6.Discussion and Conclusions
6.1 The Improvement Requirement and ISO 50001:2017 Draft
The requirement of showing an improvement to be recertified with the ISO energy management system has both drawbacks and advantages. The advantages include a requirement that in many cases means that the organization may be able to save money and that the standard allows for a lot of freedom when making decisions as an organization (Nachiappan and Muthukumar 2016). This freedom in conjunction with the improvement requirement has some complications and there are other aspects like the choice of model to be considered (ISO 2014c & ISO 2014b).
6.1.1 Possible Implications
As already explored there are financial advantages of the ISO Energy Management System (Nachiappan and Muthukumar 2016). The requirement to show an improvement could be translated as a requirement to monetarily optimise processes as was already presented in 4.5.3. It must be said that these benefits are not felt by every organization that follows the standard (Yuriev and Boiral 2018). There is also a proportion of organizations that find the investment of changing current processes or equipment is initially too expensive (Therkelsen and McKane 2013). This is rather contradictory of ISOs original goal of the democratic creation of standards and their claims of supporting equality because this requirement may not be possible to fulfil for some organizations that take this approach (Latimer 1997 and ISO 2017b). It could also be considered a barrier to trade for those organizations that are not able to afford the initial investment. Organizations however also have the option of a cultural change also leading to an improvement (ISO 2017a). Forcing this process of change for an improvement though has other TBT related repercussions in relation to the SEUs.

The process of setting up the Energy Management Systems requires the identification of the SEU and changing them when needed (ISO 2017a). Forcing this change that is related to the equipment and facilities used in an organization could be considered a way to encourage the transfer of technology (GATT 1994). It is however a requirement where if the improvement is not shown they lose certification (ISO 2014c). This could in turn be considered a technological barrier that is the reciprocal effect of financial limitation explored above (GATT 1994 and Therkelsen and McKane 2013). Due to ISOs position as one of the organizations working in the information centre in Geneva it is assumed that one can trust ISO to respect this agreement but there is little external scientific evidence about this particular requirement (GATT 1994). The requirement to show an improvement has also only been around for 4 years so the actual effects my not have been felt, yet (ISO 2014c).

Showing an improvement as a part of ISOs Energy Management System concentrates on measureable data relating to energy. The requirement focuses solely on the improvement of the EnPI not on other aspects of the EnMS (ISO 2017a). There is a use of tangible information in the form of the data about the relevant variables, the static factors and energy. The chosen EnPI is compared against itself meaning that the improvement as shown by the data is clear because it is the data against itself giving an individual reference point and context (ISO 2017a). This comparison ignores Cohen and Nagels’ (1935) system of facts but adopts the practical applications of Lewis (1929) and the concepts of pragmatism. The measurement of this data as presented in ISO 50015:2014 are also not actually fixed values but are approximations and there is a certain level of uncertainty. The mentioned standard explores approximations stating that there are cases where an approximation is appropriate and advise discretion with the use of sensitivity analysis to identify levels of uncertainty from measurements and asses them. This means that a loss in certification is potentially based on data that is uncertain (ISO 2014a). These measurements are not the only reason for uncertainty.

ISO 50015:2014 also mentions that models can be uncertain. The uncertainty of the multivariate regression model is assessed though the variation. As was displayed by Flick et al. (2017) and Dehning et al. (2017) even companies working in the same field use different methods to analyse both variation and correlation, while they also used different variables. This is an advantage because it means organizations can create their own systems. Applying this freedom to the context of having showing an improvement, it might be more practical for companies to play with the data they have than to make an actual improvement. This possibility of using data to show an improvement comes from the range of options for relevant variables, the analysis of the variation and the correlation as shown in 4.5.5. Whether or not this is an actual possibility requires more research.

This freedom of choice is one of the main selling points for any ISO system (ISO 2016). This means that the Energy Management System can include the as described model but there is also the possibility of using other models, which increases the variation of data processing methods. Another aspect of this freedom is the choice of organizational and physical boundaries of the system (ISO 2017a). This opportunity in relation to the boundaries of the system could be a loophole that organizations take advantage of. This however needs more research and information from the third party auditors to know if they have observed this behaviour.

Apart from freedom the systems requires many perspectives about energy to be taken into account when creating the system. Energy management as apart of ISO 50001:2017 Draft is not just about its use but also about efficiency and consumption. The EnMS includes many different aspects including the plans and policies (ISO 2017a). The improvement requirement doesn’t include these other aspects of the system as it only focuses on the data that is associated with EnPIs. This is advantageous because this aids in being able to show a clear improvement as a part if the system (ISO 2014c). This is neglectful of the other aspects of an EnMS but these other aspects as shown by the Plan-Do-Check-Act model in the introduction of the system should lead to the measurable improvement. The time frame for showing this improvement is 3 years, which might be long enough to see these improvements without having to reference other aspects if the EnMS (BSI 2015).
6.1.2 Conclusions
The recertification requirement of ISO 50001:2017 Draft has an emphasis on data. By having a requirement for showing an improvement, it is treating this data as a definite piece of information, ignoring the complexities of the process of modelling EnPIs and the other uncertainties with relation to measurement. It also neglects to consider the variations that the freedom of the system with which organizations are able to display this improvement allow. Such requirement may not be appropriate for a generic process standard. Use of an EnB and EnPIs does direct how the improvement should be shown but as was discussed there are still so many options when modelling for energy management. The implications of this requirement will need more exact observation over the next few years, as more and more organizations will be recertified with this requirement.

6.2 The Improvement requirement, ISO 9001:2015 and 14001:2015
The perspectives of quality management and environmental management are different when it comes to showing an improvement because of the different focuses of each standard (ISO 2015a & ISO 2015f). Quality management has to considered the constantly requirements of customers (ISO 2015f). Environmental management includes all the variables included in energy management and many others (ISO 2015a).
6.2.1 An Improvement Requirement and Quality Management
The ISO quality management system standard concentrates on customer requirements as apart if its main purpose. This is embodied in the first quality management principles. The requirements of the customers could be considered the reference point comparable to the EnB and the proportion of these requirements that are fulfilled could be a quality management indicator similar to the EnPI. The customers’ wishes are referenced in the standard itself as being subject to continuing change. This means that to showing an improvement would require a continually developing assessment system. This assessment would also require the data that is used for the evidence-based decision-making that is another principle of quality management (ISO 2015f).

Making decisions based on ISO 9001:2015 should include an analysis. This analysis is not prescribed with the same detail as with energy management, which includes this requirement to show an improvement (ISO 2015f ; ISO 2014c). If a regression model would be used as apart of showing an improvement with ISO quality management system standard, to cope with the changing conditions it would require a constant changing model and its variables. This constant change would not allow for such a comparison of a baseline time period to a reporting time period but could be somehow adapted to the customers’ requirements. One could assess the improvement differently with another type of model but modelling customer requirements as apart of a quality management system requires further study.

The requirement for improvement could be combined with the quality management principle of relationship management, which is also applicable to the start and end of the process approach with the sources and receivers. How well communication is kept between there parties could be apart of an improvement requirement. This process approach also grants such assessments for improvement at the other stages (ISO 2015f). Each stage or just one stage could be subject to a requirement of showing an improvement like with the improvement requirement of the ISO energy management system standard only being applicable to the EnPIs. Some stages could be grouped and assessed together due to their similarities. In both the source and receiver stage there is the inclusion of customers whose requirements are subject to continual change. This could have an influence on the other stages within the process approach. This revisits the challenge of showing an improvement in the same way described by the ISO energy management system standard. It would require a very complex ever-changing model.

The middle stages of input, activities and output could be the subject of a requirement for showing an improvement. Especially in the case of input and output there is a clear physical aspect that could be subject to a requirement of showing an improvement this would extend away from the process standard as it address the resources and products associated with a product standard. ISO 9001:2015 is however a generic standard that is used by over 1,000,000 organizations worldwide (ISO 2016). Even within the schematic representation in section 4.5.1 acknowledges that there is a huge variation of inputs and outputs. The activities could be subject to an improvement requirement in terms of improving activities effectiveness in the constantly changing conditions. This could be displayed with a regression model but there are many possibilities for variables that would need to relate to the other parts of the approach especially the constantly changing requirements of the customers. This could explain the claim from Hoyle (2017) that showing any improvement with this system is difficult.
6.2.2 An Improvement Requirement and Environmental Management
Environmental management encompasses not only energy but also pollution and environmental protection. Many environmental aspects should be considered as a part of the system. These aspects could be grouped as prescribed in Annex A of the ISO 14001:2015 document. Like with energy, these aspects could be reassessed in comparison to themselves before a change but as can be seen by the assessment of energy there are many different variables to be considered (ISO 2015a). There is the possibility to create regression model with an infinite number of variables like the one presented. Each variable of environmental aspects could be grouped and then assessed separately but this might be contradictory of the system. It states that the whole life cycle should be considered simultaneously so that damage is not off set. Measuring this damage is however not made clear (ISO 2015a).

The document itself is vague on many aspects related to measurement with many other details that are essential for a requirement of improvement. ISO 14001:2015 focuses on mitigation activities, which reduce the severity of something. By requiring correct measurement like that prescribed in 50015:2014 there could be opportunity for a similar Environmental Performance Indicator like and EnPI from 50006:2017 to be set up. Another focus of the system is external obligations. Following these is also a goal of the system. These obligations could be a reference point and instead of using an EnB. However in a Sub-Saharan context there is an institutional weakness that wouldn’t support the obligations being a reference point for improvement (Tene et al. 2018). The Environmental Management System would require an increased focus on data and quantitative measures for this.

One of the current focuses of 14001:2015 is the long-term impact of the organization on the environment. This is embodied in the consideration for the life cycle of products produced by an organization that would like to be certified with the ISO Environmental Management System (ISO 2015a). Recertification occurs after 3 years (BSI 2015). This might not be an appropriate time to consider the full environmental impact of an organization. This time frame however long-term is not defined within the standard this means that an assessment of whether 3 years is long enough to assess these environmental impacts and potential improvements is needed. The socio-economic impacts also need to be considered.

ISO 14001:2015 considers the socio-economical impacts of the organization with relation to the environmental management system. The socio-economic impacts such as customer satisfaction are also subject to analysis when researchers study the study this system (Boiral et al. 2018). This relates to the requirement of the system to communicate actions appropriately to the relevant parties, which includes these customers (ISO 14001:2015). This already set communication acts as a signalling tool for external parties. Requiring extra documentation and analysis to show an improvement my counter act the non-environmental benefits of the system due to the fact some organizations already find the process cumbersome (Boiral et al. 2018).

6.2.3 An Improvement Requirement and Process Standards
Only process standards were selected for the assessment of the improvement requirement from ISO. As was shown in 3.1 there is very little interaction between process and product standards. The ISO 50001:2017 Draft concentrates on data that could be the processes easiest to access physical manifestation. In comparison to product standards they don’t have the same root in the physical realm and there are many different variables to be accounted for (Navdi and Wältring 2002 & Flick et al. 2017 & Dehning et al. 2017). This is not only applicable to energy but also quality and environmental management. On the other side requiring an improvement for recertification may strengthen the signalling tool that is the certification. A third independent party already audits the standards and maybe this requirement isn’t needed to strengthen the signalling of the certification (ISO 2018a).
6.2.4 Conclusions
The quality management standard from ISO focuses on the requirements of customers that is subject to continuing change. This means that such a requirement as prescribed in the ISO 50000 family with the EnB and EnPI cannot be used in relation to the ISO quality management system standard. There may be other methods to display an improvement within the context of quality management but this will require more detailed research into this topic. As for the application of such a requirement being applied to the ISO environmental management system standard there are so many variables that would need to be included in the model to display the improvement. The system for doing this with a regression model is plausible but the same problems that are applicable to the energy management system are also applicable here. These problems are the reason why one can conclude that a requirement to show an improvement is not completely appropriate for process standards.

6.3 ISO and the Science of Improvement
To answer the third question about ISO and their standards relation to the science of improvement the main concepts presented in section 2 will be revisited. These concepts are the two apposing types of logic being deductive and inductive; the listed prepositions from Perla et al. (2013) and pragmatism. The work of Shewarts, Deming (1986), Kaizan (1986) and Langley et al. (2009) will also be reviewed. What will also be discussed is ISOs role and whether they help to improve standards.
6.3.1 Inductive and Deductive Logic
An integral part of the science of improvement is the constant reassessment of reality against prediction. This is the movement between what should be and what is (Moen and Norman 2006). In terms of the three examined standards from ISO this repetition is parallel to the constant reassessment implied in the Plan-Da-Check-Act Model and by the minatory audit every 3 years (please refer to Annex 1 ; BSI 2015). The logic behind the standards however doesn’t completely follow the deductive and inductive logic cycle though time. It rather jumps between intended and actual outcomes in the same way that the science of improvement oscillates between predictions and observations. This is demonstrated in the following figures. To ease the comparison figure 1 has been repeated and is above figure 8. The actual outcomes are the real situation like the observations, while the intended outcomes are related to the predictions as being a theoretical situation. ISOs standards can be considered a practical application of the cycle between what should be and what is.
Figure 1. The repeated graphical representation of the oscillations between deductive and inductive logic adapted from Moen and Norman 2006 (Pg. 3)
Figure 8. Conceptualization of the difference between Actual Outcomes and the Intended Outcomes

6.3.2 Propositions and Pragmatism
The systematic relationship between facts from Cohen and Nagel (1935) is not applicable to all ISO standards. This theory that a fact should be true without need for further clarification excludes the practical context of ISO process standards such as the quality, environmental and energy management norms presented in this thesis. For product standards like those presented by Nadvi and Wältring in 2002 there is the opportunity for an absolute fact because these standards prescribe physical attributes of a product. A translation of an absolute fact to being prescribed as physical change or improvement is potentially possible but this is a topic that is out of reach of this thesis because it concentrated on process standards. What is within the realm of what has already been presented is the connection between the process standards, pragmatism and truth.

The publication from Moen and Norman (2006) that summarised the Lewis’s (1929) concepts of truth with its relation to experience is reflected in the structure of ISO systems. The system is based around a cycle that requires plans to be set, completed and the results to be assessed so that further changes can be made based on the previous experiences. The systems are to be used in the practical context of an organization where the organization has the freedom to create a system that fulfils their practical needs (Annex A). This directly relates only one of ideas of truth that was presented by Lewis (1929). What the reviewed ISO systems don’t neglects to consider is that experience shouldn’t be able to change truth and the conceptual interpretation of facts being a danger to seeing the actual truth. How ISO combats this hazard is by only producing the standards and requiring a third party audit so that misinterpretations are limited (BSI 2015). This however doesn’t fully respect the complexity of the Lewis’s concepts. The age of his publication might have something to do with this or it may also have something to do with the rapid developments of markets such as nanotechnologies (Goebelbecker and Albrecht 2014).

The direct relation between ISO systems the science of improvement is demonstrated in Annex 2 where each of the prepositions from Perla et al. (2013) are addressed in respect to the ISO system generic structure. What can be deducted from annex 2 is that in general ISO and their structure has a relation to the prepositions from Perla et al. (2013).This relation comes mostly from the Plan-Do-Check-Act Model used within the Annex SL Structure. There are however many concepts from the prepositions that are not completely referenced by ISO. ISO neglects direct reference to “justification and discovery” (pg. 173) by giving organizations freedom of choice and the systems theory of a common cause must be initiated very strongly by the leadership team which may put too much responsibility on a few peoples shoulders (Perla et al. 2013 ; Annex A). This also neglects the writings of Kaizen that every member of an organization should be involved in improvement processes (Moen and Norman 2006).

6.3.3 Shewart, Deming and Further Developers
What has not been neglected by ISO are the concepts from Shewart and Deming with respect to the PDSA cycle (Moen and Norman 2006) and the questions from Langley et al. (2009). As can be seen in figure 9, which took elements of figure 2 and figure 4, there is only the minor difference between the two cycles where the PDSA has study, a word with an academic connotation, in comparison to check (Oxford 2018). The concepts behind these words are however synonymous.

Figure 9. Comparison of the PDSA and a simplified Plan-Do-Check-Act Model adapted partly from Langley et al. 2009 (pp. 24)and figure 4(on pp. 32 of this thesis)

The first of the additional questions asked in the Improvement Guide from Langley et al. in 2009 about what the organization is trying to accomplish is mirrored in the “objectives” of the systems and the “intended outcomes” displayed in the Plan-Do-Check-Act Model. The second question about how to observe if an improvement has come from a certain change is directly related to Clause 9 “Performance and Evaluation” of the Annex SL Structure (Annex A). The third question about what actions result in an improvement is directly related to clause 6 about “Planning”. This abundance of similarity shows how this predominately American way of thinking is also being used worldwide (Moen and Norman 2006). What was not covered in this work is who influenced who and when. Whether ISO gave an international perspective where the science of improvement was an American application, or ISO use ideas from these American thinkers concepts for its international standards or that the simultaneously developed and learnt from one another, it can still be said that there are parallels between the two where each has different concentrations.

6.3.4 ISOs Position and Standards in General
ISO shares a very specific responsibility that was given to them in the TBT (GATT 1994). This position in combination with the abundance of standards that they produce and that are used worldwide means that ISO has is a special actor in the world of standards (ISO 2010). They produce voluntary standards that, as prescribed in the “European Forwards” are to be translated into national standards (please refer to Annex 1). As their name full suggests they are one of the most important players within the realm of international standards even though they don’t have a direct legal influence (ISO 2010). It can be said that ISO improves international standards, therefore enables more efficient international trade by fulfilling their role as apart of the TBT and by producing many of their own standards that require a third party audit. Due to this they can be considered an effective signalling tool. It must also be said that this role could have been filled by another organization and would have been if one of the predecessors of ISO had succeeded (Latimer 1997). How well they combine positive aspects of private and public actors to improve standards could however be subject to further research.

6.3.5 Conclusions
The science of improvement and ISO have many similarities. The basic ideas are similar with ISO having direct practical uses and therefore different details to the science of improvement. The parallels between the two at least give the impression that ISO is trying to improve their standards and help any organization that use them.
7. Final Comments

One aspect where one needs to be particularly careful is ISOs position as referred to in the Technical Barriers to Trade Agreement. They are a private organization with good intentions but a lot of influence. As can been seen by their requirement for showing an improvement with ISO 50001:2017 Draft that they might not be respecting this agreement fully. This will need to be investigated further as it could have serious implications for ISO.