[原创]Management of Data Flow in an ERP Environment

Although excellent data accuracy and high data quality are one of the cornerstones for the successful running of ERP systems, some people still subconsciously don’t believe they are as important as product quality. In a sense, this is comprehensible. Product quality, after all, is much more overt and tangible, for customers will reject any products of poor quality and even demand huge claims; while poor data accuracy and data quality are not felt by customers at all. However, poor data accuracy and data quality will definitely increase our operations costs, slow our response to changing customer demands, and lose opportunities to win more customer orders and strengthen our relationships with customers.

To make things worse, there are widespread conceptual misunderstandings of data accuracy and data quality. The commonest one is that once basic data coding and data entry have been established during the successful launching of an ERP system, data accuracy and data quality are only related to data entry errors and inventory data accuracy. Actually, one of the biggest contributors to poor data accuracy and data quality is poorly designed processes and bad practices. For instance, infeasible work orders are often implemented halfway and then have to be frozen; leaving lots of semi-finished goods lie idle until the work orders are finally completed. This will plunge the current data flow into turmoil. And the distorted data profile will start misleading ERP users. What comes next? More infeasible production scheduling, blind purchasing or panic material build-up, more distorted data flow, and people’s desperate and secret turning to any “more convenient and reliable practices” that bypass the ERP system. What an evil cycle!

On the other hand, some degree of data distortion (in the fields of data accuracy and data quality) is simply an unpleasant part of reality: we can’t eliminate it completely for one reason or another. Therefore, it is our unshakeable focus to keep data distortion under control and minimize its negative impacts over our operations. The passages below are therefore mainly focused on how to manage the data flow in an ERP environment by attempting to eliminate bad practices and establish better processes (limited to the management fields of planning and scheduling, purchasing, and inventory).

u       Typical bad practices that distort data flow and how to alleviate their data distortion

What are bad practices in an ERP environment? Put it simply, any infeasible activities that can’t be implemented as scheduled, and any activities that prevent us from discovering problems and matters hidden in the system. Below are some of the commonest bad practices:

1.         Infeasible production scheduling
Infeasible production scheduling is public enemy No.1 because it only not makes our production unnecessarily unstable, wasting valuable production resources such as labor, production capacity, etc; but also deteriorate data accuracy and data quality (countless unconsumed WIP stocks scattered in numerous production lines are always a headache, for example)

2.         Untimely data updating

Data updating here refer to two things: inaccurate data, say, inaccurate inventory data, failed promised delivery dates, BOM errors, etc; the periodical maintenance of order modifiers or other important system parameters and other important basic data like BOMs. Inaccurate data is a spot data distortion. Improperly maintained parameters are systematic data distortion. Both of them will make the data flow muddy and twisted.

3.         Secret “more convenient and reliable practices” that bypass the ERP system

In some companies where an ERP system has been used for several years, It’s not surprising to find out that some ERP users are secretly using their own “more convenient and reliable practices” that bypass the ERP system. For instance, in a company, the buyer uses Excel spreadsheets to monitor the consumption of all of its chemicals and purchase those chemicals accordingly; he rarely looks at the data of the chemicals in the ERP system, for he knows very well they are actually out of use. And the ERP system is only used as a final transaction recorder.

However, based upon the “more convenient and reliable practices” mentioned above, we can’t design or support any decision supporting mechanism / systems.

Why such bad practices exist is a big question and well worth further research and study. In my opinion, lack of a holistic picture of an ERP system and poorly designed processes are two important factors. Here I just list some of the practical procedures used to attempt to alleviate those data distortion. Better solutions may then be identified to eliminate those bad practices once and for all.

1.       How to tackle infeasible production scheduling?

Infeasible production scheduling is the epicenter of a data earthquake. It occurs because of lack of understanding of its huge destructive impacts over physical operations and the normal data flow. To tackle this matter, both discipline and better principles must be followed.

Ø         Production scheduling must be material shortage free and not capacity constrained

Ø         “One piece flow” or “multiple small batches” principle makes work orders lean and nimble
Lean and nimble work orders minimize the negative impacts over production and the data flow when things goes wrong. Secondly, Lean and nimble work orders shorten the production cycle by timely entering finished goods in the ERP system generating a more transparent and near real time data flow to reflect and guide the actual operation. Thirdly, Lean and nimble work orders make the company more flexible to respond to changing sales demand.

Ø         Highlight on order items needed to support specific work orders

List all on order items needed to support specific works. Once they are not supplied as scheduled, timely adjust current production schedules.

Ø         Proactively identify and tackle potential problems

While scheduling production, always proactively be aware of any potential problems that hinder the execution of work orders. For example, when the on hand qty of an item is just equal or above the qty required to build a work order, especially when the accepted inventory different percent is big enough to make a significant impact over that work order, an investigation should be made before the work order is fixed and submitted, for it may be a bit too late to discover that the on hand qty is actually not big enough to build that work order.

       2.    How to tackle untimely data updating and secret “more convenient and reliable practices” that bypass the ERP system?

              Solutions are quite simple, here are some of them:

l         continuously improve data accuracy so that the overall data accuracy is kept under an acceptable level,

l         periodically maintain order modifiers to ensure that data generated in the ERP match the physical world;

l         stop any practices that bypass the ERP system.

However, it’s easier said than done. Some processes / mechanisms must be established to ensure that the ERP system is not only complex recorders of final transactions financially or operationally, but also efficient data analyzer, problem identifier and solver. In other words, an environment must be created where ERP users can easily visualize the complicated relationships between the various data that are always flowing in the ERP system, and they can also be prompted when abnormal data flow is identified. Furthermore, they must be continuously trained to believe that all their problems can be more effectively solved via ERP. Otherwise, dealing with inaccurate data and poor data quality will always like pressing basketballs into the water with one hand.

Below is the introduction of an attempt to help establishing those processes / mechanisms.

u       Establish an item profiling management system

ERP systems’ ability to integrate originally isolated functional data/info makes it possible for us to establish an item profiling management system, which provides valuable guidance for buyers, planners, and product engineers, etc. Below are the basic elements of the item profiling management system.

Ø         Demand profile
This portion outlines the demand side of an item (Finished Goods or Raw Materials) in the past (history demand), at present (on order demand) and in the future (forecast) in the fields of

²        Demand nature analysis
Identifying the nature of a demand: hard demand, safety stock demand and forecast, an important factor used to prioritize demand

²        Demand Trend analysis
depicting the demand trend and demand pattern of shipped orders, booked orders and forecast while providing basic demand indicators like weekly/monthly demand average, pan size, etc.

²        Demand Stability analyses
recording the frequency of order date pull in / pull out, order qty increase / decrease and/ or order cancellation combined with differences of the actual fixed order and forecast to project the stability of the demand

Ø         Supply profile
This portion outlines the supply side of an item in the fields of

²        Availability analysis
not only listing the basic info of on hand qty, allocation qty, on order qty (within a specific time fence) but also projecting its time phased ATP qty

²        Reliability analysis
providing delivery performance data of the item, quality records, rated inventory accuracy percentage to project the supply reliability of the item

²        Substitute analysis
listing not only a list of substitutes to replace an item in case of need but also alternative suppliers / equipment to form a reliable emergency plan

u       Establish an item relationship management system

In an ERP environment, all data is interwoven and interacted to reflect the physical interwoven and interacted activities. The establishment of an item relationship management system facilitates the visualization of those complicated relationships among data. Besides, the newly gained visibility of those complicated data relationships gives us insights to identify any data flow (that reflects physical activities, say, sales order confirmation, purchasing, production planning and scheduling etc) that are out of proper relationships.

Ø         Dynamic Parent (Where To Use) /Planning BOM profile
For a child item (make or buy parts), this profile lists the specific qty required by all its parents that have been demanded within a specific time fence chosen by the user,

For a parent item (Finished Goods or make part), this is actually a simulation mechanism of production scheduling in an attempt to facilitate planners’ efforts to create material shortage free work orders. The finalized work order of a specific item is definitely dictated by two factors: the qty of all its components available for this specific parent item, and the production capacity available for this specific work order.

Ø         Capacity Workload Profile

This capacity workload profile lists and monitors the of workload status of some key equipment/routines. It is again a workload simulation mechanism. It can calculate and examine the workload status of specific equipment/ routine where many work orders (Finished Goods, semi-FG) are going to be processed. (sectional workload examination). It can also calculate and examine the workload status of all key equipment/routines a specific work order is going to use (linear workload examination). The combination of the two sorts of workload examinations will determine two things (within a specific time fence:

²        The feasibility of a specific work order from the point view of capacity

²        The maximum number of work orders that the company is able to produce under its current capacity constraints

Ø         Customer/ Supplier Profile
For a finished goods, this profile lists all important info of a specific customer. For example, the importance of the specific customer, if there’s a penalty clause on the sales contact with the customer, or this item must go with another item also ordered by the same customer, etc.
For a buy part, this profile lists the overall delivery performance of the supplier; current constraints that may affect the delivery of the buy part, etc

Ø         Demand VS Supply profile
This profile monitor the ever changing demand and supply situation of an item within a specific time phase say, daily, weekly, monthly: if there exists overproduction/overprocurement, or underproduction / underprocurement, if a finished goods, or purchased/make items go to stock too early or too late, what are current trouble item / supplier / customers on which special attention be paid, if there are some customer orders that can’t be satisfied, what are the reasons of the failure to satisfy those customer orders, etc.
In short, this profile helps us to monitor and control delivery progress of finished goods, to analyze if current inventory is rational or not, and how to maintain rationality current order modifiers and other system parameters to further fine tune the data flow the ERP system.

In reality, the item profiling management systems and the item relationship management systems are closed linked with one another. They might be of little value theoretically, for all of the info is nearly readily available in a standard ERP package. However, an integrated approach to provide ERP users with the complete set of available-on-the-figuretip info to support their decisions is absolutely a must, for only well designed processes can better ensure that production operations are done as scheduled, which will be the smooth data flow reflected in the ERP system. And only with a smooth data flow, the matter data accuracy and data quality can be easily tackled and continuously improved.

Incidentally, all of the info mentioned above may be achieved by data warehouses or data mining techniques that provide multiple dimensional drill down analyses. More sophisticated info may be provided by systems like BI, CRM & SRM.