3 ms·

Hello - Andrew Hess and I are the founders of Staffjoy. We went through the first YC Fellowship class, where we launched automated contractor scheduling. Today

Hello - Andrew Hess and I are the founders of Staffjoy. We went through the first YC Fellowship class, where we launched automated contractor scheduling. Today we are excited to launch Staffjoy Boss, our automated application for employee scheduling. At the core, we have rewritten how scheduling works so that we schedule workers when they prefer to work. We have also added features like time off requests, clock-in/clock-out, and employment law compliance we make it easy to have Staffjoy run a business of hourly workers. On the technical side, we've rewritten our algorithms to search beyond optimal staffing levels for canonical optimal solutions that make employees happy. So, if one person prefers morning and the other prefers afternoon, from an optimality perspective it doesn't matter who gets assigned to which shift - but we'll try to put workers in their preferred time slot so that they are happier and more productive. Finally, today we are also announcing our first round of funding today with participation from Semil Shah, Ray Tonsing, David Ulevitch, and others. More details are available at https://blog.staffjoy.com https://blog.staffjoy.com

e12e 11y ago
Interesting. I've been thinking about this a lot - where I currently work (I work as a night receptionist at a hotel) -- there's been a few... challenges with scheduling. I ended up setting up our current rotation that was approved by our manager -- but only for the night shift. The early/late shift still have some issues (more below). I see a lot to like (from what I can read from the docs, I've yet to register for the demo to play around). I do find it interesting that your example for scheduling has such a short window (I realize this is just an example): schedule to be ready to view 3 days before week start: https://help.staffjoy.com/docs/schedule-timing https://help.staffjoy.com/docs/schedule-timing As a colleague is fond of saying: rotas and schedules aren't for knowing when you work, but for knowing when you're off (so you can plan your free time). Does staffjoy work well with most employees on a set rotation (eg: I know when I'm off in November 2016, barring any exceptional circumstance, just as if my off-hours where every Saturday and Sunday. This is especially important when working nights -- as having consecutive days off doesn't really help much to alleviate the extra strain of working nights, if the worker cannot plan their downtime (eg: have coffee with a friend, plan to take part in a group activity etc). I do see that you're not optimized for 24/7 work yet - which is fine of course (especially since you actually document it): https://help.staffjoy.com/docs/running-a-247-business-on-staffjoy https://help.staffjoy.com/docs/running-a-247-business-on-sta... Now, the reason I bring up rotations and long-time planning, is that I strongly believe predictability correlates with reduced stress for employees (and probably managers as well -- but your "pick-up" system for eg. sick days seems to a lot to help managers already). In Norway the law for overtime is different for managers and people that are able to determine their own hours - the idea is that if you can actually decide when you work, it's a lot less of a problem to work, say, a 100 hour week, than if someone else decides when you work. Finally about day/night shifts -- does your system make an effort to prefer scheduling shift changes early to late, rather than late first, then early? That is the final issue I've seen a lot at my work-place -- with seemingly random changes from late to early. Apart from being illegal (late shift ends at 2300, early starts at 0700 -- meaning that there's not enough down-time to comply with local labour law, never mind local agreements with the union) -- it's an example of a completely unnecessary extra strain on the workers due to -- IMNHO -- poor management (and/or poor tools for shift scheduling). In the general case, I think one would find that scheduling eg: 3 early shifts in a row, then 3 late shifts would be preferable to most people, rather than say 1 late, 1 early, 2 late, 1 early, 4 late -- even for those that in general prefer working the late shift. At any rate, looks like an interesting product. I'm afraid we'll never use it at my current work place, as all purchasing decisions are centralized (for a large, international chain). But I'll make sure to mention it to friends that manage smaller businesses (that also don't operate 24/7 - so a better fit overall :-).
- philip1209 11y ago
  For the example - we set the window so slow so that trial users only had to think about one future week at a time. We're working on a new feature called "fixed shifts" so that businesses can set up similar rotations week over week. We've heard that businesses with a mix of full-time and part-time workers prefer to give full-time workers the same shifts week over week as much as possible (thus establishing a a consistent base), while using part-time workers to fill expected busy times, time off, etc. We're working on it - stay tuned! With regard to 24/7 - today's release fixes previous technical limitations on 24/7 scheduling. We're amid revising our help docs to reflect this. We agree with giving predictability where possibility - by setting your preferred working hours, we build in a lot of these schedules. However, changes happen often - holidays, vacations, sick times, hiring's ebb and flow - so changes have to happen quite often. Staffjoy enforces a configurable minimum time between shifts for roles so that workers have enough time to rest between shifts.
  - burfog 11y ago
    It's not just minimum time between shifts. People fall apart from jet lag if the shifts happen randomly. This increases workplace accidents and lowers productivity. Some places require by law that there are 11 hours between shifts. To avoid jet lag, add the requirement that 11 hours of downtime happen at the same time every day, in UTC or otherwise ignoring daylight saving time. This applies even on days when employees do not work. For example, a Monday-to-Friday daytime worker needs the 11 hour downtime scheduled on Saturday night even though he works neither Saturday nor Sunday. To move that 11-hour downtime to some other part of the day, there are two approaches. One approach is adjust by half an hour each day. The downtime becomes 11.5 hours (a half hour is added on one side or the other) and then back to 11 hours (a half hour is removed from the other side). The other approach is a brutal sudden shift, but giving several days off for recovery.
    - philip1209 11y ago
      We're always working to improve how our systems work, and when we have more users operating 24/7 I think that we will be improving the "time of day" consistency. Letting users set their preferences - e.g. "mornings" or "afternoons" - does actually create consistency. Managers can also set "working hours" for workers - which, for a 24/7 business, can turn the hours they are eligible to work into a hard constraint.
DataJunkie 11y ago
Super cool. In grad school I developed a system to schedule TAs for a new student counselor. I used simulated annealing because it was related to something I had just learned in class, and because it wasn't heuristic. It's an interesting problem because there are so many ways to define optimality. Does it mean working with the same manager? Working at the same facility? Minimizing travel between facilities? Number of shift in a week? etc. The problem can be extended with reinforcement learning with employee feedback. If there are manager preferences as well, there is a two-player problem with two perhaps competing interests. I would love to work on this problem with high school scheduling for building both a master schedule and individual student schedules. No clue where I would get data though. It's a very cool problem.
- philip1209 11y ago
  I agree! We did an interview focused more on the pure mathematics side here: http://www.aistartups.org/2016/02/interview-with-philip-thomas-ceofounder.html?m=1 http://www.aistartups.org/2016/02/interview-with-philip-thom... Class scheduling for schools is a huge issue but selling to schools is a nightmare. That means that turning it into a business would be really hard.
  - DataJunkie 11y ago
    Yeah anything with schools is a nightmare. It would be more of a project than a product for me because Zangle and MMS already have the market completely saturated, at least in California.

e12e 11y ago

Interesting. I've been thinking about this a lot - where I currently work (I work as a night receptionist at a hotel) -- there's been a few... challenges with scheduling. I ended up setting up our current rotation that was approved by our manager -- but only for the night shift. The early/late shift still have some issues (more below). I see a lot to like (from what I can read from the docs, I've yet to register for the demo to play around). I do find it interesting that your example for scheduling has such a short window (I realize this is just an example): schedule to be ready to view 3 days before week start: https://help.staffjoy.com/docs/schedule-timing https://help.staffjoy.com/docs/schedule-timing As a colleague is fond of saying: rotas and schedules aren't for knowing when you work, but for knowing when you're off (so you can plan your free time). Does staffjoy work well with most employees on a set rotation (eg: I know when I'm off in November 2016, barring any exceptional circumstance, just as if my off-hours where every Saturday and Sunday. This is especially important when working nights -- as having consecutive days off doesn't really help much to alleviate the extra strain of working nights, if the worker cannot plan their downtime (eg: have coffee with a friend, plan to take part in a group activity etc). I do see that you're not optimized for 24/7 work yet - which is fine of course (especially since you actually document it): https://help.staffjoy.com/docs/running-a-247-business-on-staffjoy https://help.staffjoy.com/docs/running-a-247-business-on-sta... Now, the reason I bring up rotations and long-time planning, is that I strongly believe predictability correlates with reduced stress for employees (and probably managers as well -- but your "pick-up" system for eg. sick days seems to a lot to help managers already). In Norway the law for overtime is different for managers and people that are able to determine their own hours - the idea is that if you can actually decide when you work, it's a lot less of a problem to work, say, a 100 hour week, than if someone else decides when you work. Finally about day/night shifts -- does your system make an effort to prefer scheduling shift changes early to late, rather than late first, then early? That is the final issue I've seen a lot at my work-place -- with seemingly random changes from late to early. Apart from being illegal (late shift ends at 2300, early starts at 0700 -- meaning that there's not enough down-time to comply with local labour law, never mind local agreements with the union) -- it's an example of a completely unnecessary extra strain on the workers due to -- IMNHO -- poor management (and/or poor tools for shift scheduling). In the general case, I think one would find that scheduling eg: 3 early shifts in a row, then 3 late shifts would be preferable to most people, rather than say 1 late, 1 early, 2 late, 1 early, 4 late -- even for those that in general prefer working the late shift. At any rate, looks like an interesting product. I'm afraid we'll never use it at my current work place, as all purchasing decisions are centralized (for a large, international chain). But I'll make sure to mention it to friends that manage smaller businesses (that also don't operate 24/7 - so a better fit overall :-).

philip1209 11y ago
For the example - we set the window so slow so that trial users only had to think about one future week at a time. We're working on a new feature called "fixed shifts" so that businesses can set up similar rotations week over week. We've heard that businesses with a mix of full-time and part-time workers prefer to give full-time workers the same shifts week over week as much as possible (thus establishing a a consistent base), while using part-time workers to fill expected busy times, time off, etc. We're working on it - stay tuned! With regard to 24/7 - today's release fixes previous technical limitations on 24/7 scheduling. We're amid revising our help docs to reflect this. We agree with giving predictability where possibility - by setting your preferred working hours, we build in a lot of these schedules. However, changes happen often - holidays, vacations, sick times, hiring's ebb and flow - so changes have to happen quite often. Staffjoy enforces a configurable minimum time between shifts for roles so that workers have enough time to rest between shifts.
- burfog 11y ago
  It's not just minimum time between shifts. People fall apart from jet lag if the shifts happen randomly. This increases workplace accidents and lowers productivity. Some places require by law that there are 11 hours between shifts. To avoid jet lag, add the requirement that 11 hours of downtime happen at the same time every day, in UTC or otherwise ignoring daylight saving time. This applies even on days when employees do not work. For example, a Monday-to-Friday daytime worker needs the 11 hour downtime scheduled on Saturday night even though he works neither Saturday nor Sunday. To move that 11-hour downtime to some other part of the day, there are two approaches. One approach is adjust by half an hour each day. The downtime becomes 11.5 hours (a half hour is added on one side or the other) and then back to 11 hours (a half hour is removed from the other side). The other approach is a brutal sudden shift, but giving several days off for recovery.
  - philip1209 11y ago
    We're always working to improve how our systems work, and when we have more users operating 24/7 I think that we will be improving the "time of day" consistency. Letting users set their preferences - e.g. "mornings" or "afternoons" - does actually create consistency. Managers can also set "working hours" for workers - which, for a 24/7 business, can turn the hours they are eligible to work into a hard constraint.

DataJunkie 11y ago

Super cool. In grad school I developed a system to schedule TAs for a new student counselor. I used simulated annealing because it was related to something I had just learned in class, and because it wasn't heuristic. It's an interesting problem because there are so many ways to define optimality. Does it mean working with the same manager? Working at the same facility? Minimizing travel between facilities? Number of shift in a week? etc. The problem can be extended with reinforcement learning with employee feedback. If there are manager preferences as well, there is a two-player problem with two perhaps competing interests. I would love to work on this problem with high school scheduling for building both a master schedule and individual student schedules. No clue where I would get data though. It's a very cool problem.

philip1209 11y ago
I agree! We did an interview focused more on the pure mathematics side here: http://www.aistartups.org/2016/02/interview-with-philip-thomas-ceofounder.html?m=1 http://www.aistartups.org/2016/02/interview-with-philip-thom... Class scheduling for schools is a huge issue but selling to schools is a nightmare. That means that turning it into a business would be really hard.
- DataJunkie 11y ago
  Yeah anything with schools is a nightmare. It would be more of a project than a product for me because Zangle and MMS already have the market completely saturated, at least in California.