Data Cleansing - Use Quick Wins and save money

Duplicates in the article inventory are a thorn in the side of every company. They cost money. Unnecessary money. Money that can be saved and put to better use elsewhere. Cleansing the duplicates in the inventory would enable immediate savings, and quick wins. With modern and established methods, Data Cleansing is an efficient approach to permanently remove duplicates in the item inventory. Data Cleansing is economically very reasonable and can be realized quickly with SIMILIA.

Put an end to duplicate drawing parts, and keep your item inventory clean, with Data Cleansing for industrial companies.

Data Cleansing –the term originating means "to clean or correct data and errors".

Data storage errors creep into the item inventory of industrial companies for a variety of reasons. Often, the errors are due to careless or inconsistent data handling. This cannot be completely avoided in the industrial environment with conventional procedures. So it is a quite common effect, which one accepts with gritted teeth willy-nilly, because the effort of data cleansing, the data cleaning, is considered to be extremely high. The widespread assumption is that you have to live with it.

So far, at least.

Once you have decided to perform data cleansing, the next question immediately arises: Where do I start? How can I identify duplicates? The first step is known to be the most difficult.

Let's put ourselves in the position of a company where these errors have crept in over the years. Cleaning up this data is an arduous task. If the company also has multiple locations, multiple plants, different CAD systems, and different data repositories, it all becomes significantly more complex - an almost impossible task. Linguistic differences in the designation and naming of items further increase the degree of complexity in data cleansing. If you also have Asian characters in the data fields, it is beyond the imaginable effort to be able to realize a cleansing of the data here. This constellation is not an isolated case. Many medium-sized companies have several locations in Europe, the USA and ASIA. It's a horror vision that opens up when you think of data cleansing in this context, isn't it? So you say to yourself: let's leave it alone. Somehow this will work out. On the contrary. At the latest during the next data migration, during the next introduction of a new ERP system, during the consolidation of location data or the like, it will then fall on your feet. The benefits associated with data cleansing should not be wasted. Considerable potential benefits can be mobilized quickly, and they pay off. In the following, we would like to show you how data cleansing can be realized economically and with manageable effort.

Initial situation in companies

Today, it is common for companies to accept duplicates and multiple installations of drawing articles in their databases. This is despite the fact that they know that these duplicates are cost drivers and cause avoidable costs. It is not just duplicate data storage costs, but duplicate or even multiple efforts in the entire organization of the company. These include expenses in work preparation and in production,

• in the overall quality assurance,
• in warehousing,
• in the master data and article management,
• in procurement and purchasing, and
• in logistics.

Fast success within reach - quick wins!

A plausible example from the field of special mechanical engineering.
Also read: 10 Tipps zum Geld sparen für Unternehmen

In this industry, many drawing parts are created and also purchased externally. It is not uncommon for companies to purchase the same drawing part from different suppliers at different prices. There may even be a duplicate in the warehouse, but no one knows about it because it is under a different part number. Thus, the same drawing part exists under a different article number. The employee in purchasing cannot avoid this situation at all unless he happened to have requested the same part recently. At this point,  coincidence is not a reliable help. The more complex and decentralized the organizational structure of the company, the greater the probability of this situation. In a well-organized mechanical engineering company, the proportion of duplicates is already around 1.5 to 3 percent on average. With an average inventory of 150,000 drawing articles, this corresponds to 2,250 to 4,500 drawing articles. SIMUFORM has obtained these values from the evaluation of more than 70 data cleansing projects in the past years. This friction loss can be avoided. Just by cleansing the data stock, very fast successes are achieved already in the supply chain. This is a very simple yet very effective optimization of the supply chain management because at this point immediately avoidable costs can be saved.

• Eliminate redundant storage costs.
• Eliminate redundant costs for unnecessary procurement.
• Eliminate redundant costs in item management and service.

These cost savings can be activated immediately.

But who has the time and leisure to check all the data manually? Even an existing classification, which alone can cost the user the last nerve, is of little help here.

Read now in our free brochure which advantages DATA CLEANSING with SIMILIA offers your company.

The first step: automatic capture of duplicates and duplicates for drawing parts

If you think away from conventional procedures that rely on alphanumeric designation and structuring of data, then the situation could look different today. But what requirements should a fast and simple procedure fulfill in order to keep one's own data clean?

1)The procedure must be independent of descriptive characteristics or attributes. Naming, designation, article number or other descriptions of the components and assemblies must not become the basis of a data comparison. The procedure used must rely on self-referencing characteristics that are independent of the above-mentioned attributes.

2) A fully automatic comparison of all data in batch mode, an N:(N-1) comparison, must be absolutely possible because this cannot be done manually. The effort required increases almost in quadratic proportion to the number of drawing parts. While 20,000 parts require about 400 million comparisons, 100,000 parts require an incredible 10 billion comparisons. Only then it is possible to compare all the data and ensure that there are no duplicates. This cannot even be tackled manually. Many companies start doing this and then stop after a few months because there is no end in sight.

3) The result of such a duplicate evaluation should be able to be used EDP-wise to mark the identified duplicates in one's own database or data storage or to mark them in the ERP system. In this way, duplicates can be phased out over time, marked as invalid, and referred to as the reference part.

4) It happens, of course, that geometrically identical drawing parts nevertheless, correctly, represent different articles. This is the case if, for example, different materials are assigned or a drawing part is manufactured for a specific customer in higher production quality, i.e. is assigned to a different tolerance quality. Since the material or the manufacturing tolerance is not a geometrically differentiable characteristic, a downstream technology differentiation must be carried out in a second step.

5) Once the dataset has been cleaned, it should be ensured in the future that the "clean room" created is not contaminated again. Incremental reconciliation should also be automatic and support data cleansing.

Capture the duplicates in 3 steps - at the speed of light

Step 1: Record duplicates

The identification of duplicate or multiple drawing parts is realized by SIMUFORM with a specially developed process flow - very economical. With this procedure, for example, the above-mentioned 10 billion comparisons for 100,000 drawing parts can be realized in just 4 days (approx. 32 hours). With these prospects, no company should shy away from cleaning up its database any longer, because the first step is quickly realized. Much faster than expected. The speed of light. The result is a list with all part numbers for twins and triplets in the database, enriched with all master data and metadata from the system landscape.

Step 2: Technology differentiation

The process has one drawback. The experts in the company cannot avoid a final check of the identified parts of the drawing with regard to differentiating characteristics or attributes. But this is reduced only to the remaining twins or triplets in the data stock, which is only a fragment of the entire data stock. In this way, it should be recorded subsequently which duplicates nevertheless have differentiating characteristics, e.g. material properties, etc.. But in relation to the original effort, this is child's play and can be done in a few hours of work. Predefined functions in Exel and Co. make the work much easier.

Step 3: Labeling and permanent cleanliness

The identified duplicates are marked in the leading system, e.g. ERP. Thus, the drawing part can be marked as invalid and a reference to the reference part can be set up in the ERP system. Of course, this can also be automated. In the future, the creation of duplicates should be avoided automatically, especially since one has created a cleansed data set here. There are ready-made solutions for this, such as SIMUFORM's GATEKEEPER, which can intervene at an early stage. It automatically detects duplicates and raises an alarm in time.

Conclusion

With modern and established methods such as those carried out by SIMUFORM, data cleansing or the cleansing of drawing parts is an efficient approach to detecting and permanently eliminating duplicates in the article stock. From this, short-term benefit potentials can be mobilized, which at the same time have a sustainable effect in the company. The beneficiaries of this data cleansing are distributed differently throughout the company. It can be used by users in Purchasing and procurement, warehousing, AV, or even in Design and development. Once the clean room is created, the creation of duplicates in the inventory can be controlled automatically. This data cleansing process can additionally support the introduction of new PLM or ERP systems. This can prevent existing legacy data from being transferred to the new system and keep it clean from the start. Data cleansing is a profitable business.

Find out now, in our free brochure „Data Cleansing“, about the advantages and possibilities of the potentials you can use.

Experience DATA-Cleansing.