

The program is ongoing, with continuing research to further extend and improve the predictive accuracy of the model across the full range of livestock, production environments and processing practises. There is also growing interest in the potential to utilise the same or modified approaches in export markets and in collaborating with other countries to test the response to local consumers and product.