Our Approach


“How does your software approach filling shifts, especially in dynamic (unpredictable) staffing situations?”

This seemingly simple question turns out to be anything but when asked to product owners at Workforce Management Systems (WMS), the entities in theory responsible for filling shifts.

The reason has been covered in detail here, but the short answer is that WMS traditionally have not “filled shifts,” at least by the definition we use at HireHand: “to proactively attempt to place workers into shifts.” The reason is that they have limited themselves to static (predictable) staffing situations in which the same person does the same shift each week. I don’t blame them. Filling shifts in dynamic staffing situations is hugely complex.

However, it is this abdication of responsibility by WMS that shift-filling has been largely the domain of staffing managers who have had to work against – rather than with – their staffing systems to handle the unexpected. Some of these managers have reached “Grandmaster” levels of shift filling that we at HIreHand have attempted to emulate.

This note walks through four approaches to shift filling: (1) the Manual; (2) the Bank; (3) the On-Demand; and (4) ours at HireHand.

The aim is to show how only Hirehand is an effective dynamic staffing software at scale – responsive and intelligent in real-time.

  1. The Challenge

    Companies have to be responsive when the unexpected happens. Being responsive requires companies to have the flexibility within their operating model to make decisions in real-time amidst a shifting operational environment.

    Take a typical dynamic example such as a front-line worker having to drop out of a shift two hours before its start due to illness. In order to successfully fill that shift with a suitable replacement, a host of actions have to take place. Just to name the main ones:

    1. The ill worker has to notify the company that he/she will not be working
    2. The company has to remove the worker from the shift
    3. The company has to determine a list of suitable replacements for that worker
    4. The company has to determine in which order to start working through a replacement list
    5. The company has to work through that list until a suitable replacement is found
    6. Upon finding a replacement, the company has to put that person into the shift for payroll purposes.

    It is important to understand that unpredictability (and the required responsiveness) is not just limited to the initial event of the scheduled worker dropping the shift. The availability and responsiveness of all of the suitable replacements will also be in a state of flux, thus increasing the complexity of understanding who is likely to take the shift and then actually filling it effectively.

  2. The only way most companies can be intelligent and responsive in real-time to fill shifts in dynamic staffing situations is through manual interventions (The GrandMaster)

    The most common way people-intensive companies fill dynamic shifts is through site-managers using remarkable powers of understanding to find a replacement. I’ve now worked in or across six sectors, and every people-based business handles dynamic shifts-filling the same way. At first word of a dropped shifts (e.g.,c 1.b.1. above), a site manager will typically kick into overdrive.

    In terms of understanding, these site managers are remarkable people who, similar to Grandmaster “Chess” players, can be intelligent and responsive in realtime to unexpected events given a wealth of experience and a huge amount of understanding. These individuals hold huge amounts of information in their head to determine the individuals on their call list that will both be the best fit and most likely to agree to fill the shift. This information will range from the basic (e.g., which individuals with the requisite skill sets) to the advanced (e.g., which individuals are likely to be responsive in the mornings vs the evenings). This information will also take into account dependencies, i.e., if one individual picks up a shift today that will potentially put at risk subsequent risks in the week. In other words, a true Grandmaster will understand how a short-term fill could lead to a “checkmate in six” situation – a car-crash later in the week.

  3. WMS have had to do better of late given the growing proportion of dynamic shifts but has developed an approach that is neither intelligent nor responsive in real-time.

    Dynamic shifts are now far more prevalent than the odd dropped shift here or there. As a result WMS have had to get more involved initially in managing the growing temp spend that resulted and now in actually providing a bank function to fill these shifts. (See Why are Workforce Management Systems Bad at Filling Shifts Section 2.d.i, for more information) The problem is that the mechanism by which they actually go about filling shifts remains incredibly poor because, inexplicably, it seems to have been modeled on how temp agencies fill shifts internally.

    The typical approach in the case of a dropped shift appears to be to post the newly available shift within the shift system, alert the entire workforce of its availability, rely on a worker logging in to the system to accept the shift, and then rely on a manager to approve the worker. This approach is not intelligent because there is no thought put into whom to ask and in what order (i.e., Steps iii. and iv. above). And the approach is not particularly responsive because, especially on short notice, passively posting a shift and hoping that any worker – much less a suitable one – decides to pick up the shift (Step v. above).

    As should be clear, this “bank” approach is far inferior to the manual Grandmaster approach. And the kicker is that this approach is not automated thus meaning users aren’t at least gaining ease in exchange for lesser quality. Indeed, anecdotal reports suggest that this bank software has actually increased temp spend in dynamic staffing situations. Previous “Grandmasters” who would have found an internal fill are not longer empowered to do so.

  4. The on-demand economy offers the promise of responsive filling in real-time. It is not intelligent, however.

    In Why are Workforce Management Systems Bad, I suggested that the on-demand economy is a fruitful environment to look to in order to fill shifts in dynamic staffing environments. The reason is that the short notice and short shift lengths required have necessitated the development of a powerful shift-filling mechanism. Take Uber, the most notorious on-demand company. Once a user requests a ride, Uber’s “shift-filling” mechanism immediately notifies all drivers in a fixed geographic area of the opportunity. The first driver to accept this opportunity is awarded the “shift.” This approach is automated and incredibly responsive at scale and frankly superior to the existing bank solution (all steps save v.) above.

    It is not intelligent, however. The reason is that the majority of companies in the on-demand economy do not need to be. These companies operate their workforce in a fungible manner, i.e., all parties are considered interchangeable – you don’t care who the specific Uber driver is that takes you from point A to B. As a result, on demand economy companies have not invested time into mechanisms to ensure the best possible worker (however defined) is placed into a shift. The result, for instance, is that the order by which you might offer the shift to individuals (Step 5 from the grandmaster approach) is not even considered.

  5. HireHand is the only shift filling technology that is not only responsive in real-time but also intelligent.

    HireHand is unique in that it is from the on demand economy stable but from its beginnings was deeply passionate about intelligence – ensuring not only that the shift was filled but that it was done by the right person. The initial motivation was with the worker (not the company) in mind given the company’s roots as a social enterprise helping long-term unemployed people back into work. We were deeply aware of two things:

    1. That specific people needed a specific number of hours each week to make ends meet
    2. That specific people thrived in some work environments but not in others through no fault of their own

    Both of those points argued against the pure fungible and intangible model operated by on demand companies. And in the past we have described ourselves as an intelligent on demand platform based on our 40 factor matching algorithm and invite sequencing (See more on this here).

  6. The actual reason is the “Grandmaster” approach that we adopted to shift filling for our first 3 years

    The more salient part of our origin story is that we adopted a “Grandmaster” approach to shift-filling for the first 3 years of the business. Scott Erwin, the Founder, matched every shift on the platform personally for the first year with the same care and attention (if not more) that the aforementioned staffing. And then, in a typical apprentice fashion, he taught a series of operations managers to match with the same skill that he used. From the second year onward, the software engineering team began building initially a 20-factor rank order algorithm to determine a suitable rank order for each shift. This algorithm (and its relative weightings) was debated passionately for over a year and continues to be updated to this day. The next phase was to determine the appropriate sequencing for invitations (i.e., not only who but when to ask). And over the course of the next year (and to this day) we debate at what pace to extend invitations to give each worker individual time to consider a shift while ensuring it gets filled.

  7. Conclusion

    The proof of our success is in the pudding – a fully automated, scalable platform filling 97% of shifts on the platform with record levels of business and worker satisfaction.

    The actual secret to our success is the nuance we bring to the shift-filling process which far exceeds any other technological competitor at scale, and at a localised level, can repeatedly defeat Grandmasters at their own game (i.e. filling shifts in their own companies).

    We’re certainly not perfect but we are far better than the competition because as opposed to applying a static approach to filing shifts unsuited for a dynamic situation (like the majority of WMS), we have taken the time to learn at the feet of the masters.


1. Filling shifts in dynamic staffing situations is complex.

a. Dynamic staffing situations are more complex – especially compared to static situations – due to their unpredictability

Dynamic staffing situations are where either shift patterns fluctuate unpredictably from week to week, or the people filing shifts fluctuate unpredictably from week to week. Alternatively, static staffing situations are those in which shift patterns are the same each week, as are the people filling those shifts.

The complexity in filling shifts in dynamic staffing situations comes from unpredictability. This unpredictability is driven by unexpected changes in business activity (e.g., customer demand), and, by unexpected changes in people’s availability (e.g., last-minute sickness).

b. Unpredictability in the shift-filling process requires companies to have flexibility built into their operating models (responsive and intelligent in real-time)

d. The reason that companies have traditionally had to rely on the Grandmaster approach in dynamic staffing situations is their existing shift-filling systems (i.e., Workforce Management Systems) did nothing to help.

Most people-based businesses have typically used Workforce Management Systems (WMS) for scheduling purposes. These systems were almost exclusively oriented towards sorting predictable shift patterns well in advance with full-time employees. The reason is the main problem they were built to solve – Kronos in particular – was tracking time and attendance amongst full-time employees with an additional focus on keeping all HR-related information in one place and ensuring compliance with local labour laws (e.g., working time directives). In this context, the typical way in which WMS would deal with a dropped shift would be to be more concerned with logging the sickness in an absence management module (See Why are Workforce Management Systems Bad at Filling Shifts Section 2.d.i, for more information.) To be clear, WMS provided no help to a “Grandmaster” actually go about filling the dropped shift – at best it would log the full-time employee who dropped (1.b.i. above) and then the person found to replace (1.b.vi above).