Showing posts with label throughput. Show all posts
Showing posts with label throughput. Show all posts

Monday, April 25, 2011

TOC hiring logic from Bill Rhind

Bill Rhind (see profile on linkedIn) of P3 Consulting has kindly prepared this overview of utilizing the Theory of Constraints logic in the employee recruitment process. I found it very clear if you think of a constrained resource group, but Bill pointed out to me that this logic stands for any hiring, not just within a constrained group.  In fact, he feels the real value is in the non-constraint since there are many more non-constrained resources than constrained resources and one often doesn't think of “exploiting” a non-constraint. To top it all off he even included a small and very thought provoking case review.

Thanks Bill!

"What intrigues me is that many HR decisions are still made utilizing "local" thinking. Each hiring decision is generally made in isolation… on an individual, functional, or divisional case by case based and on the merit of each request. In addition, the decision to hire (elevating the system) is often done prior to identifying the constraint, to exploiting the constraint or subordinating to the constraint.

Why do I believe this is generally the case within most organizations? Organizations which I have been part of or visited (and have not implemented TOC) have a FULL list of personnel requests with all functional managers stating that theirs is the most important, and the managers do not have a clear way to accept or reject them. Managers don’t have a value based way to determine the priority or explain why one group was able to hire while another may not (or they hire based upon “local” efficiencies and don’t realize they are actually hurting the system’s overall performance).

I would like to inject that hiring must be considered holistically. Looking at all requests in an organization and prioritizing them based on T, I and OE (T, or Throughput, is the money or goal units generated by the system, I, or Inventory, is the money invested in the system in order to generate T and OE, or Operating Expenses, is the money the system is working on in order to transform in to T) from a systems perspective. There truly can only be one (or very few) most important hire.

In order to tackle this question, I suggest breaking the answer into three parts

1. decision on hiring
2. who (which skill) should ultimately be hired
3. the hiring process

Decision on Hiring

One necessary condition in order to know what is the most important hire is the company knows their system’s constraint (the only true way to know the impact on T). If the constraint for a system is not known, the best one can hope for is impacting delta I and delta OE (which are also often from a local perspective). Of course there may be situations where this is not the case (e.g., a key skill is missing or lost, and without it, even if the position is a non-constraint, T goes to “0”).

I think a key faulty assumption made or a missing process step in hiring is the assumption that a company truly needs another person within the constrained resource group to add capacity. A step that I believe is fundamentally missing in most hiring processes is the step to ensure that we have fully exploited the current capacity of the skill prior to moving to the elevate step. I can think of a few observations/checks which can be made to test the current state:

- What is the level of bad multitasking occurring within the group?
- What percentage of the constrained group’s capacity is the constraint (or CCR) performing the task which only they can perform (e.g., have they off loaded Herbie)?
- Are they working on the most valuable work to the business?

Prior to even considering the hiring of any additional personnel (and adding OE), I would want to ensure that we have, at a minimum, fully exploited the current capacity (it is important to make sure that the measure is not activity based but truly value based, focused on measuring only the amount of capacity where the constraint is doing only the specific work that the resource/skill can provide) and subordinated all else to ensure that the company was not wasting any of it. As all of you know, these steps alone can bring significant increase of flow (25%...50%...or more) within an organization while having a very positive impact on all of the people.

Who (Which Skill) Should Ultimately be Hired

If it is determined that the currently available constrained resources have capacity which is being stolen doing “other” work (work which does not add value from the customers perspective), then I would suggest hiring a person to remove this more general work from constrained resources versus hiring another constrained resource. The benefit of this is many fold:

1. The hired person would cost less OE (I know this is an assumption, but I believe a pretty good one)
2. It would have a much bigger impact on T (and, therefore, NP)
3. Constrained resource generally happier and more satisfied in their work

Take for example a sales group. We have seen instances where a company was interested in increasing throughput and was considering adding additional sales people. It was obvious that the constraint was the market. In which I really mean, the company’s ability to sell more to the market (the market by itself was not the constraint since the company only had about 10% of the market). To check whether hiring sales people was the right decision, we analyzed the sales process and were looking for overall effectiveness. The key here is how does one define effectiveness.

It became quickly apparent that the sales people were always busy (and one might say efficient); many working non-stop for 50-60 hours per week. Their days were filled with meetings (both internal and external), report writing, following up on sales orders, logging their sales calls, obtaining marketing materials, driving between calls, trainings and, of course, sales calls. Since sales generally are going to be made during sales calls, we asked how much of the day is spent with customers. The general answer for this client (and many others) was 25-50%. As we peel back the activities within the call itself, we asked one more question: “how much time are you spending with the decision makers?” The answer was enlightening, only 5-10%.
Would adding another sales person have an impact…of course, but, in this case, adding a person who’s responsibility would be to provide support to the entire sales organization would be much bigger. In this case, adding one support person provided for an absolute increase of 15% to EACH sales person. This is the equivalent to a doubling or tripling of the SALES FORCE. While I know that there are other possible levers for improving the T, no one of them could deliver more impact on T (while minimizing impact on OE) in the time it took to implement this improvement.

The Hiring Process


A key improvement is to consider the multitude of hiring’s as a multi-project environment. Ensuring that the hiring projects are being released in such a way as to ensure bad multitasking is minimized. Ensuring that Full Kit (strong job description is developed, availability of resources, clear end date, etc.) is fully developed prior to starting the recruiting process. Ensure that the hires are prioritized and released into the hiring process based upon their impact on T, I and OE.

I think by breaking the solution down, a company can now focus and improve the process to deliver value to the business at a global versus a local perspective. The organization can now make much better and quicker decisions on hiring and the hiring will happen at a much faster rate."

Monday, April 4, 2011

Flow to the extreme


A simulation clip for a train designed to never stop I fisrt saw over at the Goldratt House.
First off, I think it is important to think about the paradingm shift this represents. All the trains I have ever encountered work on the same paradigm - the train waits for people to board/unboard. This is, to the best of my knowledge, a main reason for train schedule flactuations. A Londoner once told me that he could catch the underground  at 7:45 and be at work within 10 minutes, or he could be a bit late and catch it at 8:00 and then be at work within 30 minutes. So it is pretty clear what is the power of this new technology and it is probably good practice to try and work through the other questions of new technology. But what keeps troubling me is - how does this thing turn around to start the trip back if it never stops?

In manufacturing we seek flow because it means inventories are moving and as inventories move in one direction they create a move of income (or rather Throughput) in the other direction.Jusr remember flow is only a means to a goal and must be viewed from the greater perspective of the entire system (here's why)

Tuesday, March 29, 2011

TOC for hiring part 2 - Sorting through the CVs

From a TOC point of view handling the classical hiring process (advertise opening, get CVS, sort, interview, hire) is entirely an operating expense (OE), it does not create any throughput, even though the end goal of the process is, indeed, an increased throughput. Then at the end of the process what you get is "marriage after a blind date", with neither sides really sure if the right decision is being made. Taking that into consideration – perhaps there is place for paradigm shift in hiring.
But assuming we have to use the current process – how can we minimize the impact it has on out throughput? As always – it is done step by step.
So, once a company announces an opening it will receive candidate CVs and most of the times it will receive a lot of CVs, much more than needed to create a meaningful selection. We can divide these CVs into three basic groups:
  • The "No-s" – these CVs present candidates that are clearly not relevant when compared to the defined characteristics for a good enough candidate (as discussed here)
  • The "Yes-s" – CVs that represent candidates with good fit to the defined characteristics
  • The "Maybe-s" – CVs that are not clear cut, perhaps there is great fit in some of the defined characteristics but not in others, perhaps something else is creating doubt
How to treat the "No-s" and the "Yes-s" is clear, assuming there are enough CVs that are not "No-s". So the only problem is the "Maybe-s". This should not be a problem, if we look at the situation from a satisficer point of view. First these should only be considered after the "Yes-s" have been exhausted.
Once it is clear that there is need to go into the "Maybe-s" pool, the first step would be to re-analyze them. There is more data and knowledge in the system now, so perhaps some of these are actually "Yes-s" or "No-s" and should be treated accordingly. Then process them in order from the most promising to the least promising. Stopping as soon as possible (when the position has been manned, there are enough eligible candidates to move to next phase etc.)
One point to keep in mind is that often times bias leads to wrong decisions in hiring. This does not have to be the "regular suspects" of bias such as gender, race, religion and so on. There can be many hidden assumptions about the position, the appropriate candidate and so on that will lead to bias.

Monday, March 28, 2011

Using TOC for a hiring process

Let us begin by defining the goal of the process as finding a good enough candidate for the job. This is how a "Satisficer" would look at this. An "optimizer" will look for the best candidate possible and thus will take much more time and effort to fill up the opening – is this beneficial for the organization? I dare say it is not, since not only is the job open for a longer time, meaning that required output is delayed, the recruiting team also spends more time on this task and thus is not available for other tasks.
To recruit a good enough candidate you first must define the characteristics of a good enough candidate. It is important to keep some slack in these definitions to allow for variability between candidates, so recruiters have a real choice. In order to get a good understanding of necessary characteristics, just get a good understanding of the job this candidate will be doing.
Adding a person to the company's staff is a process of change and should be handled like such (as discussed here). Hiring managers should consider all 4 quadrants of this change first at the general level of hiring someone and then at the specific level of hiring a certain candidate. Throughput considerations should be used.
At the general level consider:

+
-
Hire new person (change)
"Pot of Gold" - increased throughput by manning position
"Pain" – losing throughput by paying more salaries (increasing OE) without sufficient return
Don't hire
(don't change)
"Mermaid" – low OE, business as usual, no need to train and bring someone up to speed….
"Alligator" – overloaded workers, missing capabilities….
This would be valid even when the manager leading the process is very willing to hire the new person. Also notice that these are very generic possibilities and should be replaced with the specifics of each situation in real life.
At the specific level each candidate is different so a generic analysis is not relevant. One point to consider, though, is how would the analyses differ between a candidate that is an acceptable fit on all parameters and one that is a very good fit on some parameters but poor on others. Which parameters are the most important to your throughput?

Continued here

Wednesday, March 23, 2011

Road Runner Ethics

Implementing the Theory of Constraints requires everybody behave like the Road Runner (the blue bird). It is either running full speed or standing, there is no other option. It should be the same for all our resources, especially the human one.

Cost world, which demands maximal efficiency at every point, leads people into conflicts - conflicts with fellow workers (that don't want to work their hardest ALL the time but don't want to look "bad" either) and conflicts with the system that requires they achieve opposing goals (no over time, on time delivery AND maximum utilization). These conflicts lead people to slow down their pace of work, make sure they look busy (I once saw a shareware that helps cover up the fact you are not really doing anything?) and often times leads them to make the wrong decisions.

Throughput world focuses on global optima, not local optima, and thus accepts the fact that most resources should have idle time. This means that being active isn't always being productive. But to get the most out of the system we need to have all our protection where we need it - at the constraint. Anywhere else it is wasted as it is not protecting the system. Therefore, even though a resource has over capacity and is not constraining the system, once it there is work to be done - it should be done ASAP. Knowing that working hard is not penalized (=rests are allowed when there is no work) should lead everyone to that direction. So there are no conflicts between fellow workers (no one should feel bad or threatened by co-workers working hard when they are not), no conflicts with the system and the road is paved for making the right decisions.
Meep Meep