NEWS ALERT: ERP System Saves Baby!
Forgive the hyperbole but I chose it for a reason: ERP systems do one thing only and that they are very good at – they formalize the capture of transactions of cash going in and out of an organization. At least this was the definition of the financial module, which was and still is the corner stone of ERP. Over the years, ERP as a term has morphed into meaning everything a company does. Many would now include financial and human resource planning beyond capital equipment, payroll and benefits as well as supply chain functions, such as inventory management, factory floor, transportation and demand planning. For some reason, CRM has not been brought mentally into this fold but then again, give it some more time and it will merge into Skynet as well.
I would rather not get into a religious war about what ERP was and is today and where its boundaries lie. However, what I do want to tackle is the continuing perception that a) the perfect truth about a supplier, an asset or product can be found in an ERP instance and b) a single corporate ERP instance is the saving grace from a best-of-breed (aka fit-for-purpose) Wild West of (cloud) applications. This Wild West includes all other enterprise applications in your system landscape including other ERP instances as well as the pervasiveness and ease-of-use application we all have made a living from: Excel.
A modern (cloud) ERP stack was to do away with best-of-breed applications and thousands of individual users’ spreadsheets and small databases. It was supposed to make cost rolling, accounts payable and financial closing a breeze – and it did. However, what it did not do is accommodate process variation and ease-of-deployment required by certain functional departments (hence the spread of cloud apps) and especially, country organizations. Legislative and commercial frameworks may be vastly different than the often US-hosted or at least US-designed ERP instance.
One such variation may be how cost updates are being updated on a periodic basis with little to no warning for a country sales org, effectively squeezing margins or reducing sales. Another one could be how tariff codes are interpreted and aggregated via a global HQ vs local numbering scheme. What is an unfinished good for one country, may be a finished good in another requiring additional duty. Weights, dimensions and handling instructions may differ vastly based on local infrastructure, customs and regulations. A central ERP system is often either not equipped to deal with this complexity or the implementation team has made a conscious choice not to over-customize it to mitigate future upgrade efforts.
All these variations in process directly impact the data they rely on. Therefore, without fail, we continuously find unintended and yes, even intended, data problems in our profiling engagements. After all, an employee will do whatever his/her boss says to get the job done even circumventing standard processes. This is why Excel is still around.
Here are some of our data profile findings: 56-90% of measure (UOM) or handling attributes are NULL. No global part numbering scheme with 200+ different numeric patterns affecting over 50% of SKUs. We also find a single base part number with 50 or even 500 different variations, which even surprises the organization. How about an initial ERP system data seeding creating 60% of sales SKUs with no country of origin or just the word “null”. Then there is rampant use of “one-time-vendor” setups on a plant-by-plant basis creating 12-30% duplicate vendor entries. Cost attribute management from engineering to procurement to finance suffers heavily from defaults and entry errors and typically only fast-moving or large numeric variances get caught in time.
Knowing this, would it then surprise you that a mid-sized manufacturer could potentially save $800,000 annually from additional volume discounts? After all, they could eliminate one-off suppliers, roll up to a true overall spend for a supplier with sometimes 50 or more subsidiaries, understand potential alternate or excess (slow moving special order) part numbers they may be able to eliminate.
A recent client’s South American team indicated to us that ungoverned, untimely cost variances of 5-20% affect a large portion of their salable products and a global team of accounting staff spends months reconciling these errors.
We even had clients who were blacklisted by their freight forwarder due to not or incorrectly listing the HAZMAT code when component changes were not updated. Some of our clients lost weeks to months of revenue and incurred commercial penalties due to such a slip up.
Despite all these “horror stories”, these often small-to mid-sized manufacturing organizations, which have been well-served by a historical conservative investment strategy, continue to attribute miraculous data management powers to their new central ERP instance.
Leading organizations have realized that this is flawed thinking, often instilled by ERP vendors’ promises of productivity gain and cost reduction by eliminating multiple regional ERP instances and reducing the use of Excel. The finance sector has become FinTech consequently.
Today, underpinned by legions of published studies by academics and practitioners, it is a known fact that data is not only “the new oil” but as such, should be managed in a separate instance, just like ERP transactions. After all, organizations like Google, Amazon, Uber and industrial giants like GE are proof that data holds intrinsic value not just for their own operation – think “people also buy” – but also resale value – think licensing.
On the industrial side, it appears that mostly automotive and aerospace OEMs as well as petrochemical and utility giants, have internalized this truth. To their business model, data has become a route to survival. Knowing what well head to cap, what reservoir to double down on, what block to pursue, what customer to upsell, what routing to optimize or what equipment to service next has become second nature.
I sometimes plead with my clients to please do something if they choose not to work with us (yes, that happens too) but many times the enticing premise of doing nothing is just too tempting. CODN (cost of doing nothing) is too easy of an excuse for smaller manufacturing firms. I attribute this to mind set of “if I cannot touch it or fire someone, it is not real”. Well, manufacturing and its innovations like JIT, Kanban, Lean, Six Sigma, etc. have optimized processes to the point where many departments are overly lean so many times continuing productivity gains from process optimization are elusive. This leaves us with “hard” cost reduction but even here, leadership interprets it as “buying a new machine requiring less power, cheaper raw materials or labor” – something that shows up on an invoice, is tangible or the organization has done before. Revenue or margin expansion (or just control) from data management is stuff of legends.
When I then pose the question of “how do you budget for sales?”, it typically gets very quiet because they realize that this is based on assumption as well, e.g. their clients entering new markets, building new infrastructure, consumer spending, etc. Note that these are all predictions, educated guesses and assumptions.
Also, if you were to ask someone in the purchasing department about their supplier data quality and the response was, “we are 99.9% correct”, you may want to follow up by getting a definition for “correct”. This applies even more so if your accounting staffer says she finds errors all over. After all, Informatica is in the house and I am sure the supply chain lead has not called the fire department if there was no smoke – right 😉
In summary, no matter if your manufacturing business is doing well or not, start thinking about “data management” like your ERP system in terms of importance, degree of focus; and thus, a separated effort with dedicated skills, budget and leadership. Just as much as you don’t want to miss paying or issuing an invoice, booking an order or filing a regulatory report, you don’t want your data to cause these transactions to fail or require rework. It is time for manufacturing to become MakeTech!