How to Prevent Mispicks in an E-Commerce Warehouse

In a previous post, we explored the impact of warehouse mispicks on e-commerce fulfillment operations. The picture was not pretty, but your business doesn’t have fall victim to frequent picking errors. Solutions range from manual prevention methods to partial or full warehouse automation.

Here, we’ll examine some of the options you could consider. So to pick up where our earlier post left off, here’s a reminder of the very first, and possibly most important step in mispick prevention—measurement.

To Manage It, You Must Measure It

It’s not difficult to set up a key performance indicator (KPI) to measure picking accuracy, and you can even use the results of such a KPI to start reducing mispicks before implementing other solutions.

Mispicks typically result from human error, so when you can show your warehouse teams how they are performing, and combine that knowledge with a little education about the consequences of mispicks, you should see some early improvements as employees take more care over picking accuracy. You can also begin to manage performance more effectively by setting goals tied to your picking-accuracy KPI.

Of course, if KPI tracking shows a mispick percentage lower than 1% of orders, you might not need to do much. Nevertheless, by tracking the metric, you will know if performance ever falls to levels that are less than acceptable—something that might be unnoticeable in the absence of KPI monitoring.

Manual Methods for Mispick Reduction

If your enterprise has an ERP platform or warehouse management system (WMS), you should probably use those solutions to improve picking accuracy. If that’s not an option, you might have to settle for some manual steps to address mispicks in your warehouse.

You could consider any of the following possibilities to reduce mispicks. Some of them will help to improve picking speed and efficiency as well.

Improve SKU labeling to make it easier for pickers to identify the right items.
Review and, if necessary, simplify the documentation used for picking.
Optimize picking routes through your warehouse to reduce picker fatigue (as fatigue invites errors).
Group items that are commonly picked together in slots adjacent to one another
Make sure items that look similar to each other, or have consecutive SKU numbers, are not kept in adjacent picking slots.
Train warehouse operatives for accuracy first, productivity second.
Assign some warehouse employees as order checkers. You can choose to check every order after picking or have checkers scrutinize a percentage of orders at random.

Try to use manual checking as a last resort only. There are two reasons for this. First, pickers can become complacent if they know checking is in place and that their errors will be corrected (although you can mitigate this with an incentive program that rewards accuracy).

Second, as mentioned in our article about the impact of mispicks, your operation incurs extra costs from the moment a picking error takes place, and the labor involved in checking after the event makes it an expensive and rather inefficient solution.

For a Higher Level of Prevention, Think Automation

It’s worth mentioning one more time—mispicks are human errors. Therefore, the more automation you introduce to your warehouse operation, the lower the likelihood of mispicks.

Automated warehouse solutions are becoming more affordable and for any operation with WMS, ERP, or both, the additional cost to automate picking management at least, should not be prohibitive.

For example, barcode scanning and pick-by-voice solutions can be added on to most systems today, and make it much harder for human pickers to make errors. Mistakes can still happen, but with automated management, multiple factors must combine for a mispick to occur, making it far less likely.

Fully Automated Picking

Fully automated solutions, such as Automated Storage and Retrieval Systems (ASRS) and warehouse robotics are expensive to implement, but as they significantly reduce the need for human pickers, they are more effective than manual or partially automated solutions.

That is not to say that full automation will get you to 100% picking accuracy, but it does remove human error from the picking equation.

Warehouse robots look set to become the least costly solution for fully automated picking, especially since the concept of Robotics as a Service (RaaS) began to pervade the warehousing domain.

RaaS enables SMEs and large enterprises alike to take advantage of robotic picking assets without making substantial capital investments, paying instead for robotic hardware, software, maintenance, and support on a throughput-only basis.

The technology is still new, but robotics vendors are already offeringRaaS warehouse solutions, and adoption is rising as operators realize that RaaS is a low-risk, high-return solution for increasing picking accuracy, and for overall warehouse cost, productivity, and efficiency improvements.

Better Technologies Are on the Way …

These emerging technologies offer promise for removing the mispick menace, but they aren’t perfect. For example, robots are currently very limited in their ability to perform piece picking, although they are quite adept at carton and pallet picking. However, this will change as robot dexterity improves.

Other emerging technologies, like computer vision, will eventually assist robots and humans to pick more quickly and accurately, and machine learning will allow robots to collaborate for greater efficiencies, as they already do in one online grocer’s automated warehouse.

… But Can You Afford to Wait?

In the prevention of mispicks, deciding how much to invest and what current or future technologies to invest in, will depend upon various factors, which include…

Current picking performance;
The value of your merchandise;
A calculation of the cost per error;
Scale of your fulfillment operation.

As a general rule, however, a picking accuracy rate of less than 96 or 97% is likely costing your company substantially, so it will be worth taking steps to minimize mispicks as soon as humanly—or robotically—possible.