There have been many changes to the manufacturing landscape as Industry 4.0 takes hold. Manufacturing managers of today must now have a predictive maintenance/monitoring program in place to help reduce costs, decrease waste, have fewer breakdowns, and remain competitive. Many companies faced with a decision of how to implement such a solution, consider whether to build or buy the necessary software to monitor their plant conditions.
When considering whether to purchase a predictive monitoring software from an outside vendor, or to build your own predictive monitoring software internally, there are many factors that should come into play in your decision process. These include:
Uniqueness of your plant
Time to value
How much of your plant you need to monitor
Corporate culture, goals, and needs
Let’s look at each of these aspects in turn.
Uniqueness of the Plant
Often companies consider building their own predictive monitoring solution because they feel that their specific plant has its own unique requirements, that cannot be met by an off the shelf product.
It is true that certain plants may have developed customized processes, or even acquired customized equipment, however, at the most basic level of process industrial equipment, the components are often similar across many industries. There are valves, heat exchangers, pumps, and so on, and the measurements that need to be taken – pressure, temperature, vibration, etc. – are similar across many industries.
If there is truly a unique aspect to your plant that requires a customized, self-developed solution, then this could make sense for your company. However, in most cases, the benefit of working with a software company that has already developed a solution and proven it in scores of other plants will outweigh the small issue of customization.
Developing predictive monitoring requires a strong understanding of data, and the experience of what aspects of the data to ignore and knowing what is significant and needs to be explored more deeply is something that comes from many years of experience and based on implementations in many plants.
Time to Value
Let’s look at how quickly you could achieve value when you are building your own solution. You will need to invest in R&D including data scientists, additional technical support team, user interface experts, and training. You should expect quite a lengthy and challenging process.
Finding the right people to hire, and then on-boarding them, can take anywhere from one to five months. Anyone who has tried to hire a data scientist lately knows how challenging this can be, as you are competing with the likes of Google and Facebook for talent. In a recent report by LinkedIn, there is a national shortage of people with data science skills to the tune of over 150,000 people in the USA alone.
Then the building process, which, depending on the skill level of the team and the complexity of what you are trying to accomplish, takes additional time. PWC breaks down the key steps to beginning the predictive maintenance journey internally:
Asset value ranking & feasibility study
Real-time performance monitoring
Only after all these steps are completed, tested for quality assurance, and users are trained, will the solution start to provide value. We estimate this whole process to take anywhere from six months to 1.5 years.
Contrast this with buying an off-the-shelf solution which, in the case of Precognize, takes typically two weeks to install and be fully running and protected.
Coverage of Your Plant
What do you need to monitor in your plant? Can you prioritize the various processes down to the specific machines or even valves that you need to monitor? If you plan to build, you may be tempted to only monitor certain areas of the plant, rather than the entire entity and you can be vulnerable to plant glitches and problems.
In fact, in the PWC report referenced above, in the section called “Asset Selection,” the authors specifically warn “Keep it manageable and don’t try to cover your entire fleet or factory in one go.” In other words, monitor what you think is most important, or what is most likely to have a problem or failure that can cause critical damage or downtime.
However, the largest benefit of predictive monitoring comes from the day to day savings it unlocks, by identifying business leakages or places where the plant could be running more efficiently. These types of discoveries may have the dramatic flair of preventing a major shutdown, however they come more often and require monitoring the entire plant, since there is no way you can know where these important savings can occur.
Building a solution to monitor your entire plant can take years, if you are, indeed, as PWC recommends, biting off the project in small pieces. And yet, there are off-the-shelf solutions – like Precognize – that will monitor your entire plant from day one.
So, if you are looking to focus on preventing the major issues, building a solution may make sense, but if you want to benefit from the ongoing upside of having a solution that covers your entire plant, you would be better off buying a software solution.
Your Core Business & Company Culture
In some large companies, the culture dictates that everything must be developed in house. This will allow you to customize the solution 100% to your needs, and to maintain control over its source code as well as the entire development process. Knowing that it will definitely fit your current infrastructure will lead to more confidence while monitoring the plant.
However, it can also be a distraction from your core manufacturing mission. In very large conglomerates, often there are enough different departments and capabilities that software development is already well entrenched in the company, so building your own solution may be strongly considered. However sometimes the large number of people involved in making each decision slows the process down.
Software companies that develop off-the-shelf solutions are typically much more agile and responsive than large manufacturers, with the added benefit that software is 100% their core capability. As we pointed out earlier, software companies such as Precognize have implemented their solutions in scores of similar situations, have already encountered many of the problems that may come up and have found ways around them. As a result of their key personnel, core capabilities and years of experience, software companies are well positioned to ensure the success of your predictive monitoring implementation.
Making a Final Decision - Build or Buy?
We’ve put together a handy table to summarize the key points discussed in this post.
There are quite a few considerations to think about when deciding if you should buy or build your own predictive monitoring/maintenance software. Your decision should be made solely on the plant’s needs. Understand the uniqueness of your plant, the time investment you can afford to make, the company culture, and what your goals are before deciding whether to build or buy a predictive monitoring solution for your plant.
Considering whether to build or buy a predictive monitoring solution? We’d be glad to help walk you through the process, drop us a line!