Big Data-Projekte Nur valide Stammdaten führen zum Erfolg

Big data projects: Only valid master data leads to success

In the era of big data and advancing digitisation, almost all marketing and business decisions are based on data evaluations. To make those promising decisions, complete and reliable data proves to be indispensable. This starts with the master data, that is, the basic information on each individual company. Without this, a data-driven company cannot be successfully shaped.

The auditing and consultancy firm Lünendonk has determined in a study the status quo of master data management in German companies. Only 15 percent of the 134 companies surveyed rated their master data quality as “good”, (none rated their master data as “very good”), 72 percent rated it “mediocre” and 13 percent considered it “bad”. In addition, 40 percent of companies estimate their overall data quality to be “very good” (16 percent) or “good” (24 percent), while 60 percent of respondents see themselves as mediocre. The specialist magazine IT Daily concludes that, although mediocrity is not a bad thing, it is not enough to live up to a digital transformation.

Solutions to efficient data management and reliable master data here are not as complicated as you, the retailer, might think. We show you the causes of bad master data, its effects and the solutions to it.

Very poor decisions: The impacts of incorrect master data

Some 82 percent of the surveyed companies agree that high-quality master data can be used to perform quick analyses and make profitable decisions. Expressed in reverse, this means that master data of poor quality will actually provoke bad decisions. Digitisation projects without a high data quality are doomed to failure, say the authors of the Lünendonk study.

Although the majority of companies do have digital strategies, the automated solutions needed to perform this are imperfect, according to the study. Some 62 percent of companies are still collecting their data manually. Surprisingly, larger enterprises (69 per cent) more often enter their data by hand than the smaller ones (55 per cent). On the other hand, 70 percent of companies that classify their master data management as “bad” are entering their data manually.

The main reason for poor master data quality is data duplications, since these decisively determine the quality of the data. Some 85 per cent of the companies surveyed have no idea as to the proportion of duplicates in the individual data domains of their company. Harshly expressed, this means that not even one in seven companies can properly assess the quality of their own data.

Data duplicates: How do they arise and what consequences do they have?

Duplicates – or doublets – are different names for the same thing. In eCommerce, this means that one and the same product appears several times under different names in the shop. They can already spring up when (manually) feeding into the database: One time “Hewlett-Packard Printer” might be entered and then another time as “HP Printer”. If the printer is then offered in various designs – e.g. one with WiFi, one without, one in black, one in white – and yet still receives the same article number, this then makes identification of the actual product enormously difficult.

The worst thing that can happen to a low quality database is that the customer, as well as the product manager, will receive displayed that the HP Printer is no longer available in the shop. The customer will then switch their provider and the manager will order new goods, even though the printer itself is still in the warehouse – just entered under another term in the database.

Incorrect master data not only leads to identification problems within the company database, but also in competitor monitoring. In order to analyse the market, retailers have to match the product data of other providers with their own. This turns out to be difficult when the product information itself does not even match.

Web-based solutions clean up the database and standardise the master data

Our Business Intelligence software, blackbee, uses the process of product matching for competitor monitoring. This means that we compare the product data entered by our clients with the same data from other online shops. For this to happen at all, the product data needs to be standardised. Taking the example of the HP Printer and using blackbee, the product name can be standardised to “HP Printer”, all missing attributes can be added and incorrect entries such as “Hewlett-Packard Printer” can be deleted.

By this process, duplicates can also be quickly found and then eliminated. This ensures a clean database for the retailer and a high quality of the master data. blackbee generates a data integrity of 98.2 percent, among other things, due to the use of efficient machine learning. This creates added value for the retailer in both internal and external product matching. On the basis of this master data, retailers can then make profitable business decisions in their marketing and pricing.

Summary: Professional software solutions like blackbee work quickly and ensure a high integrity of master data

The manual transfer of master data between different applications and the elimination of duplicates cost both time and effort. If companies instead use web-based solutions, they can save on average five percent of their working time, as the Lünendonk study notes. For B2C companies, they can even increase this efficiency from five percent to ten.

The main source of poor master data quality is duplicates, which arise especially in the manual feeding in of data. Once these are in the database, they can only be removed again at great effort. The blackbee Business Intelligence software offers a simple solution here. The software helps to identify and remove duplicates by way of product matching. This provides the retailer with a clean database, with which they can then monitor price developments in the market.

The Lünendonk study has proved that German companies are not finding their digital transformation easy. It is in successful master data management that the greatest potential for profitable pricing strategies remains slumbering. Take advantage of this potential and remain a step ahead of your competitors with blackbee.

Do you need help in creating a database containing highly valid master data? Test blackbee now!