Chapter 4: A company like many others

# Chapter 4: A company like many others Having understood the global forces and trends that are shaping the technological evolution of the industrial sector, we'll now zoom in into the daily life of an ordinary manufacturing company. Your business will be different in some aspects, but you should still find many commonalities as you read. This is a fictional example to convey the common scenarios that I've seen in companies that haven't yet achieved streamlined and robust processes. If you feel like some of the following issues resonate with your workplace, don't be ashamed, you're in good company: most small and medium manufacturing businesses struggle with processes, and many bigger companies still have yet to come to grips with some of them. Getting data right is not easy, but with the appropriate strategy you can. On the other hand, those of you who have always worked in structured, efficient enterprises, may find some of the issues totally out of the world. It is not so for smaller realities. It's not a matter of laziness and unreliability. These are growing pains of successful, thriving companies that have yet to harden their bones, and this book wants to speak to them too. So if you already have a strong orientation to processes and data governance, hat off to you! I'm still pretty sure you can find some pieces of wisdom ahead to help you sharpen your skills and think more clearly of your next steps. Let's dive in! ## ACME Inc. ACME Inc is a company that produces professional and industrial equipment based on a few common platforms that can then be personalized according to customer needs. The company is healthy and has been growing significantly in the last few years. It has successfully weathered the pandemic and somehow is also managing to get through the raw materials shortage without losing orders. The company has a functional organization structure that loosely follows the classic Porter Value Chain Model, with a Sales, R&D, Procurement, Planning, Production, Logistics, Quality, IT, HR, Finance and Control departments. This organization is now making things more difficult, because each order has its own peculiarities and with more and more coming it's becoming hard to coordinate between all the departments regularly since each function is busy managing itself to reach its own targets and KPIs. :::warning VALUE CHIAN MODEL PICTURE ::: :::info NOTE ABOUT FUNCTIONAL STRUCTURE: Those of you who are acquainted with lean principles and those who have more academical business background already know the limits of typical functional organizations. Please stick around, as seeing those issues through the lens of Industry5.0 can provide you with further insights. ::: The increase in volumes was stretching people beyond their ability to keep up but simply adding personnel has not helped much. New people are not as experienced with the nuances of the products and of the internal procedures and there's little time to stop and train everybody properly, since a lot of energy is going into fixing daily issues coming from activities that were once simple routine (we'll see some of them in a moment). Plus the recent addition of people and equipment in the shopfloor to increase capacity has only added to the complexity of keeping things under control. For all these reasons they want to adopt a new approach to streamline activities in order to satisfy the increasing demand without having to grow the organization too quickly. Their hope is to reduce non-value-adding activities (such as troubleshooting and fire fighting) and free up resources to focus on improving lead times and quality. ## Fulfilling orders The high-level lifecycle of a standard customer order follows this process: - A customer inquiry is passed on to a sales technician who drafts product specs and generates a price quote through a configuration software. - If the quote is confirmed the technical office, part of the R&D department, takes charge to finalize the bill of materials and technical documents, and create item codes when necessary. - Once the bill of materials is defined, the purchasing office checks the inventory and issues purchase orders, supported by the MRP software[^1]. - The planning office then tries to establish a plausible fulfillment date for the order checking the following: - estimated production lead times based on data from the configuration software - purchases lead times declared by suppliers - production capacity - Once they have all necessary info they provide feedback to the sales department on the planned shipments. - The sales department confirms the delivery date to the customer and the order is finalized. - The production order is then added to the factory portfolio and purchase orders are issued by the procurement office. - Procurement, production, planning and sales coordinate at various level at least on a weekly basis to accomodate new orders coming in, handle issues and adjust schedules accordingly. As the delivery date of an order approaches several departments in production and logistics adapt their plan to fulfill the demand for sub components in time for the final assembly. - Once ready, the product is shipped to the customer and invoiced according to sales agreements. [^1]:Check the IT glossary in appendix for a definition of MRP On paper the whole process is not super simple, but it is still pretty straight forward. In reality, anybody who has had any real-world experience in manufacturing knows that it's very easy for things to go wrong somewhere in the middle. ## Common issues To make sure that orders are planned and executed with a good service level, many moving parts must act in concert and with accuracy. In discussing these parts, we'll consider the three main high-level dimensions of performance: time, quality and cost. If something goes wrong in any of these moving parts, projects will be affected in one or more dimensions, that is, the order may go late, have quality issues, incur in increased costs or any combination of these. :::warning [ DRAWING OF THE COST-TIME-QUALITY TRIANGLE ] ::: Of course nobody wants that, and companies try to avoid undesired events as much as possible. Unfortunately, besides the most evident issues, there are many subtle ways in which these performances can be damaged. I say they are subtle because they tend to be less tangible and immediate than, for example, an unwanted machine stop or a supplier delay. Most of these issues, as we'll see, are not related to the production process in itself, but in the supporting processes around it, especially at the seams of different functions. For this reason these subtle issues are often overlooked and accepted as a matter of fact, but be warned: even if these risks are not evident, and any single event may not have huge consequences, it does not mean that their impact on global performance is negligible. In fact, they may even have a _greater_ impact than those that are immediately visible due to their **repeating and cumulative effect**. You can think of these issues like poorly inflated wheels on a shipment truck. You may not notice the impact at first, but you'll discover a much higher fuel consumption and thus cost later, which may eat in your margins in a very unhappy way. These seemingly small issues happen mostly in what we earlier called _low-level coordination activities_ (LLCAs) and addressing them will likely yield a return that is far higher than what you would expect. We'll now go into more details on a few of these potential pitfalls, drawing some more generalized conclusions at the end of the chapter. ### Making a quote The first step after receiving a sales inquiry is making a quote that will inform the customer of the pricing and fulfillment date for the project. This will not be an issue for "make-to-stock" manufacturers: producing larger batches of the same product makes figuring out cost drivers and marginality easier; plus lead times tend to be close to zero as shipments are often made from already available inventory. This is not to say that make to stock is easy to handle: many other complex decisions must be made regarding mix, volumes etc, but this is outside of our scope. In other context, foreseeing costs and times is not so easy. There can be tens or thousands or even millions of possible configurations starting from the same product platform. A given variant may see one or two pieces produced in a year (if any at all) and at different times and different conditions for the company, making it difficult to use simple statistics to calculate standard margins. Plus in the B2B space there's usually significant negotiation involved over the price, which makes predicting profits even more difficult, and can even bring the deal below break-even if costs are not well accounted for. In these contexts, the quoting step becomes crucial because it decides two important things: the positioning on the market and your marginality. Price too high and you may lose deals to competitors, price too low and you will lose money; set a deadline too close and you may miss it and incur in penalties and lose trust in the market, set it too far and, once again, you may see the customer choose somebody else. One should not fear the consequences of a quoting mistake or two in the middle of hundreds and thousands of offers, but if they start to happen regularly they may create serious harm, just because orders were not timed or priced correctly. And I know you don't want that. So how do you get quotes right? This is not a book about positioning and pricing strategies, but whatever approach you pick, you must have a sufficiently accurate knowledge of your production costs and execution times in advance. In general, I expect that all this is no news to you. But I point it out because in order to estimate times and costs, a LOT of information is required. Let's make the list. You need to know: - what items went into the products and how many/how much of them, i.e. the Bill of Materials (BoM), or Parts List (PL)[^2]; - how much the specific items cost you _when you purchased them_ ; - how much labor and machine time you spent producing the machine; - how much labor and machine time you have available [^2]: Different companies may give slightly different meanings to these two terms. The definition we'll use is that the BoM is the master component tree, defined as a general reference for a specific product, while the PL is the sum of elements consumed in a specific work order, which may differ from the master BoM. Inventory consumption will always be driven by the contents of PLs. No. 1 (BoMs and Parts Lists) seems easy, but you should not take it for granted. Multiple input items or materials can serve similar purposes but have different prices, lead times and quality levels. Which one are going to be used? Besides this, in less structured companies it's not rare to have incomplete or mistaken parts lists. They are often born reflecting the R&D or sales perspective and may not include every little detail necessary for the production process to run smoothly. No. 2 (material cost) can become a severe issue depending on the type of input you rely on. Semiconductors and raw materials like metals are a great example of how price can vary significantly in a short window of time, so much to make historical values and averages unreliable. Energy and oil price changes can also drive input costs up or down due to the variation they generate in manufacturing and transport costs. And if you buy from abroad, currency exchange rates also will play an important role. No. 3 (processing time) strongly depends on the complexity of the transformation that ACME carries out on input material, but it is never negligible. For heavily automated processes execution time is pretty consistent. On the other hand, for more labor-intensive activities (e.g. manual assembly, testing, setup, etc.) you can manually time execution, but if you work in a high-mix, low-volume environment, you can't really study all the possible products in advance, especially in cases when you provide manufacturing services and don't make your own products. No. 4 (production capacity) is also not as easy as it may seem. You always start with the total theoretical production time available (whether from labor or machine), but that is of little use. You cannot know how many pieces you can make in a shift, day, week or month until: - you know how much it takes to make one - you know when and which people are available - you know how much time machines are actually available to produce (excluding maintenance and setup times).[^3] [^3]: Some of you may be familiar with Overall Equipment Effectiveness (OEE). It is a popular measure to gauge the productivity of equipment. What we talk about here is the *Availability* component of the OEE. In high mix low volume contexts all these vary. Processing times are not always known in advance (see point 3). Equipment setup times are hardly predictable based on the order portfolio alone, unless specific action is taken to standardize operations in this sense[^4]. It takes effort to measure machine stops, especially unwanted ones, and even more to feed this info back into planning. As for people availability, besides accounting for sick days and holidays of personnel on payroll, ACME must pay the price of flexibility earned through the use of contingent workforce, covering 5 to 10 percent of labor costs depending on the time of the year; hourly workers may reduce overall fixed costs, but they also tend to have lower performance and they make it more difficult to establish production capacity. [^4]: Lean practitioners know all too well the importance of reducing setup times through standardization of setup procedures and the use of tools such as jigs and fixtures to increase efficiency. Specific techniques have been developed in time about this, under the umbrella name of "Single Minute Exchange of Die", or SMED. (Note that the goal is not actually for setups to last 1 minute, but less then ten, that is a single digit number of minutes). Does it seem complicated enough? Well, it is. And ACME hasn't even received the order yet! But there's good news. Complexity has its rules and can be tamed with the right strategies. But I don't want to skip ahead, so let's stick for the moment to understanding other possible risks and ambiguities in the management of ACME's processes. ### Production Planning and Control The planning department used to coordinate all other functions to make sure the deliveries happened correctly and on time and in time people in ACME have come to expect planners to be on top of everything every time. With sales and the organization growing significantly in the last years, it has become impossible to manage each order one by one like they used to, but expectations haven't changed and today every issue seems to be their fault. The problem however is not the people, but a non-scalable, centralized process to keep track of things. The planning department is in fact the sole responsible for the management of work orders: the status of each one must be manually changed by them on the ERP; each change in the parts list must be validated and recorded by them. The delivery date of each quote must be vetted and confirmed by them after coordinating with the purchasing office. Because they are responsible for work orders and the only department who relates regularly with purchases, sales, production and logistics, everybody relies on them for updates, asking them questions frequently each day, only adding to the already huge pressure of having to carry out their actual planning work. It's not a surprise that things are starting to fall through the cracks. The _process_ simply is not up to the task. In particular, one of the things that is creating the most issues is the sequencing of orders. Production is divided into several departments and material flows do not follow a straight line. Some departments make sub-assemblies for multiple product lines, others work only on a single one, and most items cannot be produced to stock, but are specific to one order or the other. Thus a lot of coordination is needed to make sure material is available when and where needed. The _when_ is the trickiest part. The ERP does not provide any insight on the feasibility of orders within a specific timeframe. The only constraint it considers is the availability of material, but once that is covered, the system supposes that you can produce anything immediately, regardless of how much capacity you have. This assumption is referred to as _Infinite Capacity_, which is obviously not a realistic thing. Usually people figure out rules of thumb to understand when too much is too much and define delivery according to that, but this works only when you can coordinate closely with all stakeholders, on all orders, regularly. As soon as you receive orders faster than you can discuss them, then this kind of process fails and orders will start to be placed where your capacity has already been saturated, leading to delays. When they reach this point, some companies try to approach planning using _Finite Capacity_ algorithms, which take into account many variables to schedule work orders and optimize routing on the different work centers in order to maximize utilization or lead time or whatever combination of goals they're setup with. If you think this is the final solution to capacity planning, think again. There are two main reasons why _Advanced Planning and Scheduling (APS)_ systems, as they're called, may not help, and neither because of their own fault. The first reason is that there's nothing an APS can do for you unless you can model your production correctly. They may be great at optimizing, but if they do based on the wrong assumptions, whatever result you get will not be a correct one*. In ACME, the rules of production flows are reasonably and logically structured from a high-level point of view, but if you dive into the details you'll find many situations where people make choices based on their practical experience and will have a hard time pinning down an explicit rule to cover all cases (e.g. which piece of equipment to use to process a specific part). Plus, as we've seen earlier, recorded lead times may not always be correct and the system will have no way of knowing it. And by the way, you have to explicitly tell the system the lead time for each phase of production for each product: good luck in being accurate with that when many orders are for products being produced for the first time. You may have heard a variant of the saying "garbage in, garbage out" from some IT guy. This is one example of that. The second reason is due to the volatility of production plans on one side, and the rigidity of optimized schedules on the other. Imagine playing Tetris and sticking every piece of the puzzle in the plan with as little empty space as possible. This is what optimization algorithms do. Now you know that it may be enough for a single piece to change shape or move for the whole puzzle to fall apart. That's what happens with optimized schedules. The system of course allows you to re-optimize and re-schedule when things change. In ACME however, changes are the norm, not the exception, and the idea of changing plans daily or even more frequently would make people in various departments quite unhappy. In summary, in companies where products are standardized, are produced in medium to large batches, and the factory has a pretty simple material flow, APS may work fine. But most job shops and make-to-order environments will have a hard time implementing them with success. So what should planning do? Currently they keep dealing manually, or rather verbally, with capacity considerations, through a bi-weekly meeting with department forepeople, where they discuss the status of each order for the coming six weeks. These meetings are usually a high-stress situation in which planning reluctantly tries to push deadlines (which in turn are being pushed on them by sales), with forepeople pushing back to ensure feasibility. Sometimes it feels like a tough old-school horse market negotiation. Not really a motivating job, but there's too much work to even stop and think. Any minute lost may mean an order not updated. ### Inventory Control and Planning Do you know any sizeable manufacturer that does not physically count inventory at least once a year? There's a reason why counting is necessary: even with the most advanced software systems and processes, you will always have some level of discrepancy between recorded and actual stock levels. That's because moving things around cannot always be tracked automatically with accuracy and people will always make mistakes to a certain degree. In some cases this can be a serious issue because the stockout of a tiny part costing less than a burger may lead to significant delays of entire shipments. There are many reasons why this can happen in ACME: - defective items due to poor incoming quality controls; - untracked consumption due to e.g. urgent deliveries of spare parts or scraps in the process; - low-value, small-dimension and high-volume items such as bolts and nuts and screws and others which can be tricky because people often lose a bunch of them in each production and there's no way of tracking how many; - items that are consumed in length or weight rather than pieces, which are harder to measure and may be consumed at a higher rate than expected; - lastly, one of the most frustrating culprits are inaccurate bills of materials and parts lists. Besides pure mistakes of missing and excess components, issues can arise when the BoM levels are not structured according to the actual production process, that is, sub-assemblies which are a reality in production but are not treated as such in the ERP or viceversa. This makes so that consumption of some items is delayed or anticipated compared to reality which can create confusion especially when sub-assemblies remain in stock for a while. Another shade of this is the management of "complementary" items, such as shipping material and accessories: sales and operations are still discussing whether to include these as additional order lines or within the product BoM, and more than once a complete machine could not be shipped on time because of a lack of such items. But inventory issues don't stop at the discrepancies between recorded and reality. With the high pressure of the many orders coming, delays in supplier shipments sometimes are not always spotted enough in advance, with warehouse personnel finding shelves empty when they go to get the picking list done. This leads to production having to switch order or, worse, to stop altogether. Can you see how tricky these issues are? Solving them requires carefully coordinating multiple departments, but when you're always chasing urgencies it gets really hard to concentrate on these things: who has time for trivial problems such as a 5 euros component or a 10% difference in expected lead times when you have Super Important Client X calling about a late order? The paradox here is that urgencies usually come up _exactly because_ of issues like these. **Most urgencies are the cost of not dealing with small details that are difficult to deal with**, like: - setting up policies for procuring low-cost components and shipment material, - finding out ways to empower front-line workers with recording important data such as changes in the parts used, - involve production personnel in the review of BoMs - think of a light-weight system to coordinate capacity evaluation and allocation - etc. All these may seem like something that can be postponed to tomorrow, until you find out it was needed two months ago. These are some examples of the Low Level Coordination Activities (LLCAs) which can drastically increase time-to-value and the overall energy of the company. Most companies do know it would be highly beneficial to streamline these activities, so what do they do about it? ## Half solutions which are half the problem We've listed a series of complexities that commonly arise when manufacturers grow, especially if fast. In these contexts managers often resort to two types of solutions to face complexities, both of which have their pros and cons. ### Reporting and coordination tasks The first strategy is adding data gathering and reporting activities to people's tasks. For example, when it comes to the use of alternative items in production, operators in ACME must inform planning about the change for a specific work order through a paper form. These forms are collected daily and must then be recorded in the ERP. In other cases, people are assigned with periodically updating and sharing reports about various kinds of information, such as production status, inventory, efficiency and so on. Managers need these pieces of information, but can't collect or handle them personally because it takes time and attention to each single case so they delegate these tasks to their collaborators. This is a very common scenario and a good sign that value is being placed on real data. On the other hand, exactly because it takes time to carry out these tasks, people gradually get saturated with tracking, control and reporting activities and coordination meetings: this shifts their focus away from value creation and daily problem solving, which alienates them from their role and creates frustration. That's why, when reporting tasks become a big enough burden, companies resort to hiring people specifically for reporting purposes. A similar issue happens with coordination activities, those that require communicating with multiple parties to get things done: the Project Manager is probably the archetype of roles dedicated to coordination activities. ACME has hired people as _expeditors_, a role whose main job is to support communication and coordination, avoid details falling through the cracks, and to push people and suppliers when things get hot to make sure specific orders are delivered on time. It's similar to the project manager, except they do not take decisions, and are dedicated only to gathering and sharing information. For operators, expeditors are those that sometimes come and say they need to switch production to another order because there are issues with x and z. Outside of ACME, lean companies have peculiar figures which are dedicated to these kind of activities. One of such figures is the _Value Stream Leader_, whose role is to coordinate and monitor at 360 degrees the activities and requirements (from inventory to production to shipping) dedicated to a family of product or similar aggregation. You can see that, while reporting and coordination is crucial in any business, there's a huge cost to them, whether in the form of hiring dedicated resources, or in the time people would otherwise spend in their core tasks. Just start counting the hours you have been in long meetings or the times you felt you didn't have enough time for your own work because of the dozens emails you had to reply to; you get the idea. ### Rules and policies Another strategy to handle complexity is to reduce it by defining in advance what decisions need to be made in different situations through specific rules and policies. This is also a common strategy, but not as spontaneous as the first one, because just defining these rules is already quite a feat for companies relying on implicit knowledge, like most small and medium companies do: these rules must be formalized in documents and presentations, then they must be communicated, repeated, repeated again, verified and sometimes enforced with decision. It's not always easy. Businesses caring about quality standards (like ISO9001) will also need to keep track of changes in the procedures, who made the change, when and why and so on. Besides the burden of implementing the policies and managing the documentation, by definition these rules limit the freedom of operation and, while this does reduce complexity and increases control and reliability, it may make the business more rigid and slower to adapt and respond, sometimes even incapable of doing so in certain areas, since the options to choose from are limited by definition and have been defined in the past, maybe at a time when current conditions where not even thinkable. This is not to say rules and policies are bad. Quite the contrary. No sizeable company can work properly without a certain amount of them and the right ones can be a boon for productivity and value creation. But too much of a good thing is also bad: the stereotype of the big, slow but reliable company that turns neither right nor left may have worked in the past, but today's markets demand speed and agility, joined with quality service. ## So what? The above solutions are not the _only_ things companies do, but they are the most common ones, and regardless of the strategy a company chooses to grow and structure itself, reporting, coordination, rules and policies will always be part of the foundations of how businesses work and thrive. So it's not a matter of _if_, but _how._ Being so essential, the downsides of these elements could be considered a necessary evil until a few years ago. But not so now. The performance trade-off between scale and reliability on one side and speed and agility on the other can shoot well beyond the established ones. And that is thanks to the enabling Industry 5.0 technologies and practices. :::warning Show performance curve shift ::: Now you should have a clear picture of how _your_ business can benefit from a digital transformation and how it can help you overtake competitors, or help them overtake you. Industry 5.0 can completely change the rules of the game by breaking out of the path of the traditional business lifecycle that sees a company becoming rigid and slow as it grows. On the contrary, business growth can go hand in hand with the ability to serve customers better and adapt to their needs. If that's something you'd like to achieve, and you're asking where to go next? And where to start from? These are the questions we'll answer in Part 3. Before that, we need to create a better understanding of the elements that constitute an efficient and thriving digital business.