Although the analytical skills required to make use of data play a role in many professions, few are built upon them like computer and electrical engineering. It’s difficult to complete even the most basic of tasks in this area without these skills, so everybody with real ambition works on developing them from an early stage. A thorough knowledge of the basics will suffice at an early stage, but the more they can be improved, the easier it is to advance.
This article explores some of the contexts in which they’re needed and looks at why they matter.
In many areas of business, market research is carried out by a separate department that has very little interaction with product designers, but this is rare when it comes to computers and electronics. This is because asking the right questions and making sense of marketing data requires a degree of technical understanding of the products or product types under investigation.
This means that even before getting to the design stage, a business working in this area needs input from engineers who understand marketing data. This is particularly important in B2B contexts where the customers are themselves technicians or engineers, and in contexts where you’re bidding for public sector projects with pitches that will be scrutinized by expert teams.
Just like marketing, budgeting, which is often handled separately in other industries, generally requires input from engineers in this sector. It takes an in-depth understanding of engineering issues to make realistic assessments, but in order for budgeting to be done successfully, data from similar projects needs to be analyzed.
In other words, it is impractical to keep finance and engineering departments wholly separate. Each of the issues discussed in this article will need to be considered, and related data assessed, when drawing up budgets. Only experienced engineers can properly understand the costs involved and resist pressure to cut corners where doing so risks compromising vital aspects of design.
Establishing development and production lead times
It’s notoriously difficult to make reliable estimates of development lead times in this sector, but it’s quite impossible without drawing on data about similar work done in the past, which is the only truly useful indicator of how long each stage of the process might take.
Engineers without a background in data or statistics often make the mistake of thinking that they can estimate development time based on their personal experience on other projects. However, this is an approach prone to bias and routinely fails to adequately factor in how unforeseen hurdles are dealt with. When it comes to production, an intimate knowledge of the product is required to anticipate the challenges of assembly, and to recognize that small-scale operations cannot be smoothly scaled up because of the importance of assembly line workers being able to focus effectively on detailed work.
Calculating the failure rate in the manufacturing process
No matter how accurately you have managed to measure production times, you’re also going to need to account for a failure rate. With electronic goods, this is much higher than in other areas of manufacturing, and there’s usually a trade-off between how fast you can produce goods and how many of them will have faults. These could be major, leaving you with products that have to be disposed of – or they could minor enough that you can still put the products up for sale, but there’s a risk of reputational damage because they make unwanted noises or don’t run as smoothly as intended.
For example, the Pentium FDIV bug affected such a specific function of early Intel Pentium processors that the company went ahead with release because a relatively small number of customers would be dissatisfied and they didn’t want to miss a great market opportunity. Without being able to predict how many users would be affected, they wouldn’t have been able to make that decision.
Therefore, an ability to determine the frequency with which faults occur and the number of users they are liable to affect, and how severely, is essential to making good decisions about product release, and only a data-literate engineer can do this.
Understanding end use
Product design isn’t just about creating something that can be understood and used effectively by engineers – it always has to take the end user into account. What’s obvious to someone with a technical background may be far from obvious to the average member of the public. In this situation, data from studies of public engagement with similar products, or from surveys exploring consumer preferences, helps to fill the gap.
It makes it easier for engineers to ensure ease of use and also helps to create an understanding of end-user workflow. This can highlight unexpected difficulties (e.g., around downtime as a product boots up or shuts down), but can also spark new ideas that help to make it even more appealing to prospective customers. Good interaction design also takes into account the fact that consumers are not monolithic, ensuring proper access for disabled users and sometimes resulting in the development of more than one version, as with game controllers designed for smaller or larger hands.
Establishing the lifetime of equipment and products
As a rule, customers are very unhappy if the products they buy stop working after just a short time. Companies, meanwhile, may be unhappy with the idea of a product that lasts forever if that inhibits their ability to make future sales. There’s a sweet spot to be found here, and in the field of computing and consumer electronics, it’s further complicated by the fact that technologies are rapidly evolving – so a product that continues to function as well as it ever did may lose its usefulness if enough people switch to using other products that have superseded it.
Backwards compatibility is not a given, and slower speeds may put a user at a disadvantage. With this in mind, predicting the lifetime of a product is a complicated business that requires extensive analysis. Computer and electrical engineers also frequently find themselves asked to assess equipment that their employers are thinking of purchasing, with the same issues in mind. This can be anything from a single machine to a new set of computers for every employee in a company. Therefore, a great deal of money can be involved, requiring accurate assessments to be made under considerable pressure.
As well as the lifespan of a product, it’s important to know how long it can be expected to maintain optimum function. This helps software engineers to determine roughly when, and how often, patches will be required so that development time can be assigned to them. It also reveals when customers may be in the market for new products.
Furthermore, if the product is going to function imperfectly some of the time, how often will that be the case? In many end-user contexts, there is an accepted failure rate. For instance, testing and measuring devices may still be useful if they don’t achieve perfect accuracy every time, as long as they’re being used in contexts where multiple tests or measurements can be taken.
Knowing the failure rate, then, makes it easier to determine an accurate figure. This means that an electronic device doesn’t always have to be perfect as long as it’s better than the competition – at least, after factors such as price have been taken into account. Engineers need data skills to identify failure rates and assess overall reliability.
Identifying physical operating parameters
Before a piece of equipment can be purchased for business use, or a product brought to market, it’s vital to understand the conditions under which it can operate. Does it need to be stored or used within certain temperature parameters? Does it have to be kept dry? How tolerant is it of different pressures?
Real-world data is required to understand how a design that may be perfect in the lab will transfer to the context in which end users need it to work. At the very least, appropriate information will need to be supplied with it. It may need to be ruggedized to appeal to a particular market. There is sometimes a trade-off between the breadth of conditions in which it can function and other aspects of the design, such as accessibility or sensitivity, so engineers will need to understand consumer preference data in order to work out the best way to balance this.
Establishing safety requirements
Before any product can be released onto the market, it’s vital to make sure that it’s safe and, where appropriate, to be able to provide information to the customer on how to use it safely. There are very few electronic devices that can’t cause harm if users handle them inappropriately, and you will also need to know how the product is likely to behave when it breaks.
If you undertake the ECE master’s program of study at Kettering University Online, then in just two years, you could develop the in-depth understanding of data analysis needed to analyze safety hazards on a large scale and make good decisions that will keep consumers – and your company’s reputation – safe. You’ll also be in a better position to set safety parameters for developing artificial intelligences that will need to be able to make decisions autonomously in critical safety contexts, such as those that need to prioritize where multiple risks are involved in driverless cars.
Once a piece of business equipment has been purchased or a product sold, how often will it need to be repaired? Failure to properly understand the data on this places a business at risk of unpredictable expenses and makes finding the right insurance a difficult task. When making a big purchase, customers will expect to have some understanding of how often maintenance will be required. Many companies carry out repairs as part of the warranty they provide, so they need to know how much of this work they are likely to have to do when working out how to price their products.
The EU’s pending introduction of right to repair laws on electronic devices means that pretty soon, this will be a requirement for anybody selling to that market. All of this means that good data is needed to work out when items are first likely to be in need of repair, how much repairs will cost, and how many times they are likely to be necessary before the product reaches the end of its life cycle.
It’s difficult to market a product today without first establishing how (or if) it can be recycled. With electronic products, this is often a complex process. Some materials and components can fetch a good price when recycled, but they may need to be sent to different places – in which case, you will need to determine how much of the deconstruction process you can do in house and how much you want to outsource. In some cases, you may want to recycle components directly back into your own production lines. Decisions such as these need to be made on an informed basis, taking multiple factors into account. The data about the most economic options will constantly be changing, so it is useful for there to be a data-literate engineer involved in the process.
With all these issues to bear in mind, it’s easy to see why there’s a demand for computer and electrical engineers with an in-depth understanding of data analysis. It’s also a skill that has direct applications in product development, as, should you pursue this career path, you may well find yourself designing data processing systems of one sort or another. AI may be getting better at organizing data, but no-one knows better than computer engineers how far away we still are from developing artificial intelligences that are capable of analyzing it as usefully as humans – and when it comes to identifying new possibilities rather than just answering questions, they may never outsmart us. In a role such as this, you’ll be taking the best advantage of the capacities of the human brain. What could be more exciting than that?