Building on the previous blog post about Business Process Automation ("BPA"), let's now explore some ideas on how to identify processes that could (or should) be automated, and also how to evaluate the potential benefits of doing so. This is important, because there will be at least some level of cost associated with any BPA implementation, even if it's just an initial investment of time.
Given that we're talking automation, it stands to reason that the starting point in identifying processes that could benefit from automation would be examining manual processes. This extends beyond simply looking at activities like manually keying (or re-keying data), and includes processes like manually walking pieces of paper from the office to the warehouse, and stockpiling transactions (even on a computer screen) for a person's review before processing. It's important to note that some of these steps are necessary - you would probably want to retain human review of orders held for credit problems, for example - but in many instances these bottlenecks can be alleviated with thoughtful implementation of BPA.
One of the most useful tools I've found for kicking off an analysis is to get employees to make two lists:
- Repetitive activities that require very little thought process or decision making, but that have to be performed regularly
- Regular activities that are time consuming and involve some type of manual data entry
In each case, I ask them to estimate the time they spend per item on the list, either daily or monthly. Those will provide a good and objective starting point, and we investigate further from there, placing a little more emphasis on the more frequent and more time-consuming items. It is important to note, however, that one person's list can impact another's. If Joe has a task that doesn't take much time - perhaps just a few minutes daily, but until he completes that task, it prevents Sue from performing her tasks, that's a bottleneck.
Problem:
Here is an example of a bottleneck that was identified, and resolved, in a simple Business Process Automation project.
A distribution company was moving from manually, individually generating and emailing invoices in PDF format, to automatically emailing PDF invoices right after shipping. In the existing manual process, the last step before printing invoices was a review of the shipped sales orders to ensure that (a) shipping charges had been updated and (b) tracking numbers had been inserted. The controller was adamant that the same checks would need to be performed (by him) before invoices could be emailed. This would result in almost no time and effort being saved.
Solution:
The solution was to automate not just the emailing of invoices, but also the checking process. Instead of order status going directly from "Shipped" to "Invoiced", an additional status "MissingFields" was inserted in between these. Now, as soon as an order is marked as "Shipped", an automated process checks the order to ensure that the shipping charge and tracking number fields are populated, and if not, it flips the status to "MissingFields". The controller has a screen showing just these orders with the missing information, and he can insert the missing data and manually flip the status to "Shipped" for just one or both of those items - typically fewer than 5% of the orders on a daily basis. So now, over 95% of the orders shipped result in automatic invoicing with no stops or human intervention.
The next question is, how do we evaluate the benefits against the costs? In the next post on this subject, we'll explore some options for doing just that.
