What Makes a Process Slow?

By: Stephen G. Smith, Senior Vice President & Managing Executive, Rummler Brache Group

Every organization believes there is additional speed that can be squeezed from its core processes. Are all processes then inherently slower than they should be? If so, what slows a process down?

There are many reasons in which a process will perform slower than it could. Here are some of the key reasons I have observed over a number of years that will put the brakes on a process:

Handoffs

Handoffs between departments in an organization have a significant impact on process performance. Primarily, handoffs impact processes in three ways:

	Each handoff adds ramp-up and ramp-down time to the overall cycle time of the process. The upstream person performing the process must invest time and energy into getting the work item prepared for the hand-off. This is no different than preparing an e-mail to transmit a document. The downstream performer must spend time understanding an output they receive (which they did not create) before they can add value to it.
	Handoffs can introduce error into the process due to translation problems from one performer to the next. Important information regarding the process may be lost or misinterpreted.
	Handoffs are vulnerable to loss or misplacement of work items. If the process handoffs are not well managed, things can "fall through the cracks."

Approvals & Sign-Off

Approval steps add cycle time to the process and should be avoided if at all possible. To determine the value of an approval step, consider the following:

1. Is a separate approval required by policy or regulation?

2. Are the criteria for approval quantifiable and do they yield consistent results?

3. Is the approver available at least 95% of the time the item is ready for review?

4. Is the reject (non-approval) rate greater than 5%?

The more "no" answers you have to these questions, the greater the likelihood that the approval step can be reduced or altogether eliminated.

Process Mixing or "One-Size-Fits-All"

Process mixing occurs when fundamentally different process inputs/triggers are forced through the same series of steps and controls, regardless of the true output requirements. For example, the process for purchasing office supplies should not be identical to the process for purchasing new office space. While it may be true that both transactions follow the same process model (procurement), the actual detailed workflow will differ because of the differences in the magnitude of the transaction. The purchase of new office space entails greater risk, greater expense and greater disruption in planning to receive the asset than buying office supplies. Therefore, one should expect different business rules, controls and even procurement methods for the acquisition of each item.

Re-Keying

Despite the fact that most organizations today are awash with data entry systems, manual keying of data that has already been captured elsewhere is still commonplace. In general, data should be captured at the earliest possible opportunity in the process by those that are closest to the source of data. For example, if a sales representative is the first point of contact for a new customer, then customer profile data should ideally be captured at that time, rather than transcribed to a form to be later entered into a system by someone who has no direct knowledge of the interaction.

Unnecessary Serial Processing

Parallel processing can greatly reduce process cycle time and increase its capacity. You should check to see that process steps in series are truly required to follow one another. The dependencies between upstream and downstream process steps should be clear and intuitive. If they are not, then the sequence should be questioned.

Inappropriate Batch Processing

Batch processing is often used to optimize resource utilization, especially when resources are costly or scarce and processing cost is the primary performance goal. A simple example of this principle is found in many homes. Most people do not run their dishwasher every time they dirty a glass. Instead, they wait until the dishwasher is nearly full. In total, this saves water and energy, resulting in a lower cost per item processed. However, if cycle time is at a premium, batch processing is often ill advised. This is especially true if:

	The time interval between the arrival of new transactions equals or exceeds total process cycle time
	The cost of delay is greater than the cost of sub-optimal resource utilization
	The resources in question cannot be used for anything else while they are waiting for a batch to complete

If any of these conditions hold true for your process step, you should consider whether batching is your best option. By maintaining an awareness of the things that slow a process, you can appreciably increase the speed of the processes in your organization. The key is-be vigilant.

Steve Smith can be reached at: (214) 239-8501 ssmith@rummlerbrache.com