Ibukun Onifade

Cost avoidance measures are action taken to avoid having to incur costs in the future. The benefits from cost avoidance can not be shown on financial statements and cannot be reflected in the budget. They are more difficult to estimate. EXAMPLE In a manufacturing setup, management may decide to reduce spending on preventive maintenance activities because the current frequency does not appear justifiable. Some parts may have been scheduled to be changed 6 times a year, perhaps from manufacturers’ recommendations or as outcomes from previous improvement projects. This high frequency maintenance activities might be resulting in huge repairs an maintenance expenditures for the firm. Reducing the frequency will be a cost reduction measure. This will be refelected in the operation expenses portion of the financial statement. On the other hand, after some time, it might be discovered in the same company the unplanned downtime has increased in the factory and requires urgent and drastic action. Preventive maintenance activities might need to be reviewed again. This time around, maintenance expenditure will be increased to avoid the higher cost that will be incurred because of the high unplanned breakdowns. The financial benefit due to decrease in unplanned downtime will be cost avoidance. Further examples of cost reduction Ø Reduction of overtime hours Ø Reduction of manpower Ø Negotiation of lower rental costs for equipment Further examples of cost avoidance Ø Reduction of proposed price increase of a raw material Ø Negotiating long term contracts with price protection provisions Ø Negotiating additional services from service suppliers at no extra cost

FEATURE CREEP As a project progresses, a lot of additions may be suggested to the project ostensibly to make the outcome better. These additions sometimes become so much that the product becomes more complex than planned and eventually is not acceptable to the consumer, this phenomenon is referred to as feature creep. Feature creep is a quite common phenomenon most especially in software development projects. EFFECTS OF FEATURE CREEP ON THE CUSTOMER – A product with feature creep is usually very cumbersome to use because of all the unnecessary features. This will mostly lead to a bad user experience. ON THE DEVELOPERS – Unplanned additions will lead to inability to meet up with project timeline in most cases. It is also very difficult to stay within budget once there is feature creep in a project. HOW TO PREVENT FEATURE CREEP 1. PROPER PLANNING – Product creep is often the result of poor planning. Before a project is launched, it is very important to make extensive research on the needs of the end user. There should be no assumptions at all, all outcomes of the initial research should also go through a validation process. 2. SIMPLICITY – It is always tempting to add some new features to a product with the belief that it will make the user more delighted. This is not always the case, most user prefer simplicity and ease of use to sophistication. Caution should always be employed before addition of new features. 3. CLEAR COMMUNICATION – Before the beginning of any project, the entire team must be clearly communicated on the plan and purpose of the product. Inclusive deliberations should be done to ensure that every member is on the same page as regards the scope and details of the project. 4. TRAINING – Proper training should be given to team members on feature creep. They should know how to identify and react to suggestions to add features to a product. Members must be trained to stick to the timeline of the project, this will keep a constraint on addition of new features during the project.

Just-in-time, as the name implies, is an approach to operations that seeks to ensure that only what is needed by the customer is produced at any point in time. A JIT system seeks to minimize inventory levels of all materials; raw materials, work-in-progress, and finished goods by continuously determining the downstream requirement of each material and supplying the exact quantity needed. JIT was first implemented at Toyota by Taiichi Ohno, it has since become a globally accepted approach to optimizing operations. To successfully implement JIT, there are some basic requirements without which there can be no effective deployment of the system. These are listed and discussed below; Reliable Equipment and Machines; Frequent and long breakdown on machines will lead to stockout at various points in the supply chain since minimum inventory is being maintained. All machineries have to be in top condition, maintenance must also be highly prioritized for JIT implementation to be successful. Well designed work cells; poor layout, unclear flow, and a host of other issues can all be cleared up by the implementation of 5S within your production. Implementing this will make process flow smooth, JIT can never be implemented effectively if there is no smooth flow. Quality Improvements; High level of rejects because of product or service defects, and the consequent rework (where applicable) make operations very inefficient. Quality must be built into the process for JIT system to work. Standardized Operations; Defining standard ways of working for all operations will help to ensure that your processes are reliable and predictable. There will be too much variation in quantity and quality of output if standards are not well defined and rigorously implemented. Pull Production; Just in time does not push raw materials in at the front end to create inventory (push production), it seeks to pull production through the process according to customer demand. It achieves this by setting up “supermarkets” between different processes from which products are taken or by the use of Kanbans which are signals (flags) to tell the previous process what needs to be made. Single piece Flow; the ideal situation is one in which you will produce a single product as ordered by the customer. This, although impossible, should be the ultimate goal of operations. To achieve this, work has to be done on reducing batch sizes significantly through the use of Single Minute Exchange of Die (SMED) which seeks to significantly reduce the time taken for any setup. It will also often require the use of smaller dedicated machines and processes rather than all singing all dancing mega machines. Flow at the beat of the customer; The time interval at which additional pieces of products must be made to meet customer demand is referred to as takt time. You need to ensure that your cells and processes are organized, balanced and planned to match the pull of the customer. This is achieved through Heijunka and Yamazumi charts.

Overall equipment effectiveness is the most acceptable measure for the assessing the performance of any operation. There are 3 major components of OEE; 1. Availability – Downtimes happen in the course of operations mostly when least expected. Downtimes could be as a result of equipment failure, material shortage, shift changeover, product changeover, etc. These downtimes result in a direct loss of output for their entire duration. Availability is calculated by subtracting the downtime from the total available time, then expressing it as a percentage of the total available time. Availability = (Planned operation time-Downtime)/(Planned operation time) 2. Performance – This represents how efficiently the machine is utilized during the uptime. This is normally derived by calculating the standard output of the operation based on standards set by the process designers (e.g machine manufacturers give standard speed at which a machine should run and how many parts it should produce per unit time). The actual output can then be expressed as a percentage of the standard output. Performance losses happen mainly because of 2 reasons; a. short stops in the process that are not tangible enough to be reported as downtime like jamming of parts, obstructions, etc. b. Running lower than standard speed, this would mean that output per unit time is no longer up to the set standard. Performance= (Actual output)/(Standard output per unit time x uptime) 3. Yield – Producing defects affects the effectiveness of an operation adversely. Defects happen due to machine malfunction, human errors, bad raw materials, etc. Defects can also be categorized based on time of occurrence, some defects happen at the start up of the process, while some happen during the course of the operation. Yield= (Total output-Defects)/(Total output) OEE is a obtained by multiplying the three major components; OEE=Availability*Performance*Yield It is practically impossible to get 100% OEE because there is no way we can completely eliminate the losses identified above. For example, we may reduce changeover time as much as possible, but we cannot completely eliminate it. Downtime also can be greatly reduced by proper preventive maintenance and close machine monitoring, but this also cannot give us zero downtime.

CONWIP (Constant Work In Progress) is a method of implementing pull in an operating system. CONWIP is a combination of pull and push systems. The entrance of an item into the system is based on pull, but its movement through the processes towards the end is based on push. The inventory in a CONWIP system is controlled only at the end of the line, there is always a fixed amount of finished goods inventory that can be kept at anytime. This system is maintained by the use of cards, a completed product frees up a card which will be attached to the new entrant into the system. In CONWIP, nothing is allowed to enter the system without the card. The number of cards in they system is equal to the maximum work in progress inventory. CONWIP differs from Kanban in 2 major ways; there is a pull system for every stage of the process in a Kanban system, every stage requires a Kanban signal from the succeeding stage before producing the next lot. CONWIP implements pull for the entire system at the beginning only, the only signal is to pull material into the system. Also, the cards used in Kanban carry specific item tags, a Kanban card can pull only one particular item in the process. This is not the case in CONWIP, the cards are generic, any card can pull any item into the process. The items to be picked are determined by scheduling based on the existing backlog. ADVANTAGES OF CONWIP OVER KANBAN 1. Kanban will be very effective only in a high quantity, low variety environment where production is made to stock. Consider a setup where items are made to order and the components of each order are not similar, Kanban cards will need to be made for every new order which will be difficult. This is not required in CONWIP since the cards are generic. 2. It is easier to control the sequence of production in a CONWIP system than in a Kanban system. Since a backlog system is used in CONWIP, production planning personnel can always rearrange the sequence in case of any change in product requirements since only one pull system is place at the beginning of the process. Rearranging sequence in a Kanban system can be done only on the production floor DISADVANTAGES OF CONWIP 1. A CONWIP system is more sensitive to bullwhip effect. Since the sequencing is manually done as against the automatic sequencing in a Kanban system, misjudging the requirements is more likely. The system can end up with many items at the end of the process that are not currently needed. 2. CONWIP involves sorting of backlogs and matching the backlog to CONWIP cards. This involves more manual work and could also result in errors.

Planning poker is a consensus-based, gamified technique used often for estimating the effort required to accomplish a certain goal or relative size of development goals in software development. In planning poker, members of the group make estimates by playing numbered cards face-down to the table, instead of speaking them aloud. When all have played their cards, the cards are revealed simultaneously, and the estimates are then discussed. By revealing the figures at the same time, the group can avoid the cognitive bias of anchoring, where the first number spoken aloud influences the estimate other participants will make. The design of this process was to help software organizations more accurately estimate development timeframes, build consensus among the members of the cross-functional team, and more strategically plan the team’s work. The steps involved in planning poker are explained below; Step 1: Hand out the cards Participants are all given an identical deck of cards, each with a different number. The most common sequence used is the Fibonacci sequence, each number in the sequence is the sum of the 2 preceding number in the sequence.The decks are limited and well staggered, because the goal is for all participants to reach a consensus number for each story. This is to increase the efficiency of the process. Step 2: Read the story The product owner will read each story out loud to everyone in the group. Step 3: Discuss Since everyone is now aware of the story, it is time to discuss about it. This also a time to ask questions so that confusion will be eliminated. Participants will describe the ideas they have about how the work will flow, how many people they estimated will be required to get the job done, which skill sets will be required, and what if any obstacles they envision slowing progress. Step 4: Estimate and share When all discussions and questions have been concluded, each person will secretly choose a card from the deck to represent their estimate of story points. They will then reveal the numbers on their cards simultaneously. Step 5: Work toward consensus If all participants’ reveal the same card, then that number becomes the consensus. The group can move on to the next story. If the numbers are different, there will be another discussion where people at the extremes will explain their choices, more insights will be shared among the team members. Another round is then conducted where participants will make new choices. This process will be reiterated until consensus is built among the participants.

Filter bubble as a phrase was first used by Eli Pariser in his 2011 New York Times bestselling book, The Filter Bubble: What the internet is hiding from you. It is a term used to describe the personalization of individual online experience by the use of computer algorithms. These algorithms customize what an online user sees based on his online history. For example, if a user has been reading articles that support a particular viewpoint, this algorithm will filter his feed such that he will be seeing mostly articles that support the viewpoint he holds. The user find himself in a kind of world created for him by this algorithm, a world where almost everyone leans towards his viewpoint, a world where the alternative viewpoint is silenced, this world is the filter bubble. Online marketers are already taking advantage of the filter bubble through targeted advertisement. If you check about an item online, you will soon be bombarded with several adverts about that item. This has made marketing more effective as advertisements are now directed towards interested people who are more likely to make a purchase that any random person coming across the advert. Businesses must first ensure that they a part of the filter bubble of a sizeable part of the market. They must maintain a regular and aggressive online presence. Web analytics can also be employed to get reports on people that have us as part of their filter bubble, and more importantly who are those that do not have us. We can then strategize on how we can break into their bubbles.

Paralysis by analysis – data driven decision making is becoming increasingly popular in organizations. But to what limit should we depend on data analysis to make decisions? Scenarios abound where the main objective of data analysis end up being defeated because of nothing else but excessive analysis. What should ideally bring clarity becomes the cause of confusion. Typical situations of paralysis by analysis are; a. Dialogue of the deaf – To get approval for certain actions, much data analysis is done to justify such decisions higher management. Unfortunately in many cases, the higher hierarchy already has a strong opinion about the subject matter unknown to the subordinates. People continue to prepare more reports which are more likely to be ignored. In some cases, managers may even find it easier to request more and more analysis instead of directly rejecting proposals. b. The vicious cycle – In situations where there is wide participation and a diffuse power structure, where many people at the same level have equal say in decision making, paralysis by analysis often manifests. Proponents and opponents of a decision conduct many analyses to support their position, and it becomes very difficult to build any consensus because the parties involved don’t even trust each other’s data. c. The decision vacuum – This normally happens when the decision maker and the data analyst are not in sync. Many organizations create departments solely for the purpose of data analysis, but people in such departments don’t really influence any decisions being made in the organization because there is a disconnect between the decision makers’ priority and that of the data analysis team. Extinct by instinct – This is the opposite end of the spectrum, where decisions are taken without sufficient analysis. Decision makers form a kind of solidarity, refusing to explore negative possibilities in the process of decision making. Hasty decisions are often poorly thought out, and are most likely to result in a dead end when eventually implemented. Examples are discussed below; a. The dominant leader – Subordinates are afraid to question the thoughts of a dominant leader. CEOs for instance wield a lot of power in organizations, and people under them may find it difficult to challenge whatever they say especially when such leader has a reputation of not welcoming opposite opinions. Overtime, members of the team become “yes men” and just rubber stamp whatever the leader decides. b. Parallel power – This happens when some individuals or group even though not high in the hierarchy, are powerful enough to decide and implement whatever they want without recourse to others’ opinion. Such power come sometimes because of privileged access to the highest authority or because of some specialized expertise. For instance, Senior managers may accept ideas from lower-level managers that are not necessarily in the company’s interests, either because they do not know enough to ask the right questions, or because opposition would not seem legitimate.

Instruction creep is a common problem especially in large organizations. Standard operating procedures are usually modified frequently once any lapse is perceived by people responsible for maintaining order in the process. This continuous addition often leads to a complex guideline that most workers find tedious to read through or implement completely. In fact, most workers do not read their usually voluminous SOPs not to talk of following the contents. Many guidelines have been rendered ineffective because of this anomaly. The following steps can be taken to avoid instruction creep; 1. SOPs should only be created only when there is a real problem to be solved, not just by someone’s perception. A review committee can be established to review any new SOP and give approval only when the SOP is targeted at a real challenge. This will eliminate unnecessary accumulation of SOPs . 2. Precise language should be used when creating SOPs to eliminate confusion during implementation. Volume of words should be kept minimum, unnecessary repetitions should be avoided when creating guidelines. Guidelines can be translated using visual aids if possible. It is said that a picture is worth a thousand words, this has been proven to be the most effective way of communication whenever a sizeable number of people are involved. 3. Policy makers need to involve the workers that will implement a policy when drafting it. Their inputs should be given consideration when drafting SOPs. Doing this will make them have a perfect understanding of what is expected of them, it will also improve their level of compliance and cooperation when the SOP is implemented.

Reporting bias happens when a researcher’s report is influenced by the nature and significance of the results. This sometimes defeat the objective of research. 1. Multiple publication bias: This happens when the results of a study are published multiple times most especially when they are favourable. Data is also sometimes duplicated within the same study. Studies should always be subjected to rigorous reviews to avoid this kind of bias. 2. Location bias: This happens when pulling results from different databases. Studies stored on some databases are already influenced by some other bias, extracting secondary data from such database will introduce location bias into our study. Studies should span through all relevant databases to remove this bias. 3. Citation bias: We sometimes look into the reference section of a research to get more research results about a particular subject. Studies have shown that researchers sometimes select references that confirm their biased judgement of the subject. Hence, analysing citations from the reference list of one study may introduce citation bias. Selection of research materials from a single reference list should be discouraged address this. 4. Language bias: We find it more comfortable to consider results from studies that are reported in a language that we are familiar with. The implication of this is that our assessment will not cover studies in other languages which leads to bias against data in studies conducted in other languages. Translation software can be employed to solve this problem. 5. Outcome reporting bias: Results are sometimes filtered based on how favourable they are to our expectations. The conclusions from such studies are therefore misleading because it was done only to justify subjective conclusions. To avoid this kind of bias, sampling procedures and sample selection should be well reviewed, live data reporting can also be adopted.