4. RESULTS AND ANALYSIS
4.2 Existing measures
4.2.5 Data reliability
ERP data reliability was probably the most sensitive topic around the organization. A lot of dissatisfaction were observed from most of the interviewees. When it comes to pro-curement and purchaser’s responsibilities, the data reliability refers to the item data, mainly lead times and purchase prices of the purchased items, and the arrival dates of the purchase orders. On the other hand, purchaser 2 says that the material need dates and loaded amounts cannot always be blindly trusted. There’re examples of the loadings that haven’t realized even there’s free stock available. Purchasers also notice if a loading is irrelevant or if unrealistic assembly volumes are loaded for too narrow time period.
Controller 1 had created a tool to follow procurement performance. The tool showed the following measures:
Key supplier rating
Dynamic scoring 80%
Quality related performance
40%
Delivery related performance
40%
General scoring 20%
Less dynamic factors 20%
1. Number of alerting purchase proposals.
2. Number of confirmed purchase order batches that are more than 7 days late.
3. Number of unconfirmed purchase order batches that were ordered more than 7 days ago.
4. Number of open purchase order batches without a material need.
5. Number of purchase order batches that are confirmed to be delivered later than the materials need is.
6. Value of purchase order batches that are confirmed to be delivered more than 2 weeks earlier than the materials need is.
7. Number of open purchase order batches.
8. Value of the open purchase order batches.
The tool showed many logic incoherencies, but the numbers themselves didn’t really express the whole truth. An alerting purchase proposal was a signal that the item’s ma-terial need date is earlier compared to when the items would arrive according to their lead times. However, many of the alerting purchase proposals were because item’s stock level went under its predefined minimum stock. Usually purchaser waited for more load-ings before ordering. The alerting purchase proposals may appear due to loading mis-take, because sometimes the loadings themselves had material needs that weren’t needed or the timing was incorrectly marked. Sometimes more info was needed from the other functions considering to the component specifications or drawings before the or-dering could take place. Only marginal amount of the alerting purchase proposals was alerting for a relevant reason and purchasers were able to order them.
Purchased batches without material need weren’t validate measure either. Some of the purchases were made without a loading but they were still agreed to be purchased. Many of these items had constant supply and if the loadings varied a little, this alert popped up very easily. In some cases, the loadings were deleted and then this alert showed cor-rectly that the order was really arriving without a need. Sometimes these orders can be cancelled and sometimes it is too late, and the part arrives to the stock. However, it is important to clarify how the information flows from the withdrawn loading to the purchaser and finally to the supplier, because currently the information didn’t always reach the pur-chaser and thus this measure was perceived to be important.
The purchase order batches that were confirmed to be delivered later than the material need is were not an informative measure either. Usually the purchaser and supplier had already done their best and still the confirmed arrival date is not early enough. If the loaders always updated the material needs and adjusted them with the confirmed deliv-ery date, the alert could disappear and only the order batches that had to be reacted by the procurement or by the loaders were left in the measure. However, this kind of change in updating material need dates upraised a lot of resistance from the loaders side. Also, this measure excluded all the not confirmed purchase orders from the view, even they are equally important.
The deliveries confirmed to arrive too early need also a little more examination. This excludes also the not confirmed orders, which is a significant amount of all the orders in case company’s situation. However, the value of too early arriving order batches is a relevant way to examine this type of orders. The greater the value of the too early arriving order is, the greater effect it has on the company’s cash flow. Thus, the more valuable the order is, the more important it is to get it arrived just in time.
Some purchasers did very systematic follow-up of their open purchase orders, asked for the most recent delivery plans and updated the confirmed delivery dates to the ERP.
Some purchasers again didn’t update the dates that often. The purchasers had different ways to handle their orders and especially updating the most recent information to the ERP was many times left undone. This was also noticed by the production manager, as he mentioned category-specific differences considering the arrival date correctness. Ac-cording to COO, the ERP data is not maintained, and the performance measures are thus not upheld.
The ERP system allows a lot of exceptions, but at the same time the controllers say that these exceptions mess up the big picture. If the exceptions are allowed, it arouses errors and irrelevancies in the performance measures. Purchaser 1 says that the performance measures should illustrate reliable values. Head of global procurement even said that it is secondary what kind of performance information is available, if t’s first reliable. If the raw data behind the measure were more reliable, the measure show also more reliable results.
Because the measures had a lot of pitfalls and the procurement didn’t find them to be relevant, they weren’t used and followed. Thus, better internal performance measures should be developed. When developing the performance measures, it is equally im-portant to develop the tools for the procurement they can use in order to improve the
performance. For example, a single value of too early arriving orders is not very useful for the procurement if they can’t sort a list of orders that are about to be delivered too early.