5.2 Assessing Impact

Each issue that has been created needs to be studied to determine the impact it has already had or potentially may have on the corporation. Somewhere along the line someone will ask the "so what" question about an issue. It is important to justify development and disruptive efforts to deploy corrective actions. It is important to document the value returned to the corporation for the time and cost spent pursuing issues.

This needs to be updated from time to time. It is usually impossible to compute the costs and benefits up front. One approach is to look at the facts and theorize on possible impacts. A brainstorming session with data analysts, business analysts, and others may be helpful. This will lead to activities to prove that the impacts have already occurred. Because impacts have not occurred does not mean they will not in the future. As the issues are worked through the entire process, additional information about impacts may become apparent. These need to be added to the impact section.

Impacts Already Happening

The impacts may not be obvious to anyone but may be very real. For example, an issue that states that suppliers exist in the supplier's database multiple times may lead to speculation that you are not getting large enough discounts for volumes purchased over a year. Investigation may uncover that this is true (one department orders under one supplier ID and another department uses a second supplier ID for the same supplier). You can easily compute the discount difference, the volume of purchases made, and the value lost to the corporation. The cost of this type of inaccuracy is totally hidden until the issue is identified and pursued.

Sometimes an issue is created from an outside-in investigation and the cost is already known. Tying the external cost to facts is part of issue definition. For example, the external manifestation might be that the accounts receivable department spends x amount of people time per month correcting wrong information on invoices. The facts are the number of blank or inaccurate values found during data profiling. The facts back up the assertion that invoices are not being prepared properly.

Further investigation may reveal that not only is time being wasted but that payments are being delayed by a certain amount for two reasons: one is the lag in time in getting invoices out, and the other is that invoices sent out without corrections get rejected by the purchasing company, causing yet further delays. In fact, there may be a group of invoices that are never collected due to data errors on the invoices. This is an example of a single visible cost leading to facts about inaccuracies, which lead to the discovery of more hidden costs.

One point to consider is that a significant accuracy problem on a data element may indicate a bigger quality problem. In the case of the missing supplier ID, it is clear that if 30% of the values are missing, there is a real possibility that the process is flawed and that the supplier ID is not available at the time the data is entered. It is unlikely that data entry staff are that bad at their jobs. It is also clear that this field is not involved in making the purchase or subsequent payments (it appears to cause no harm). The harm is all done in the secondary uses of the data. It is easy to speculate that if the data is not available at entry, data entry staff may also be entering wrong but valid values. The problem may be much larger than it first appears.

This is why you need to match inaccuracy facts to known manifestations. By seeing the actual data values in error and the data elements containing errors, you can often speculate about hidden costs that may be occurring.

Impacts Not Yet Happening

The most dangerous impacts are those that have not yet occurred. Seeing the presence of inaccurate data can sometimes lead to speculation about problems that could occur. These can have greater impact than those that occur on a regular basis but cost little to correct.

A simple example is the inaccurate birth dates of employees. There may have been no costs that have occurred yet for a new company that hires mostly young people. However, as this population ages, all sorts of government regulations about reporting, pension programs, and changing medical benefits when an employee reaches age 65 are at risk of occurring. These errors can also make decisions about hiring practices inaccurate and lead to wasteful efforts to adjust the company's mix of ages.

A business rule may require that a fast mode of shipment be used to ship certain materials that have the potential to spoil or decay. They may require refrigeration or avoidance of temperatures above a certain number. It may be that errors in the orders have caused a number of shipments to be made that violate the rule and no dire consequences have occurred. All values are valid individually, but the shipment mode rule for the product type is violated. By speculating on the potential for costs, the issues team may speculate about returned orders, merchandise that cannot be resold, and lost customers. However, that speculation may lead to the potential for real lawsuits, as the corporation may be liable for damage done to the purchaser trying to use spoiled merchandise.

This example may have been saving the company money (lower shipping costs) but creating a potential liability (lawsuits) that could severely damage or even destroy the company. This is why speculation on potential impacts is so important.

The process of assessing impacts will crystallize issues. It may result in issues being broken apart or issues being combined. As participants gain more experience, they will be better at sniffing out impacts both real and potential. As new participants join the process, they can benefit from the documentation of previous issues as a training device.

It should also be apparent that the documentation of the impacts of issues is highly sensitive information. The issues management process should provide for a high degree of privacy and safety of the information.