Inventory Management and Warehouse Operations.


Multi-modal Data Collection in the Warehouse on the Shop Floor

 Bar codes, RFID, Voice, Light-directed, Touch Screens and more.

By Dave Piasecki

Back in 2003, I wrote in my book on Inventory Accuracy " I think using voice combined with other technologies (bar code, RFID, light systems) will provide the ultimate in accuracy and productivity." In more recent years, the industry has come to call this combination of ADC technologies "multi-modal data collection" because apparently it takes a cool sounding name to get some of you folks to pay attention. Just imagine the next time you're at a trade show or seminar with a group of your peers from other companies, and as some of them are bragging about their voice or RFID or bar-code systems, you get to smugly sit back and mention how you feel multi-modal is the only way to go.  Oh yeah, you're the man!

All kidding aside, combining various data collection technologies is becoming more popular—and  for good reason. And if the industry wants to call this multi-modal, well, that's ok with me.  Or is it?

In reality, multi-modal data collection is not exactly a correct description anyway. The term “data collection” implies we are simply using these technologies to collect data, but in reality we are using them to direct and execute tasks. Which is why you will also run into more specific descriptive terms like multi-modal order picking, multi-modal putaway, multi-modal manufacturing, or more general terms like multi-modal systems.

What is multimodal data collection?

Multi-modal data collection simply describes using more than one data-collection technology to accomplish a task. It can easily be argued that we have been doing multi-modal data collection for decades. Technically speaking, entering information on a keypad is a form of data collection, so in past years when warehouse workers used a combination of a keypad and bar-code scanner they were using multiple technologies to collect data. But clearly the more recent hype is not about combining a keypad with a bar-code scanner.

What has really driven the multi-modal movement is voice technology. The most likely combination being voice and bar-code, but voice and RFID, or voice and light-based systems are also becoming popular. And, of course, you don’t need to limit your multi-modal system to only two technologies  (if you did we’d call it bi-modal, but that just doesn’t sound as cool). Nor should you assume voice needs to be a part of multimodal. Light-based systems combined with bar-code technologies also work well (and have been used for years), and parcel shipping systems have been combining bar-code technology with weighing technology (and cube measuring technology) for many years.

Why multimodal?

It simply comes down to there not being a single perfect technology that can accomplish these tasks. Every technology has its advantages and shortcomings, so rather than having to accept the shortcomings of a technology because it does part of the task well, we can utilize additional technologies to accomplish these other aspects of the task.

The key to effectively utilizing multi-modal is understanding the tasks you need to accomplish and the capabilities of the various technologies available. Here’s a quick rundown of some of the advantages and disadvantages of various technologies.

Bar-code systems.

Barcodes are a great technology for capturing discrete data. You point the scanner at the barcode that has the data you need, pull the trigger, and hear a nice little beep that tells you it worked (hmm? A barcode scan with an audio beep, that’s multi-modal already). It’s very fast, very accurate, and the cost to produce barcodes is very low.  

Downsides? You need to find the barcode to scan it (requires line-of-site), can only scan one barcode at a time, workers using the technology need to carry some type of scanning device (some are more cumbersome than others), and you typically need other technologies to accomplish a task  (you need something to direct the task such as a paper document or computer screen).


Doesn’t require line-of-site and can scan multiple tags at once. Can potentially carry more data than a barcode, and data in tag can potentially be updated/changed as part of the process (depends on type of tag).

Problems with RFID include difficulty in reading tags in certain circumstances (with liquids and metals), higher costs of tags (when compared to barcodes), and difficulties with discrete scanning. The realities of the “barcodes versus RFID “ debate is there really is no debate. They function so differently that there are very few applications where they can effectively compete with each other.

Hand-held portable computers with keypads and screens.

A barcode scanner is only an input device, so in applications such as warehouse processing you also need something else to direct the workers and collect the data. Historically this has been hand-held portable computers with small keypads and small screens. These devices are fairly inexpensive today and are capable of performing a wide variety of tasks and have become fairly easy to program. They can either be used with an integrated input device (barcode scanner) or a separate one.

I’ve never really liked hand-held computers in the warehouse. Sure, you can do a lot on them, but they are very cumbersome to use in a warehouse environment. Years ago when I still had perfect vision, I didn’t like the small screens and tiny keypads. I haven’t exactly changed my opinion now that I’m older and require reading glasses. This isn’t to say I don’t see a use for these, I’ve just never been all that happy with the form factor. They’re a productivity killer when you have warehouse workers reading off of tiny screens they have to constantly scroll or change to get the information they need, and entering stuff on tiny keypads. Not to mention the fact that they are tying up one or both of their hands to carry and use the thing. However, if you can organize their tasks and the programs in a way that minimizes screen changes and keypad use (and please do not make them use a stylus), they can be effective with certain tasks.

Vehicle-mounted portable computers with keypads and screens.

When your warehouse workers are constantly on vehicles such as forklifts, order pickers, or turret trucks, the option of a larger vehicle-mounted portable computer gets around many of the problems associated with the handheld devices. You can have much larger screens (preferably touch screens) and even full-size keyboards. Properly designed touch screen applications can be very effective when used on a vehicle-mounted computer. Another nice thing about vehicle-mounted devices is they are typically powered by the vehicle so you don’t have to charge them separately.

Downside of vehicle-mounted computers? They can be a little pricey when compared to a low-end hand-held unit, and, of course, you need a vehicle.

Voice systems.

The big selling point for voice systems (speech systems) in the warehouse is the hands-free eyes-free operation.  Your warehouse workers aren’t handling a device or looking at a computer screen.  This can greatly increase productivity.  Voice systems are considered “wearable systems” since the worker does not have to handle the device as he executes the tasks. Voice is also a big plus in freezer environments where screens and keypads are problematic.

They do tend to be more expensive than non-voice systems. And while you can get high levels of accuracy with voice systems through the use of verbal check characters, it still may not be quite as accurate as being able to scan a barcode on the item itself (depends on the operation). You also need a very consistent workflow, and special circumstances like capturing a serial number or lot number can be problematic. Some voice systems can be a bit slow though, either from the processing speed of the programs themselves or through poorly designed prompts.


Light-directed is probably the fastest manually order picking technology available. Downsides include cost since you usually require a light device for each picking location (there are variations of this that reduce equipment requirements). It also works best when you can see the light next pick from the current pick. This makes it more effective for faster movers than slow movers. In addition, it can get complicated if you want more than one picker picking in the same area. And like voice, you may not get the same level of accuracy as scanning a barcode on an item and have no means of capturing additional data such as serial numbers or lot numbers.

If you use your imagination, you can easily see uses for various combinations of these technologies.  Adding a wearable ring-type barcode scanner to a voice system gives you the ability to confirm or acquire item-level details such as UPC codes, lot numbers, and serial numbers. The same can be done with a light-directed system. Or you could do a voice/light-directed combination where light-directed is used in the fast mover area and voice takes over when you leave that area. Or, in a pick-and-pass operation, RFID could be reading the totes as they  move from zone to zone, then voice or light-directed or portable computers with barcode scanners take over once in the zone.  Or, a lift truck operator can be voice directed, but also have a touch screen and barcode scanner to execute specific parts of the task.

From the worker's perspective, multimodal should be seamless. The worker does not have to "switch" between one mode and another. For example, in a simple voice/barcode ecommerce order picking combination where the items being picked have UPC codes and you want to scan each individual item to get the ultimate in accuracy, the voice system will direct the worker to the pick location and tell him how many to pick, the worker will then scan each UPC code as he picks it, then the voice system sends him to his next pick.

Will multi-modal cost more?

Probably, but it really depends what it is replacing (or what the alternatives are). Multimodal may require more equipment or more expensive equipment and more complex programming. But none of this should be assumed. Integration is becoming so much easier these days that some of this could be very simple. Then again, don’t get carried away with multi-modal. There’s a lot to be said for being practical when you approach these types of projects. Not every task will benefit from multi-modal, and even of those that do, not every task will benefit enough to justify the cost. I always recommend starting small with any data collection project. Trying to go too big too fast will typically result in mistakes being made, that is, if you ever get the project started. The biggest problem with going too big is it just never happens—either due to sticker shock or just a lack of resources.

And for God’s sake, don’t go multimodal just because you think it sounds impressive.

Also see my article on ADC Basics.

More Articles by Dave Piasecki.

Dave Piasecki, is owner/operator of Inventory Operations Consulting LLC, a consulting firm providing services related to inventory management, material handling, and warehouse operations. He has over 25 years experience in operations management and can be reached through his website (, where he maintains additional relevant information.


BookBook Banner02