The Impossibility Triangle in Talent Management

I have taken to asking an unfair question in interviews in recent months: as a manager, how can you make sure you help people play to their strengths, while making sure the organization meets its goals on time? The question is unfair because I believe the right answer is you can’t. I haven’t yet gotten that answer from anybody. Most of the time, I get carefully noncommittal answers couched in terms of “there are always trade offs” or faux-pragmatic ones of the form, “well at some point people have to get on the same page and realize that a company is in business to make money.” Both answers are misguided. By trade-off people usually mean deterministically balancing coupled motivations in the presence of some sort of scarcity; they do not usually factor in uncertainty. Talent management introduces competing uncertainties, which are more complex beasts. The second answer is faux-pragmatic because everybody already understands tautological statements about companies being in business to make money. The realization makes no difference at all in how the psychological calculus of strengths and motivations plays out. In other words, yelling business realities out from the rooftops won’t help you attract talent or prevent you from hemorrhaging it.

So let’s keep ourselves honest here, by beginning our search for a next-generation talent management theory, Theory W, by acknowledging the fundamental limits in play. I represent these in the Theory W Triangle (if you are curious, this is inspired by the well-known Pizza triangle and the less well-known Spreng triangle, which I discussed before).

Let’s dig in.

The Three Basic Uncertainties of Talent Management

The Theory W triangle captures 3 coupled uncertainties that relate rather like position and momentum do in the Heisenberg uncertainty principle. These three basic uncertainties constraining talent management are:

1. What uncertainty: You may know nearly exactly what you will attempt to do, to deliver business value (goal-setting), or you may only have a broad set of ideas/directions (opportunity creation).
2. Who uncertainty: You may know whether each individual will perform at nearly their maximum potential (strengths casting), or you may not (needs casting — where you’re never quite sure until you try whether you’ve shoved square pegs into round holes).
3. When uncertainty: You may know exactly when you will realize the value of your investment in people (returns management) or you may not (expectations management).

Note that even at the vertices, there is uncertainty. The key to the triangle is that you can only minimize one of the uncertainties at a given time (i.e., as a manager, you can only choose a single point in the triangle, at any given time). To illustrate:

• You insist on specific goals being achieved. You will be forced to do needs casting: putting people in roles dictated by the logic of the fixed goals, which will lead to unpredictable performance. You will also find it necessary to leave deadlines flexible, since needs-cast people deliver over unpredictable time horizons (aptitude is sometimes defined as the time it takes somebody to learn something. When individuals are not playing to strengths, this becomes a crapshoot).
• You insist on a strengths philosophy — you will only give people assignments that play to their strengths. This will in general mean that you will need to give up fixed goals, and instead maintain a large opportunity set to allow all individuals to find strengths-oriented niches with high probability. So you won’t know what you will actually deliver until people actually find their way. You also (again) lose control over time of value delivery, since that is related to what people choose to take on. You can only expect that a certain proportion of individuals might deliver.
• Finally, you can insist on timeliness of value delivery (returns management) — you want delivery by a certain date, and you want a certain amount of value delivered (commensurate with headcount), in some not-wildly-inappropriate metric for comparing apples and oranges (patents, sales targets, number of software features or lines of code for example). In this case goals will slip in a qualitative sense, as will the match between people’s strengths and roles, since the time-pressure will not allow people enough time to explore and find their ways to predictably-high-performance niches. If you loosen your grip on value delivered per unit time, you will switch to expectation management: your returns will become more variable per unit time, or there will be variability in deadlines.

Here is why this theory  is complex, but necessary.  Most people today only appreciate binary uncertainty constraints. For example, you can get at the right side of the triangle as follows: give up fully-defined goals and work with an opportunity set. Now you can trade off uncertainty in the match between people and roles and the uncertainty in value delivery schedules. Paraphrasing Lincoln, you get a case of you can satisfy all the people some of the time, some of the people all of the time, but not all of the people all the time.

Similarly, if you give up strengths-casting altogether and move to pure needs-casting (i.e. treating people as fungible up to nominal resume-level skill descriptions), you are still left with the goals-time tradeoff. This is acknowledged in software development in particular: agile/Scrum style methods stick to timeliness and give up goals when necessary, while waterfall methods deliver against defined detailed requirements at the cost of heavy deadline slippage and overspending. This is the bottom side of the triangle.

Finally, if you reduce time pressures, you get the classic trade off that drives R&D organizations, represented by the left side of the triangle. You can let individual researchers wander to rich and fertile domains of intellectual property that match their strengths, at the cost of potentially wasting money discovering stuff that your company is not positioned to take to market (open innovation methods – licensing, selling and trading IP, only partly mitigate this), and making it much harder to form coherent larger teams. Or you can accept some level of top-down cascading of strategic imperatives arising from current business goals and dictate, thou shalt research nanotechnology rather than quantum computing.

Think these three binary tradeoffs are bad? The triangle is telling you that even those are too good to be true. You need to get to at least three variables to get to a minimally reasonable model of reality.

The Good News

The triangle is not all dismal news however. Each vertex represents an uncertainty minimized to its lowest possible level, but its corresponding side (the one opposite it) does not represent a worst-case scenario. It represents an average case scenario. This means, for instance, that if you do needs casting, you won’t automatically put everybody in roles that are their worst nightmares. You’ll get some fits perfectly right, perhaps by accident, some will be partly right and some will make for some very unhappy people who will start looking for a way out.

This gives you several levers to influence the situation, even within the constraints of the triangle:

1. Movement: Make sure you keep moving your location within the triangle, matching your chosen uncertainty-trade off point to the situation. This will allow you to selectively drive down the uncertainty that is currently hurting you the most. If you see a lot of potential in somebody and that person is currently needs cast in an unhappy situation, do some tactical strengths casting. If your organization hasn’t delivered in years, risk losing people by moving to need casting, so you deliver something. If you just got a big win, use the temporary reprieve to let deadline pressures go and meet goal-clarity and strengths-casting needs. This game is rather like the keep-multiple-plates spinning metaphor I used earlier in a different context.
2. Bench dynamics: Even within a fixed point (at least within the interior of the triangle), you can keep uncertainty levels the same by swapping people in and out of ‘active play’ roles, while making sure the same proportion of goals stay fixed, at the same level of time uncertainty. This obviously means you are financially healthy enough to be carrying some slack in the first place. But if you are playing with zero bench strength, you are probably on the brink of bankruptcy. No fat is anorexia.
3. Money: It goes without saying that the more financial freedom you have — discretionary funds that aren’t earmarked for anything — the more you can break the constraints of the triangle. Google can create such a great work culture (and accept a high failure rate and lack of value delivery from a big proportion of its people) because it has a cash cow that produces money to burn.
4. Hiring for the triangle: You can explicitly use hiring to move you to parts of the triangle that your current set of people cannot get you to, or even move boundaries up (which improves the average case performance).  Here’s how it works. To raise the ‘average case’ performance of all three of your boundaries, hire multi-talented people who can contribute in multiple ways in terms of direct content of work, and play multiple roles in terms of management (both lead and follow) and are comfortable with both deadline-oriented and exploratory environments. If this sounds like a panacea, it isn’t. Most people who are multi-talented aren’t deeply good at one thing, which means you’ll lose your ability to go after goals beyond a certain level of technical complexity. There are, of course, those superstars who are world-class on multiple fronts (Michael Jordan on scoring and assists for instance), but those are so rare that you can’t rely on being able to find and hire them.

Updating the Definition of Talent Management

So let’s update the classic definition of talent management: putting the right people, with the right skills, in the right jobs. Most existing theories of talent management fit within this fairly sophisticated definition. Peter Cappelli’s magisterial Talent On Demand, reviewed the history and these models, why they used to work, why they don’t work anymore and what we can do about it. Cappelli’s argument can be further simplified for our needs as follows: traditional talent management no longer works because the world has become a more complicated place, and our solutions must become more complex as well. Without further ado, here is my revised definition, based on the triangle of impossibilities:

Talent Management: dynamically managing the returns on human resource investments by exploiting the coupled uncertainties involved in who, what and when decisions.

We’ll explore more as I can think of more things to say about this. In the meantime, anecdotes welcome, since as you can see, this series of articles has been heavy on concept and light on examples (partly because I can’t share the examples I personally encountered, that inspired these theories, because they are too recent, and I can’t think of ways to anonymize them. Presumably, those of you with more years spent managing people can find anecdotes that are distant enough to be safely shared!).