Voice technology can be implemented either as a tactical tool for improving tasks, or as a strategic initiative with the potential to empower the workforce. The simplistic view is task-centric: that speech recognition is an alternative means of capturing data, in effect a humanised “talking scanner”. The strategic view is more broadly people-centric: that voice technology can optimise floor logistics.
Floor logistics describes the activities of warehouse workers. It is the sum of all the things employees do – moving about, handling products, coordinating activities, reporting on progress, and dealing with exceptions as they arise. The implications of optimising floor logistics are profound as the real pressing business issue in the warehouse is not a given task such as picking or replenishment – it is the effectiveness and contribution of the workforce as a whole. Optimising a task produces an incremental improvement; optimising the workforce has a broader and more far reaching impact.
The human factor
The workforce’s contribution in logistics operations is crucial. Labour is the single biggest cost factor in warehousing, yet it is the least leveraged by technology. Distribution centre (DC) workers have a huge impact on the ultimate bottom line – they either operate at top efficiency or waste time, make wrong decisions and other costly mistakes. Marginal improvements result from streamlining tasks, but breakthroughs come from a focus on people.
Dynamic Workforce Optimisation (DWO) is a major evolutionary step in warehousing technology. Powered by a new kind of interactive voice-based software application, it is a people-centric solution to the challenges of labour use in the warehouse. DWO, through the operation of key technology enablers, improves floor logistics and maximises the contribution of the front line workforce.
Many organisations push for high performance by striving for uniformity in all warehouse processes. Assembly line manufacturing has proven that efficiencies are improved when workers perform pre-optimised tasks repetitively. Unfortunately, this approach often falls short in the warehouse because:
Floor logistics activities are more complex than assembly line tasks.
While tasks might be uniform, the people who perform them are not.
Logistics activities rarely go exactly as planned. The day-to-day reality of exception conditions means totally uniform processes are never fully achieved.
The technological factor
This combination of complex processes, exceptions, and human workers with variable abilities adds up to a huge management challenge, and a major opportunity to leverage technology for an operational breakthrough. Voice recognition is an ideal technology vehicle to address this challenge, but it must be applied in conjunction with people-centric software that has intelligence about the entire range of activities warehouse workers do.
All of DC activities are accomplished by people. But people are not interchangeable components – they differ from one another. They are the backbone of any warehousing operation, and effectively managing them is a challenge, because they have varying:
experience and abilities.
education and motivation levels.
job roles and/or multiple job roles and sometimes varying languages or dialects.
Often, the workforce is a diverse group, perhaps speaking different languages and coming from different backgrounds.
In some parts of the world, a trend toward multilingual warehouses is well-established. For instance, in the US more than half of all warehouse workers speak Spanish as their first language, whereas most managerial personnel are native English speakers. In Europe, multiple language workplaces are common in virtually all border areas. Also, the reality of cross-border labour means a DC may operate with workers from widely differing home areas who may speak a common language but with very different accents.
Another issue is the variety of experience and abilities among the workers. Few warehouses are staffed with nothing but highly motivated “super workers”. Just like any other workforce or population, the larger the group the more likely you are to have a “bell curve” of experience and abilities (Fig 1).
In a typical warehouse operation, at one end of the curve you will have some workers who are relatively new (who presumably are less effective). At the other end of the curve are your “super workers” who out-work and out-produce everyone else. Somewhere in the middle is the bulk of the experienced workforce.
But in reality, the shape of the workforce experience curve for many companies varies with employee churn and business seasonality. Many enterprises bring on temporary workers to handle expected volume increases during the busy weeks. At those times, the experience curve is weighted more towards the left-hand side of the scale (Fig 2), and supervisory challenges are intensified because inexperienced workers naturally make more mistakes. Just when the enterprise needs especially high performance – during the busy season – what often happens is a drop in the effective contribution per worker.
In any warehouse, supervision – or span of control – is one of the key challenges. How is such a diverse workforce trained? How is it possible for a single supervisor to give all individual workers the on-the-spot assistance, correction, and direction they need?
As the size of the warehouse increases, so does the management’s challenge. The larger the warehouse, the more skilled the supervisors must be – but finding highly skilled operations managers is difficult. And, while excellent supervision is a major asset, it isn’t free – leading to the classic trade-off: too many supervisors will drive costs up while too few will result in inadequate supervision and costly workforce mistakes.
One potential benefit of voice-based DWO technology is to effectively increase the span of control apart from adding more supervisory personnel. This can be achieved because each worker has a “virtual supervisor” on his shoulder throughout every activity. While no technology can replace supervisory personnel, real-time workforce optimisation capabilities can extend the reach of the supervisors you have.
The issue here is that workforce optimisation technology must take all of the variation inherent to the “human factor” into account. The mission is to “move the curve to the right” (Fig 3) – to provide leveraging technology that benefits each employee and makes any given employee’s contribution more effective than it otherwise would be – regardless of where he or she falls on the ability/experience curve. We want temps to quickly get up to standard, experienced workers to produce like the experts, and experts to be taken to new levels of effectiveness.
If one dimension in workforce optimisation is the variation and diversity of the labour pool, a second dimension is the range of complex activities the workers perform. Marginal improvements can be realised when a new technology makes an existing task easier to perform, but breakthroughs come from “thinking outside the box” to revolutionise a business process instead of making incremental changes to it.
The introduction of voice technology into warehouse operations is a classic case in point. Some enterprises are implementing voice as a more efficient means of capturing data during picking and replenishment activities. But all this does is turn warehouse workers into talking scanners. While accuracy and productivity improvements may be sufficient to justify the investment in voice technology, an opportunity for greater benefits is not realised. Why?
Warehouse workers do more than capture data – estimates suggest that data-capture represents less than 19% of what the average warehouse worker does. Floor logistics encompasses much more. The point in time when data is captured (whether via voice or a scanner) is only a small part of the overall activity. Data capture also depends on the immediate presence of a physical label or identifier of some kind. Yet, when voice technology is being employed, the worker is “wired up and connected” throughout the entire activity – why not capitalise on this fact?
DWO is a major evolution in voice technology for the warehouse because it focuses on everything the worker does, and leverages the fact that the worker is always in constant, real-time communication. Examples of issues this more strategic approach can address are:
Situational issues. A worker is directed to a given location (Fig 4), but finds the aisle is blocked by several colleagues. DWO means the worker can inform the system and be immediately redirected.
Utilisation issues. Site management determines that it would be effective to have certain (but not all) selectors help out in Receiving during unexpectedly busy times. DWO software allows authorised workers to temporarily switch job roles (from picking to receiving), do a new activity, and then resume the work that was in progress.
Coordination issues. Workers put products into totes on a conveyor (Fig 5). Each worker has an assigned zone of lanes from which to pick. Fast workers finish tasks before their colleagues and stand idle until the conveyor moves again. DWO balances the workload, varying the number of lanes assigned to each worker with each new pick wave, so that nobody is idle.
Qualitative issues. Workers can identify why certain exception conditions occurred, such as a short pick because some products were damaged. This allows the WMS to be kept properly updated and synchronised.
Floor logistics is a complex set of orchestrated tasks. To deliver breakthrough benefits, DWO software must successfully address multiple dimensions of the floor logistics challenge: workforce diversity; process complexity and sophistication; and exception conditions.
The ultimate goal is to have each worker contributing at his or her maximum, and to reduce labour cost as a percentage of warehouse operating expense. In order to achieve these goals, there are certain technology “must haves”.
Front line workforce
DWO typically uses a wearable computer equipped with voice recognition and application software. Whatever intelligence the solution has about worker activities and dynamic optimisation is found in the software – not the hardware. While it is important that the hardware be rugged, robust, and capable of operating in demanding industrial environments characterised by high noise and extreme temperatures – the key enablers for DWO are nevertheless software-based:
Continuous one-to-one management;
Personalised performance enhancers;
High quality supervision is key in any effective warehousing operation. If cost was no object, the ideal solution would be to assign a supervisor to every worker – but supervision is too costly when supervisor-to-worker ratios dip below 1:13.
Voice technology – coupled with the right software – gives the opportunity to put a supervisor “on the shoulder” of each worker. The benefit realised is a lower cost of supervision for a given number of workers (Fig 6). Continuous one-to-one management gives workers the “perfect supervisor” who:
Speaks many languages, so can communicate with workers in their native tongues.
Is always there, throughout every activity, from start to finish.
Knows the worker’s abilities, experience and job roles, and can provide assistance that is tailored to the specific worker.
Is simultaneously working with all other workers in the, and can coordinate staff and tasks.
Is expert in the logistics processes used.
Has total visibility into all aspects of the warehousing operation as each shift progresses.
When warehouse workers use voice technology, they are always “connected” and available. What a wasted opportunity, when the software they have simply uses their voices to capture data. So much more is possible.
Continuous one-to-one management combines knowledge of the worker with knowledge of the logistics process, real-time access to the WMS, and visibility into warehouse conditions to give superlative direction and optimise what each worker can achieve. For instance:
Workers are only assigned work they are capable of performing.
Authorised workers can be directed to switch – either temporarily or permanently – from one function to another.
Work tasks are distributed to teams of workers based upon what each individual worker can accomplish, enabling the entire group to work at its maximum output.
Errors are detected via interactive dialogs and corrected at the time when it is easiest (and cheapest) to fix them.
End-of-shift optimisation is possible, whereby remaining work is assigned to workers according to their productivity rates – thus eliminating the problem of faster workers having nothing to do as the shift is ending, while slower workers are paid overtime to complete unfinished tasks.
Training and “help” are provided in-line with worker activities, enabling workers to immediately ask what they should do next without having to find a supervisor.
Continuous one-to-one management also gives site supervisors immediate visibility into logistics activities as they occur. Workers requiring personal attention can be quickly identified, thus making supervisors more effective in the use of their own time. Also, an individual can be interrupted via a special message (e.g, “Come to the Office”) without having to go out to the floor to search for the worker.
Often, productivity tools have a “lowest common denominator” quality because they don’t take an individual’s personal characteristics into account. DWO technology solves this problem with personalised performance enhancers (PPE) – software-based facilities that understand each worker’s language, abilities, and job role. Some personalised productivity enhancers:
An “expert mode” that enables more highly skilled and experienced workers to take shortcuts and work more productively. Assignment of higher work volumes to workers who have a track record of being able to accomplish more than the norm.
The ability to immediately switch from one job role to another.
Recognition that a given worker may operate in different physical environments (e.g. a very noisy freezer versus a quiet ambient area), with the ability to enhance system operation according to the actual location at any given time.
The combination of several factors makes voice technology ideal for optimising worker performance. First, it is continuously active, no matter where the worker is in the facility.
Second, it is the most immediate two-way communications vehicle available – information of all types can be sent and received at any moment (in contrast to scanning equipment that can only send information when in proximity to a label).
Third, it is potentially the most natural interface and requires no intervention to use: workers simply listen and speak while their hands and eyes are free.
Fourth, speech recognition enables workers to “multi-task” – listening to prompts and responding while they physically accomplish tasks using their hands and arms.
However, as we’ve already noted, voice recognition by itself will not optimise floor logistics unless it is integrated with DWO software. A characteristic of that software is activity-based streamlining. This is the ability of the application to react to conditions as they occur, directing and coordinating workers to accomplish tasks in the most effective way.
Three areas where activity-based streamlining can be applied are: workflow coordination; exception management; and error correction.
Floor logistics must be well-coordinated. activity-based streamlining improves the throughput achieved by groups of workers when it takes into account their individual abilities and the mapping of the processes they do. It’s like having an “industrial engineer in a box”, constantly adjusting the group’s work tasks for greater efficiency, but it does depend on logistical intelligence in the application software.
Activity-based streamlining provides a centralised WMS with a flood of up-to-the-second information. In general, the WMS will initiate process and inventory related tasks, while the voice application optimises worker activities. Tasks to be performed by co-workers such as drops, replenishment, and cycle counts can be detected and initiated more quickly than with any other approach. Another element of activity-based streamlining is the provision to management of detailed, real time visibility into floor logistics operations. Using a web interface, managers can track activities in one or more facilities from virtually any location. Real-time visibility enables management to pinpoint situations where immediate supervisory attention is needed, and thus use supervisory resources more effectively.
In the foreseeable future, labour will continue to be both a key resource and a major cost factor in logistics operations. Leading companies that strive for excellence in their operations are looking for ways to maximise the resource while minimising the cost. Voice recognition provides a real-time, interactive interface that leverages what is arguably the most natural and immediate form of communication – speaking and listening. Tactical implementations of voice emphasise picking and data capture – but they leave many opportunities untapped.
The real pay-off from voice technology comes when it is applied strategically – with a people-centric focus on floor logistics as opposed to individual tasks. When combined with PPEs, Continuous One-to-One Management, and Real Time Activity Optimisation, it can take the workforce to new levels of effectiveness.
Stephen Gerrard is vice president of marketing at Voxware. Tel: +32 50 313 666 (Brussels).