#49 The real cost of investing in data quality?
Many companies invest a lot of money in finding resources for data governance, data lifecycle, and data operation management, they do holistic data management, and they produce better data quality. But what aspects of data quality should be specifically invested in and what does it cost? Let’s separate the idea of data quality from the rest of the data management and lifecycle, just talk about what kind of cost effects the use of data has in its various scenarious. Let’s start with the idea that how many times have you sat in a meeting looking at a report or discussing data analysis can you trust this visible data and figures? All of us.
Let us first take a situation where data is collected from different sources and the company that does so will at this stage invest in data quality control (such as accuracy, completeness, consistency, validity, timeliness, security, accessibility, trust, regulations, ownership, standards, etc.) and check " all the data to be collected, classifies it on the basis of quality arguments by automation, models it into a suitable data model and then uses only the classified data in its various functions. This is a big job and it pays off. It is said that it costs even X euros.
Another company collects data and stores it. The stored data is then started to be inspected, categorized, and considered how the previously mentioned factors are utilized at this stage. There are a lot of tools and solutions for this job that everyone pays for. In addition, this is largely manual work and requires a whole lot more attention with ever-increasing masses of data. However, this is a fully-fledged task and many companies are at this stage when data quality has been revived in recent years, when large amounts of data already exist without quality control. Usually the cost at this stage is 10x euros compared to the previous company.
The third company, on the other hand, collects all the necessary data in the data warehouse, assuming that all the data is always in order, with no exceptions in the data streams, system versions, or human actions. At this point, this company has spent zero euros on data quality assurance, but instead this is exactly the company that in meetings thinks about whether the available data can be trusted and then starts looking at individual data sets and their quality manually one by one. This leads to a constant spiral where no one dares to find data in principle, data quality is not managed holistically and the real cost is almost 100% manual work. Compared to the first example, up to 100x the cost when it comes to checking the quality of all the data. And why not talk if the intent is that all the data a company has is valuable and is utilized in data-based decision-making and the data economy.
The fourth and the most miserable example is a situation where a company has ended up making business decisions based on data and made bad decisions because of bad data quality. Whether it is said that the wrong place has been optimized in the optimization of the production plant, according to the market research, a new office has been set up in the wrong place or the shopping experience of one hundred thousand customers is being developed in the wrong direction. In these situations, it is clear that there can be no equal estimate of the extent of the damage and the costs it will bring, but it can be agreed that in these the actual cost is always higher than the costs of any other example above.
Which of these examples do you want to be in? As is known, high quality and reliable data has the greatest value but it can also be built in many different ways. Data Product Business is a messenger of high quality data and we will be happy to consult on the current situation of your company, build a roadmap ahead of traffic and only through that will open the door to the data economy (hopefully no one will buy poor quality data).