A working guide. Four steps, and a calculator at the third.
This takes a single job in your organization and works out what the person doing it
produced, in money, after taking off what somebody ordinary would have produced anyway and what the role
costs to employ. Then it carries that figure through to what it is worth to the value of the business, which is the
thing finance means by enterprise value.
Who this is for. It is written for the person who has been asked what the
workforce is producing and has not had an answer to give. You do not need a finance background to use it.
Where a term is one finance uses rather than one HR uses, there is a small i
beside it. Press it for a plain explanation.
Step one
Choosing the role
Two questions, and the answer to both has to be yes. If the answer to the second one is no,
the role still creates value. It just cannot be priced yet, and the table below says what would have to
change.
Question one
Does it matter more than most jobs, and do people in it perform very differently from
each other?
If everybody in the role performs about the same, there is no difference to measure and
nothing to attribute to any one person.
The formal term is an A position. Huselid, Beatty and Becker set it
out in Harvard Business Review in December 2005 under the title “A Players or A Positions?”.
Becker, Huselid and Beatty worked it through in full in The Differentiated Workforce in 2009. Their test is a job with
disproportionate impact on strategy where performance varies widely between the people who hold it.
They put it at under twenty percent of a workforce.
Question two
Is the output of the work already counted somewhere, in money?
Already in dollars, before anybody converts anything. Not a survey result, not a
performance rating, not a score. A figure that already sits in a system the company runs for its own
reasons, such as billing or sales.
If the role creates value by
Then
What would have to exist first
Bringing in revenue or cash, and it is counted
Price it now
Nothing. Go to step two
Preventing something bad from happening
Not yet
Records of those events and what each one costs, and an agreed way to say a particular
person prevented one
Effects that are spread out and slow to appear
Not yet
The benefit tracked to a named project with dates on it, so it can be tied to a period
Work a whole team does together
Not yet
A way to see each person’s share inside the system that already records the work
Keeping customers rather than winning them
Not yet
A comparison group of similar accounts, to show what would have been lost otherwise
Start with one role. Not a program, and not an exercise covering the whole
workforce. One job where the output is already recorded in money against a named person. The second and
third roles are easier because of the first. And a second role that fails teaches you more than a second
role that works, because it shows you where the method stops.
Step two
The information you need, and where to get it
This is the part that stops people, and it is not difficult. It is administrative, and it costs you a
few conversations.
The figure is not one anybody carries in their head. It exists, in a system somebody else looks after.
What you need
Where it is
Who to ask
Twelve months of output for that role, person by person
The system that already records the work for its own reasons. Billing or collections for
recovery roles. The sales system for booked revenue. The time-and-billing system for professional
services. The claims or policy system in insurance.
Operations, or whoever looks after that system
What the software did over the same period
Which tools were switched on, and when. Anything the software produced is not the
person’s, and it is the first thing an outside reviewer looks for.
IT, or the supplier’s own business case
A comparison group
The same job, the same pay band, the same kind of workload, over the same twelve months.
You are looking for the middle of that group rather than the best or the worst of it.
The same system, the same extract
Base salary
Payroll, or the HR system of record. The individual’s actual salary, not the
midpoint of the band.
HR
The cost on top of salary
Benefits, employer taxes and a share of overheads, expressed as a multiple of salary.
Do not estimate this one.
Finance owns this figure
The multiple the business is valued at
Whatever the company is currently held or modeled at. Only the part of the contribution
that repeats year after year earns it.
The CFO, or the investor’s model
If twelve months of output does not exist anywhere, that is your finding. The role cannot
be priced yet, and you are back in step one.
Who decides what. HR proposes the comparison figure, because HR is closest to the
people doing that work. Finance checks that it is believable and cautious, and Finance owns the cost side.
Two signatures, and neither function moves it on its own. That hand-off is where these exercises succeed
or fail.
Step three
Working out the number
Fill in what you found. Everything calculates as you type.
What the role produced
Twelve months of output for one person, in money, after taking out anything the
software produced
$
−What somebody ordinary would have produced
A competent, unremarkable person doing the same job with a similar workload over the
same twelve months. Take the middle of your comparison group.
Finance calls this the counterfactual. It is what would have happened
anyway, without this particular person. It is the most important figure on this page and the one an
outside reviewer goes to first, because without it you are comparing a result to nothing.
The lower you set it, the more you have to prove. Take it from the middle of the comparison
group rather than from a view about how good somebody is.
$
−What the role costs to employ
Salary, times the figure Finance gives you for everything on top of it
Finance calls this fully loaded cost. Salary is only part of what a
person costs. Add employer taxes, benefits, and a share of the overheads they use. Finance expresses
it as a multiple of salary and will have a figure they already use. Charge all of it. Leaving
some of it out is enough on its own to have the whole figure dismissed.
$
×
Total cost of employing them
—
What is left, and attributable to the person
—
Finance calls this net attributable contribution. It is the money this
role added that would not have arrived without this person, after the business has paid for them in full.
This is the figure you can defend, and it is smaller than the headline number on purpose.
×The multiple the business is valued at
Only count the part that repeats every year
Finance calls this the exit multiple. Companies are bought and sold
for a multiple of their annual profit. If a business is valued at twelve times profit, every extra
dollar of lasting annual profit adds twelve dollars to what the business is worth. A one-time gain is
worth its cash and nothing more, so only apply the multiple to the part that repeats.
×
What this one role is worth to the value of the businessThe
repeating contribution, multiplied out
—
The point where the claim fails
—
Enter what the role produced, and this shows the figure at which your claim
would be worth nothing.
This is the break point. Set the comparison figure this high and the whole
claim is worth zero. Publishing it, in writing, alongside your number is what separates a claim
somebody can check from one they have to take on trust.
Nothing you type here is sent anywhere. It stays in your browser.
Step four
Five things it has to pass before you use it
When somebody outside the business examines a workforce claim that does not hold all five,
they do not argue the figure down. They take it out altogether.
The method calls these the Five Conditions: a stated counterfactual,
de-overlapped, full cost charged, conservatively bounded, and auditable. They travel as a set. A claim
holding four of them is not eighty percent defensible, because the one that is missing is the one that
gets tested.
1
You have written down what would have happened anyway
In writing, before you quote the figure to anybody. A number compared to nothing is an
opinion, not a measurement.
This condition is the stated counterfactual. It is the figure you
entered in step three for what somebody ordinary would have produced. The condition is not that you
have one. It is that you wrote it down before you quoted the number. A baseline produced after
somebody challenges you is not a baseline, it is a defense.
2
Nothing is counted twice
Other people in the business may already be claiming credit for some of the same money.
Added together, the same dollar gets counted more than once.
This condition is called de-overlap. Suppose a software rollout reports that it recovered
a million dollars. Your role’s figure includes some of that same recovered million, because the
person and the software both touched it. A process change in another team claims a slice as well. Add
the three together and the business appears to have recovered the money two or three times over. It
did not.
How to find it. Before your number goes anywhere, ask three
questions:
What else in the business is currently claiming credit for money from
this same source?
Was any new technology switched on during the twelve months you
measured?
Is this role’s output already sitting inside a figure somewhere in the plan or the budget?
The check that settles it. Add up every claim that draws on the
same pot of money, and compare that total against how much the pot moved over the period.
If the claims add up to more than the pot moved, you have double counting, and you have found
it before anybody outside did. It needs no knowledge of how to divide the money up, only
whether there is something to divide.
The rule for sorting it out. The claim that can trace the money
to a specific transaction owns it. Everybody else states their benefit after taking that part out.
If some of it belongs elsewhere, your number gets smaller. A smaller figure that survives being checked is worth more than a
larger one that gets removed.
Where it gets harder. When two claims can both trace
the same money, dividing it is a judgment rather than a calculation, and a guide cannot settle it for
you. Write down the split you made and the reason for it, and put it somewhere Finance can see it.
An undocumented judgment is the one that gets found later, by somebody else.
3
The full cost of the person is charged
All of it, not just salary, and the individual’s salary rather than the band midpoint.
This is the fully loaded cost condition. The test is that nothing was
left out to make the figure look better: employer taxes, benefits and a share of overheads, on the
person’s real salary. Charging the cost at the margin, or using the midpoint of the pay band, is
among the easiest things for somebody outside to recalculate, which makes it a poor place to be
optimistic.
4
The figure is a floor, not a best case
Where you had a choice, you took the lower number. A floor you can defend is worth more
than a total you have to argue for.
This is called conservative bounding. Where you had a range or a
judgment call, you took the end that works against you. The condition is not accuracy, it is direction:
every error should push the figure down rather than up. Built that way the number is a floor, which somebody
can underwrite rather than audit.
5
The point where it fails is published alongside it
The figure at which your own claim is worth nothing, written down next to the claim.
This condition is auditability, and the break point is how you show it.
Somebody should be able to follow your figure back to the records it came from, and see the point at
which it stops being true. Published before you are asked, it reads as method. Produced after you are
asked, it reads as a concession.
A number that says where it fails is a number somebody can check. A number that does not
is a number somebody removes.
What you have at the end. Not a workforce strategy, and not a dashboard. One
sentence you can say in a board meeting that holds up when somebody checks it, and a method that will do
the same thing again on a second role.
Garrett Walker · Workforce Value Creation & Advisory