• Accounting Systems
  • Cash Flow Processes
  • Advisor Coordination
  • Budgeting & Forecasting
  • Production & Operations
  • Banking & Capital
  • Strategic Financial Planning
  • Risk Assessments
  • Revenue
  • Operating Rhythm
  • Succession or Exit Planning
  • Readiness & Attractiveness
  • Price/Multiple
Summit Insights
Aug, 21

Manufacturing ERP System Implementation Challenge


“We know we need better information, but this software just doesn’t work,” the purchasing agent replied to the question – "What are the biggest issues you face in doing your job?" He was the last of the team we interviewed, and he was repeating the same thing that we had heard from Engineering, Accounting, Sales and Operations. 

Ben’s was a successful equipment manufacturing company that had grown from literally nothing in Ben’s barn to over $50MM in sales in a new state-of-the-art manufacturing facility. They bent over backward for their customers and they had a reputation in the market for building great equipment and working well with customers. The company made large (3,000 cubic foot) custom production equipment. While each piece of equipment performed the same basic function (chemical processes under heat and pressure) most units were unique in several ways (the pressure rating, gauges, the coating of the outside shell, valve manufactures, etc.). A typical unit contained 1,000 components (some purchased, some fabricated onsite) and took 1,000-man hours over six weeks to complete. Some of the more unusual purchased components had 12-week lead times. 

We had been called in to help, after a steep drop (40%) in their market unearthed working capital issues which showed up in the form of a cash flow crisis. After spending several weeks focused on the urgent matter of “right sizing” the business to match current market conditions (the subject of another case study) ownership and management – posed the question – what’s wrong with our ERP software? 

Ben told me five years ago, they were doing fine, but they were starting to get “stretched” on hitting their deliveries. When the industry hit a huge growth spurt, what was “stretched” became broken. Their suppliers and other friends in the industry told them that they had outgrown their systems and that a $50MM company needed an Enterprise Management System (ERP) to take control of their business. 

Ben and his team had no experience with this kind of system, and they relied on the software vendor to make the right choice. A very reputable and expensive software system was chosen, financed and installed. It had been available for three years by the time we were invited to help– but it was never trusted and was not used as an ERP system should be utilized. The company was suffering because of it. Even before the industry downturn gross margins had been deteriorating and the company experienced huge annual book to physical write downs (5% to 7% of total inventory) at year end. Their financial stakeholders, vendors and banks, were losing confidence in the company’s ability to pay them back. 

So, we spoke to everyone on Ben’s team that used the software separately and asked them what they thought was going on. Everyone (literally) blamed the software package’s “bad information” for the decreasing margins and increasing write-offs. “The software doesn’t work for us; it can’t be fixed, and it should be scrapped”. 

Engineering was responsible for loading the Bill of Materials (BOM). They told us that Sales tended to be optimistic on deliveries and that it was company policy to allow change orders up to 4 weeks prior to unit ship date. With the downturn in revenue, sales said customers were demanding these kinds of concessions. In response, for each new unit ordered Engineering created a Pro-Forma BOM – that they knew be about 80% correct as “a place holder” in the system. When the final print was received by the customer around a month before shipping, the one engineer, who was an expert at the system, would load up the final print and correct the BOM and attach it to the job order that had just started on the shop floor. If any of the new parts were unique, it would take extra time to load these new individual parts into the system. He would walk the final print over to the Shop Floor Scheduler. Sometimes based on how busy the engineer was (or if he was on vacation) the final BOM would not get loaded for a week or so after receipt. Final engineering approval of the new parts and data load of BOM with corrections became significant bottleneck. This was a big issue because many of the last-minute changes were technical in nature. Only an engineer could understand the changes and only one engineer “knew how the ERP system worked”. If he was unavailable or tied up with something else – data load had to wait. If the process had to be rushed to make sure the promised delivery would be met, purchasing and the shop floor scheduler had to do “work arounds” while the final BOM was being loaded. 

Accordingly, the purchasing agent (PA) spent his time massaging the buy order requests produced by the system. He was uncertain of the validity of the requests because he knew they were only originally created by “Pro Forma” BOMs. Unless he was sure it was a good request, he was reluctant to release these as POs despite engineering’s assurances they were 80% correct. But the myriad of system buy- order requests created by the Pro Forma Jobs disrupted his visibility into inventory requirements. To overcome this the PA created a workaround where he released these buy order requests (attached to the specific job, so they would not impact inventory) as POs and then held them in his office. Because they were not “real” POs he did not pay any attention to the price or the delivery on the PO. This way he removed the “noisy bad information” from the ERP system. The PA would then regularly confer with sales, engineering and the shop floor manager to see what they “really needed”. Based on their input the PA would search through his pile of issued but unsent POs and find a PO that was what he needed, or close to what he needed. He would then phone in that PO to the vendor, adding or deleting a few items as necessary. Typically, he did not have time to correct the price and delivery, but they usually were marked for the correct job. Because of misunderstandings and mistiming the incorrect parts would be purchased. Then purchasing had to rush in the “correct” parts, typically from an alternative vendor at a higher part price with expedited freight costs to avoid delaying the shipment of the final unit. The PA was proud that he could get “whatever was needed” to get the product out the door. Over time the PA had accumulated a whole stack of POs (mostly for closed sales orders) that had never been actually sent to the vendors. Periodically he would try to update, reconcile and correct the orders in the system, but usually “he didn’t have the time” so he would wait until the sales order had shipped in its entirety and would cancel any of the paper POs he had created in the ERP system but did not actually give to a vendor. 

The shop floor scheduler’s routine became extremely complex. He had to make sure that he received the final print on jobs issued to the floor as they began production and assembly by talking to the salesman in charge of the account, ensure that engineering had taken the time to update the BOM versus that final print before he sat down with purchasing to discuss parts that were needed but not yet received. 

Shop floor personnel were encouraged “to do what it takes” to make sure that sales orders shipped on time. They had very low confidence for the information in the system and what they were told by engineering and particularly purchasing. When product arrived from vendors, they knew that the sales order referenced on the packing list wasn’t always the correct one. With the ship at all costs attitude, many times common components like gauges and valves that were equivalent (or better) than the ones called for on their final drawing would be grabbed out of receiving and installed into a unit being assembled for final shipment – even though those parts could have been meant for another job. Sometimes this part substitution would be noted in the production traveler or the internal shipping documents and sometimes it would not. All of the kitting was supposed to be loaded into the ERP system by the shop floor supervisors, which they regularly did. But they were not given permission to “kit” components that were not in stock in the ERP Inventory system. Also, jobs could not be completed until all the components were kitted. Units could not be shipped (the Shipping paperwork would not print) until they were completed in the ERP system. Customer trucks could be waiting, so all parts that did not show in inventory were “written up” by the shop floor scheduler to the necessary quantity to be able to get the paperwork done and the product shipped. These write ups would get investigated if there was time – many times the issue was caused by undocumented component substitution. But, if the cause of the inventory inaccuracy was not found immediately, they would stop looking and go on to more urgent matters. As you would guess if the substituted part did not get relieved from inventory correctly and still showed in stock, there would be another emergency 3 weeks later when production went to pick the component for the job it was meant for and it wasn’t there. Another expensive rush order to save the day and more proof that “the ERP system doesn’t work”. 

In production, the ERP schedule was ignored and an excel spreadsheet on the production managers PC and a white board in a conference room were used to manipulate the schedule and display it. These got updated once a week (in a labor-intensive effort by the production manager) for a weekly production meeting. For anything out beyond a couple of weeks you had to talk to the shop floor scheduler or production manager to get an estimate. The sales team would take these disparate promises and bring them into the production meetings a couple of weeks later and they would hash out who got what. On the shop floor on top of the issues of kitting components to the correct jobs, clocking in and out of jobs was haphazard because “nobody looked at it anyway”. 

And in accounting it became harder and harder to just process a vendor’s invoice. Almost every item had an exception that had to be chased down. The cost, quantity and order number on the PO couldn’t be relied on completely so accounting spent more time on the phone speaking with frustrated vendors demanding proof of delivery and sales order acknowledgements for verification. Variances were sent to purchasing for review but there was a substantial delay getting any resolution because the PA was so busy. Most times the vendor was proved right because the Company’s position could not be substantiated. To make matters worse, after the downturn, the AP staff was cut and vendor invoices, POs and receiving paperwork sat in stacks – typically unreconciled until the check was cut for an angry vendor demanding payment. Some very big surprises popped up and it was extremely difficult to determine what actually happened two months after the components arrived in the building. Closing the books became more complicated as the process became bogged down. Margins swung wildly from month to month and there was no way to find the root cause. Accounting just went from one fire to the next – a special rush check for a vendor holding product, a report for the bank or the owners, closing the books, etc. etc., so they never had time to try to address their process. 

Everyone blamed the software. 

We began the journey to create a process. This constituted a required weekly meeting where everyone involved in the ERP process met on Thursday morning. The only reason you were allowed to miss this meeting was you were on vacation. The purpose of the meeting was to establish quarterly goals (what does success look like?) and set actions to achieve those goals. The lead person in each department was responsible to complete (ascertaining what manpower, training, money and other resources were needed) the goal and give an update every week on the progress towards the goal. 


Before the first meeting, we had to confirm the owners’ commitment. Ben said the right things, but I was not sure he wasn’t giving us lip service. So, we talked about whether he was willing to push back against unrealistic customer demands and more importantly set clears goals and processes to attain them, train his people if necessary, and finally hold his people accountable if they do not perform. Ben had to be willing to let employees who couldn’t or wouldn’t do the job go. Ben agreed. 

With that commitment in hand, we focused first on promised lead times and change orders. I pointed out to Ben that the company had actually set itself up to fail at the outset by allowing changes up to two weeks before shipment on an equipment job that takes 6 weeks to complete. The ERP system had a change order routine, but it was not used. We decided this would be a good place to start. This change allowed us to tackle a fundamental flaw in Ben’s process – incorrect and incomplete BOMs being loaded into the ERP system. We were able to establish that no order could be accepted and entered into the system until there was a complete and accurate customer blueprint (signed off on by the customer) and list of components. If the customer wished to make a change – they could but they would get a new price and a new delivery date. The “clock would restart” with every change. They would also be charged for custom components that could not be used that were ordered for them that could not be returned. Late deliveries and shop floor “rushes” dropped considerably. Also, many customers remarked that this “new” policy from Ben was what they actually expected and had wondered why it had not been instituted sooner. The salesmen’s predictions that we would lose business due to our “lack of flexibility” never occurred – we had taken the time to warn customers of the new policy and made it clear to them that we were still going to do everything we could to get them exactly what they wanted – but they would have to pay for extra change costs and incur delivery delays for component changes when they were due to a customer change. 

We sat down with the entire team and went through the process and included an ERP system upgrade meeting right after the weekly scheduling meeting. The owners sat in on this portion of the meeting. They admitted that they had been a big part of the problem by saying it was OK to allow changes without consequences and announced that it was going to be different going forward. In this meeting each leader identified the biggest issues that they saw related to the ERP system in their department. They proposed a solution and a time frame to fix it. 

“Over 90 days” 

In engineering there was a huge new bottle neck. The extra time to get the orders right in the beginning was being combined with the time it took to process old orders that had to be fixed. An extra engineer was dedicated for the next 60 days to help. Issues with costumer drawings were identified and resolved with the customer, at the beginning of the process, before the final quotation/ order.

Manufacturing discovered they could do costs per job – and created a model showing the actual quote in real time. Clock-in and kitting were reviewed real time, by department leads, and the owners had a build report for every job in process. The owners started reviewing every job that was below quoted margins to learn why it had not been quoted correctly. They visited the shop floor and talked to the welders and learned why there were over-runs and spoke to material handling etc. etc. The schedule in the system became accurate so the work around the excel spreadsheet was abandoned and everyone looked in the system to check delivery. 

Accounting stopped falling further behind but needed outside help to catch up on the months of backlog. They brought two temps in for a month and got caught up – material variances were now discovered when the invoice was received – and were sent to Purchasing for review before payment was due. Many times, issues where we had to defer to the vendor in the past (because we did not have the time to figure it out) were now resolved in the Company’s favor. 

Weekly meetings showed progress everywhere except Purchasing where the PA continued to use work arounds. It seemed that he had gotten used to being the hero and enjoyed being the only person that could “make it work for the customer”. He insisted that the ERP system was still inaccurate and refused to admit he had made mistakes and purchased the wrong components. He hid the fact he was still using verbal orders and that he did not use the buy order requests produced by the ERP system. It became evident that the PA could not or would not fix his issues – he was the only guy not making his weekly targets and the rest of the team was offering to help. And when he did not take the help and continued to miss his deadlines the rest of his peer group started riding him to get it right. 

Finally, it was obvious that the company had to let him go, even though he was related to the owner. It was a very difficult decision for Ben, but one that was necessary for the team and the process to be able to grow. 

A new PA was hired and within a week she found 150 “old” POs that were not right and canceled all of them that did not tie to an order. She reached out to vendors and asked for their help in transition. The vendors were happy to work with the company. Within 30 days the process was directionally correct and purchasing, engineering, sales, manufacturing, and accounting were on the same page. The company finally had gotten to the point where there was just one version of the truth. Everyone on the team could work in the same direction. 

The weekly meeting continued, but new goals aimed towards improving the company processes were being implemented. The owner’s vision and long-term objectives became the litmus test for quarterly goals. It took some time (six months), but continuous improvement became an expectation. Sales, profit margins, on-time delivery and customer satisfaction all improved. The annual book to physical write-down was reduced by 80%. The company now had a platform for future growth – good information and a team dedicated to using it to achieve their goals. 

The takeaway to this story is that we have to remember that an ERP system is like a chain – it is only as good as its weakest link. It is NOT just software and hardware. It is a process where the people are the most important component. It only works if the team is trained and committed to working with the system. And that commitment has to come from the top of the organization. 

It takes time, money and owner commitment. The process works best if ONE person has to lead the implementation. It takes a lot of time to implement one of these systems and you still have to ship product while it is being done. With alignment of the team from top to bottom, a leader holding everyone accountable and a weekly meeting process with achievable quarterly goals, it can be done and done well. 

Once it is done right – you will wonder – “How did we ever work without a good ERP system?” 

Are you interested in learning how a Fractional CFO can help your business with ERP implementation? Reach out today for a complimentary consultation.

Greg Gens is an Area President with FocusCFO based in Cleveland, OH.