Save time and money with good data
A few years ago, I read an article in the Harvard Business Review that begins with a phenomenon that I often encounter in practice. Namely, that data is verified and corrected before use by the relevant department but is not updated in the source system that is being worked with.
As shown in the article mentioned above and also in later studies, employees spend half their time looking for data, verifying data and restoring data.
Underlying causes are processes that are not (clearly) defined, responsibilities that are not clear and insufficient knowledge of which data is important for processes later in the chain.
An example of data quality
A good test to see how the data quality is doing is to name 10 to 15 essential fields and then look at the last 100 created records. For example, products.
In those 100 records you indicate which clear errors there are. For example, empty fields, sizes that have clearly been reversed, typing errors in e-mail addresses, etc.
Then add up all the records that don't have errors in them and you've got your percentage of records that are created correctly. Be strict with yourself, so don't count records correctly just to get a higher score!
What is an acceptable margin of error?
My opinion is that at least 95% of the records should be created correctly. Depending on the industry, this percentage can be higher (think healthcare for example). This percentage reflects the extra work that must be done to correct the errors. Depending on where in the chain the errors are discovered, more or less departments/people spend (extra) time on it and the chance of wrong decisions, the risk of angry customers and extra costs are higher.
We now have the number. But what is the cost of an incorrectly created record?
Some say 10 times as much because of extra work, others calculate it in time with a standard cost of € 100,- per hour. To keep it simple I choose the first option here:
Each correctly created record costs, let's say, € 1,-. If all 100 records are created correctly, that is €100. A wrong record costs 10 times as much, so €10.
Let us assume that 89 records are correct, then the costs are (89 x €1,- + 11 x €10,- =) € 199,-.
In this calculation only the extra work is included. Not yet the costs of loss of customers, bad decisions or damage to the company's reputation.
The above example makes it clear that good, correct data saves costs. But how do you improve data quality?
First of all, make sure that it is clear what the processes are, who is responsible for what actions, which data is crucial and which is nice-to-have.
Communication is the most important factor here, not technology.
Technology supports and in itself says nothing about data quality.
Defining responsibilities
With the establishment of responsibilities, an often understudied factor is the responsibility for data quality. There are three options:
A centralised system with a master data person/department that monitors the processes from that central position and forms a central point of entry;
a decentralised system where the heads of department or branches themselves are responsible for data quality;
A mix of the top two where a central department monitors the processes and sets standards and the implementation is delegated to departments/branches.
When choosing a central system or a mix of a central and decentralized system, it is advisable to take into account who the stakeholders/owners are in the process and good data quality.
By placing the responsibility with the department(s) where the data is created and used, there is a direct interest. The ICT department can support the person(s) responsible in making the lack of quality transparent.
Establishing the processes and responsibilities will prevent the input of incorrect or incomplete data. The next step is to clean up the 'old' data and further optimize the agreed procedures and data quality.
This is never 'finished', but an ongoing process that leads to the professionalisation of the organisation. Now that a lot of attention is currently being paid to the GDPR, this is also the perfect time to start improving data quality.
Would you like to start? Then contact DBHeroes to discuss the possibilities.
For questions or advice please contact Lotte van Lith.
lotte.vanlith@dbheroes.eu
Telephone 088 888 6060
Or contact us via the contact form.
[ back ]