A request arrived last month in the plainest possible form: soft skills training for forty engineers, two days, covering communication, teamwork and time management. Every word in that sentence is defensible. None of it can be observed. That is the whole problem with the vocabulary in this field, and it is a problem created by the buyer long before a vendor is chosen.
Three labels pointing at different objects
The hard and soft distinction came out of military training, where hard skills were the ones you could score on a range and soft skills were everything left over. The residual definition stuck. Soft skills became the drawer that everything went into when it was about people, and a drawer defined by what it does not contain has no edges.
Behavioural training names the target differently. It does not describe a temperament. It describes what a person does in a defined situation, at a level of detail where two observers would mostly agree. “Listens without interrupting, then repeats back what they heard before answering” is a behaviour. “Good listener” is a compliment.
Competency development adds a third element: a scale. A competency is a small cluster of behaviours, written at levels, tied to a role. It answers not only what a person should do but how much of it the role requires, and where this person sits against that line. It is the only one of the three that supports comparison across people, which is why it is the language of a competency framework and the other two are the language of a training request.
The word you write decides what gets measured
| Label | What it names | Usual instrument | What you end up holding |
|---|---|---|---|
| Soft skills | A residual category | Feedback form, end-of-day rating | A satisfaction score |
| Behavioural training | Specific observable acts in defined situations | Observation against written anchors, before and after | A description of what changed |
| Competency development | Behaviours mapped to a role and a proficiency level | Assessment centre or structured assessment against levels | A position on a scale |
A vendor cannot be held to a standard the buyer never stated. Write soft skills and the only deliverable anyone can check is the feedback form. Write that supervisors give a two-minute structured handover at shift change naming status, risk and next owner, and you have created something an observer can score. The difference is not in the training. It is in the sentence that bought it.
The word itself also carries four assumptions, and they are usually unexamined.
- That the skill is a trait, and traits are hired rather than taught.
- That it is secondary to the technical job, so it gets the second budget and the second facilitator.
- That no preparation is needed, because there is nothing specific to prepare for.
- That no follow-up is needed, because there is nothing specific to follow up.
Each of those is a decision, and all four are made by a single adjective. The engineering manager who says “we just need them to communicate better” is not being lazy. The phrase is the lazy part, and it will cost him a programme that produces nothing he can point to.
Why the label changes what gets assessed
If the target is a disposition, the natural instrument is self-report, because there is nothing else to look at. If the target is performance in a situation, the instrument is observation against anchors by someone who sees the work. If the target is a role standard, the instrument is an assessment against levels, and the output is comparable across people.
Take forty engineers, two days, a fee of roughly eight lakh (illustrative arithmetic, not a quote). Measured by reaction, the return is an average score to one decimal and a set of comments about the facilitator. Measured by behaviour, the return is a count: how many of the forty now do the three specified acts when observed in a real meeting. Measured against a role standard, the return is a distribution: how many are at the level the role requires, how many are below it, and by how much. Same fee, three very different qualities of information.
A label is not a description, it is an instruction about what to look at, and soft skills instructs everyone to look at the person rather than the work.
I still use soft skills in conversation. With a plant head who wants a shorthand for why a brilliant technician is struggling to run a team, it is the clearest phrase available and there is no harm in it. It is a useful phrase for talking and a poor phrase for procuring. The test is the document, not the conversation.
Choosing the word that fits the job
Three questions do most of the work.
- Does the phrase name something an observer could see? If not, it is a topic, not a target.
- Is it tied to a role and a level, or floating? If floating, you can train it but you cannot staff against it.
- Would the same sentence still make sense in a job description, an appraisal and a training brief? If it only survives in one of the three, it is the wrong sentence.
For a one-off programme, behavioural language is enough and competency language is overkill. For a framework that will govern hiring, appraisal and development across a company, you need the levels, because without them every conversation about a person becomes an argument about adjectives. Mapping those levels to roles is a separate job with its own failure modes, covered in competency mapping and its application.
What to do on Monday
Take the last training request you wrote and find the phrase: better communication, teamwork, ownership. Then do four things to it.
- Name the situation where it fails today, and who notices the failure.
- Write the observable act in one sentence, starting with a verb.
- Decide who will look at it, when, and against what anchors.
- Only then choose the label, the duration and the vendor.
The test at the end is simple. If two managers reading your sentence would disagree about whether a named person did it last Tuesday, the sentence is not finished. Rework it until they would agree. That exercise takes twenty minutes and it is the difference between buying a programme and buying an outcome.