A keeper for when my startup gets to the “25-75 employees”.

Cost per Employee

For each additional employee, I see some more costs: When go from 99 employees to 100, ballpark each employee costs 1% of the building, janitorial, maintenance, utilities, security, insurance, the, say, TGIF catered lunch, etc. So, have to add up those costs and pro-rate them for each employee.

Sure, growing from 99 to 100 may not make those additional costs obvious, but growing from 25 to 75 or 100 to 200 will. Indeed, one cost of those last 50 employees is the HR department itself and also the time the CEO spent before and on his week on the beach after finding how to delegate HR.

And each new employee needs a desk, chair, phone, computer, e-mail and Internet access (likely from an ISP), purchased software, routine computer support, cloud access, desk supplies, printing and scanning, business travel, and more.

A technical employee may need books, videos, conference attendance, consulting, e.g., calling a paid expert at Microsoft to resolve some tricky issue in system administration of a high end version of SQL Server in an hour instead of working alone for a week.

And, as in the classic Fred Brooks, The Mythical Man-Month, “adding people to a late software project makes it later” because of the quadratic growth the number of person to person communications about the work. His claim is an exaggeration and is specifically for software but does have a point that is more general and significantly valid. So, adding people can slow down the work.

Sure, maybe don’t hire someone until there is some very valuable work that is well identified and described, quite doable, and very much needs to get done and where the other employees are so overworked that they are desperate to welcome a new employee to help with the work. Then hire for just that work, no more and no less. Okay. In the short term, this is fine. But in the longer term, the work to be done tends to change, and there is a bigger challenge, the “main point” just below.

Job Description

Really, the main point in a job description has to be, see what needs to be done, find a good way, maybe a really good way, to do it, and get it done.

Case Study #1

E.g., at a university with 55,000 students on campus, there was a computing center for the faculty and students. The center had an Amdahl mainframe with 6 million bytes of main memory and a processor clock of 3 million ticks per second. Nearly all the input was via punched cards, but there were also some typing terminals, dot matrix DEC Writers, typing on paper at 30 characters per second and a few video terminals that looked to the computer just like typing terminals.

The group had their own building, a staff of 80 persons, and a CIO.

The business school had a person who helped students, faculty, and staff use the keypunch machines and submit jobs to the mainframe.

The business school was starting a new MBA program and wanted some better computing. They had a committee, from mostly the accounting department, that had met for two years with no good ideas.

A new faculty member arrived, saw the situation, concluded that the place was in the dark ages and far behind what he had administered on a part time job in US national security while in grad school, and, after two weeks on campus, in the first faculty meeting, stood and suggested a super mini computer from Boston, DEC, Prime, or Data General.

He led the effort. The CIO bitterly objected. A year later the proposed computer was running, and the business school was thrilled. The site was wildly successful for years. The new prof served on a committee to pick a new CIO for the university. E.g., some students in the new B-school computer center worked out some database logic and some word whacking software, and the school alumni office, then, had daisy wheel printers going 40 hours a week building relations with the alumni. Presto, bingo — big ROI on the cost of the new computer. Others on campus wanted the same, and the system was borrowed, polished, and became a biggie for the whole campus. Biggie ROI again.

The struggling word processing department was disbanded, and the new computer became the center of word processing, for research, teaching, administration, etc. There was dial-in access, and faculty could type from home. The site became the world site for distribution of Knuth’s mathematical word processing software TeX on the computer used (a Prime, maybe the same as Bloomberg was using then).

The new prof gave a special, very popular, MBA course in computer selection.

So, there in computing the real job description was, see what needs to be done, find a good way, maybe a really good way, to do it, and get it done. That’s what the CIO and the B-school staff person were not doing and the new prof did.

Case Study #2

Once a guy was working in a software house with customers at a research lab of the US Navy. There was a competitive request for a proposal (RFP) for a software development project. But one existing project was in trouble, and the guy was asked to go for a week and try to save the project. By the afternoon of the first day, it was clear that there was no way to save that project in a week.

But there was also a problem with the RFP — in an important sense in estimation of the power spectrum of ocean waves, what the customer was asking for was impossible. From the classic Blackman and Tukey, The Measurement of Power Spectra, written at Bell Labs and Princeton, the accuracy the customer wanted would require hundreds of hours of data.

To illustrate this point, in the next four days of that week, the guy typed in some software to generate sample paths of some stochastic processes and calculate and graph the sample power spectra, over and over as the stochastic process continued. The software clearly illustrated how early on the sample power spectra were wildly inaccurate but, over time, converged to exactly what generated the stochastic process to begin with.

So, the Navy really could get the accuracy they wanted, but they needed more data and had just learned how much more.

The software house got sole source on the development RFP.

So, again the real job was, see what needs to be done, find a good way, maybe a really good way, to do it, and get it done.

HR Recruiting

Back to HR, there is a danger, a huge danger, a great way to shoot the whole future of the whole company with a .308 shot straight to the gut, right along all totally without sound or pain.

The HR shop is usually a distaff ghetto of psycho majors who have no chance at all of any insight at all into the main qualification of the jobs — see what needs to be done, find a good way, maybe a really good way, to do it and get it done — or how to recruit such people, especially for 2016.

Indeed, for likely the most important work for the future of the company, the usual approaches of HR will strongly filter out just the people who are good at the main qualification.

How? The HR people are, without exception, highly sensitive social butterfly types, natural applied sociologists. They all would get at least a B-, without study, in Erving Goffman’s obscure, classic The Presentation of Self in Everyday Life.

Then with a nerd, just terrific at the real work, suddenly saving the day with some fast work in, e.g., power spectral estimation, the HR people will, even in just a first phone screen, get a headache and a backache, do an upchuck, and feel worse than having a barbed wire enema. All their central life belief system, highly cherished, exquisitely sensitive, psycho-social, Victorian garden party, touchy-feely sensors will sound loud alarms at which time the phone screen sophomore college, psycho major, junior assistant summer intern will report, with some strained facial expression, a huge no fit, and that will be the end of the chances of that candidate at that company and, also, an nice step to the end of the company itself. Now, the situation might be better if the business of the company was catering dances for sorority parties at the Seven Sisters, at least as long as the catering staff didn’t have work with anything as technical and dangers as 120 V, AC extension cords!

In general, usually there is just no way the desirable candidates for a company in 2016 will make it through HR recruiting. No way. Not a chance.

There is an even bigger challenge: Fundamentally, the goal has to be to recruit people who can do really good work on really important work — see what needs to be done, find a good way, maybe a really good way, to do it and get it done — where the present staff doesn’t know about the work and doesn’t know how to do it.

So, how is the current staff going to recruit people they don’t understand? And if the current staff does start to understand, will they feel threatened from the competition and reject the candidate? Uh, in B-school courses, in psycho-social terms that’s called goal subordination.

In picking someone to do a really good job painting the ceiling, say, picking Michelangelo, whom do you ask? A bunch of house painters? Some people who have no idea at all how to conceive of a masterpiece?

Really, finding good people is, first, the job of the CEO, and as the CEO “delegates” that he has to find a way so that the company will continue to find good people.

Uh, maybe right along he should have some friends send in their resumes and, if they get that far, have the phone screen just to see if HR can recognize the good stuff when they see it. Likely they won’t.

So, for HR in recruiting, for a quote,

HR staff members are to smile, be nice, offer water, soda, coffee, tea, smile, be nice, hand over the benefits packet, help with the interview schedule, help with travel and lodging arrangement and expense reimbursements, smile, be nice, give directions to the rest room, offer snacks, smile, be nice, offer a phone to call home, help arrange lunch, smile, be nice, and, did I mention, smile and be nice?

For a candidate, under no circumstances is anyone in HR ever to (A) ask anything about qualifications or (B) form or express an opinion about qualifications.

Evaluations

Somehow I suspect that the recommendations for employee evaluations will become too rigid and emphasize simplistic formality over essential reality.

Danger

Even HR people understand that they would like to get more power. So, they can suck up to the CEO and COO and suggest that they, HR, can for employee and management evaluation provide a more objective, alternative to the usual management hierarchy. Then, too soon, HR can be the become the main power in the company, putting their psycho-social, touchy-feely, fuzzy-bunny play time standards ahead of any and all the real expertise in the rest of the company.

A CEO or COO who permits that deserves to have their company die.

.
There are a lot of nifty pieces of software/apps that will cut a few corners for you. There are a lot of retired HR folk who can work very reasonably on Craigslist or other sources.

BRC
http://www.themusingsofthebigredcar.com

What do you recommend for outsourcing/delegating HR. Something like an all-in PEO or more piecemeal?

Here is how to do just that.

http://themusingsofthebigredcar.com/human-relations-crawl-walk-run/

BRC
http://www.themusingsofthebigredcar.com