These controls are related to the input screens on the data entry operators. The basic premise of these controls is that the data entry personnel OR over-the-counter front-end executives can be trained only to a limit. This is because of fast changing processes, higher man-power turnover and outsourcing of data-entry. Therefore, system managers have to take onus of making the system 'Idiot-resistant', if not 'Idiot-proof'.
The Drop Down Master Data controls for Entry of Data
To have the drop down options for master tables. Instead of asking the operators to enter, provide a drop down of the locations, states, existing customer, job codes and product codes etc.. Alternatively ensure that the entry is checked against the master tables. Many a times the location codes etc. are not important to processing, but they will be important for analytics (For example location-wise customer preferences).
Typically the items, which are important, do have a form level OR database level check. The lack of the validation (either at form level OR checking it out with database post entry) of other items (location, age, sex) will result in an inaccurate customer data. This kind of data is extremely difficult to correct at the later stage.
TIP- Try to have the codes or descriptions in the drop down from a standard as much as possible. For example, I would use the locations list from US postal address system standard. it makes your coding as well as the description easy to create and handle.
The Data field validation controls for Data Input
To ensure that the values entered in the fields are as per the organization business rules. Starting from the basic checks like having
- mandatory input
- only positive amounts for money
- maximum age being 100, to complex rules like
- not accepting less then 200 check amount for a specific class of bank account deposit.
The lack of these data field validation controls will mean invalid data getting into the system resulting in cascading downstream impact. The data entered through online inputs is the trigger for all processing.
These data input controls can be implemented in two ways. One has to have checks build in the form, so that an operator is given a message the instant invalid data is entered. This is generally done for simple rules. For complex rules, the same checks need to be placed in the database design itself, which automatically checks the conditions, when we try to append/update the records in the database.
TIP- Create a repository of business rules related to field validation and set-up a process so that all applications are adhering to those rules.
TIP- Try to follow a standard set of principles, on:
- The validation which has to be applied on a data element. For example, if age is mandatory in application, keep it mandatory in all applications. Without it, you may face issues, when you integrate the data for the sake of analysis.
- The validation which has to be applied on the front-end or on the data-base end. If one application has created a database level control, other applications may leverage on it instead of creating their front-end checks. This approach However, need not be followed hard and fast, because front end controls can increase productivity.
TIP- When you are using 3rd parties capturing your data in offline data-capture systems, one has to ensure that your masters and other validation logics in the database are synchronized on highly frequent basis.
Data Entry form alignment with the screen design
You data entry form and screen design for data-entry should be well-aligned. This means that sequence and placement of fields should be similar. An example of misalignment is when you have the customer address in the top of the data entry form and it is at the bottom end of the long data entry screen. This alignment not only raises the productivity, but also helps you to avoid data entry mistakes.
Online customer searching and matching capability for data input
This is an important control. Customer Data quality is a big issues with the organizations. The whole concept of Master Data Management has been mostly motivated by the need for maintaining a quality customer data. When we say customer data, one can include vendor, employees, agents etc. A quality system should have the capability to do an online search for checking if the new customer being added already exists in the system. This is achieved by having customer data searching and matching programs linked to your forms. As soon as you enter the 'identifier fields', the programs scan the systems databases and throw-up possible records which could be matching with the new record that you are adding.