Perhaps you’re feeling a little overwhelmed trying to run a manufacturing operation in an increasingly competitive, digital business landscape: perhaps you can’t get any, enough, or the right data from your factory floor; maybe none of your data systems talk to each other. Meanwhile, that ERP/manufacturing execution system (MES)/big-software vendor has painted a beautiful picture of how all your data could work together—if you just scrap all those disparate systems and start over. Just sign on the dotted line, and the pain will go away…
You can make this pain disappear without giving massive amounts of organizational time and money to big software vendors (and their armies of system integrators). Instead of throwing everything away, why not integrate the systems you already have?
At 3AG, we have worked with every kind of data system, and have yet to encounter one that cannot be made to talk to any other system. As data engineers, it is literally our job to help organizations like yours make more efficient and effective use of corporate data by cleansing, aggregating and organizing raw data into usable, structured formats. This is what big ERP vendors do as well, but they take the expensive and generally unnecessary step of deconstructing and disposing of perfectly good, if “rusty,” data systems first.
Instead of uprooting old data systems and replacing them with one brand-new, overarching system, connect them with “digital glue” into a central data warehouse. Data warehouses, when properly designed and built by data engineers, make all your data easily accessible for reporting and analysis. For example, instead of exchanging manufacturing software, planning software, and accounting system for a single ERP, a central cloud-based data warehouse that pulls data from all these systems will be equally effective.
Via the “digital glue” approach, users would continue using old systems without disruption, while analysts would almost immediately benefit from data centralization and standardization. And you would, of course, save time and money by not replacing a perfectly good handful of data sources with a whole new setup.
Here are 3 ways good data engineering can extend the life of your assets:
- You can connect any database/MRP /ERP/finance system to any other system
- It’s significantly cheaper than the big-ticket IT system
- You will minimize disruptions to your operations
1. You can connect any database/MES/ERP/finance system to any other system
You may be surprised to learn how flexible data system integration can be. How flexible? You can literally pull data from any existing data system. By literally, we mean literally. At 3AG, our preferred approach is to apply a data warehouse, which is exactly what it sounds like. Just like a physical warehouse, a data warehouse serves as a secure central repository for storing your data.
There are many advantages for companies with scattered data systems to apply a data warehouse:
- Data extraction can be automated
- Non-cloud data sources can be uploaded to the cloud
- Users can trust a single source of the truth (vs. endless debates about the validity of a number)
- Formulas and rules can be applied to raw data for intermediate numbers
- Both modern and legacy data systems can be connected
- Data can be sent from one system to another without the need to program a unique path each time
Generally speaking, the companies we work with have a large number of data systems, ranging from on-site historians, MES and material requirements planning (MRP), ERPs, homemade supervisory control and data acquisition (SCADA), spreadsheets, and accounting systems, to name a few. The commonality here is none of these systems talk with each other. Rectifying this is where data warehousing shows its immense value.
By pulling data from all these systems into a single location, users will know exactly where to look for specific data. Through the process of normalization to reduce data redundancy/duplication and increase data integrity, users can be sure they are dealing with consistent units, time stamps, and even spacing when they read data.
Driving consistent performance is a hallmark of any good manufacturing operation. Consistency is just as important when it comes to accessing and using data, especially when data is scattered across multiple operations—especially since data is one of your organization’s most valuable assets.
2. Data engineering is significantly cheaper than the big-ticket IT system
For mid-size manufacturers, the typical estimated cost of an ERP system is $150K-$750K—and there’s no guaranteeing costs won’t exceed expectations. Much of a big-ticket ERP’s final price tag is integration costs, as vendors work to customize generic ERP features to fit each client’s particular requirements.
There are major issues besides skyrocketing costs associated with this approach. First, ERPs are designed mainly to record activities. Data tends to go one way in an ERP, and is traditionally difficult to extract. And because it’s highly unlikely a big new ERP system will be perfectly set up from the beginning, any future changes to your system will result in…more trips to that army of system integrators.
ERPs and similar systems also tend to be all-in-one. In theory, they work well when you switch everything over to the monolithic system. The challenge is, this requires the wholesale change of a wide range of business and shop-floor practices at the same time; there is significant cost resulting from the resulting productivity loss and process disruption.
By contrast, data warehouse integration can be performed in stages, one system at a time. Instead of revamping everything at once, a well-designed data warehouse provides a foundation for incrementally and strategically adding new systems. Data warehouses allow flexible integration of data both to and from different systems, so it is far less disruptive to existing operations.
Making incremental upgrades to your overall data system means you can spread project costs over a long period of time and make changes when doing so is most convenient. This is much less risky than throwing everything into a single massive project that may not be fully operational after the ERP vendor’s first installation attempt.
3. You will minimize disruptions to your operations
In our decades of experience, we’ve seen firsthand how ERP and other big data project integrations work (and how they don’t work). Either they require the client company to change its business processes to conform to default ERP settings, or they require significant integration work from either a third-party vendor or that in-house army to codify current business processes into the ERP.
Another common mistake companies make when adopting an all-encompassing ERP system is this: They try to fix business processes and code their ideal state parameters into the ERP simultaneously; this is analogous to building the plane mid-flight. If an organization’s ideal state doesn’t precisely mirror what it can do in reality, significant problems will result. For example, if your team logs a particular manufacturing step on paper instead of in the ERP, but the ERP is programmed to expect this data, inaccuracies will begin accruing. Either the real-world process or the ERP digital representation of the process will need to change—perhaps repeatedly, depending on how well employees adapt to using the new system.