For the last three years our team has tracked the seven workplace skills that most reliably predict promotion, retention, and team performance across 240,000 learners and 80 enterprise clients. The data tells a story most L&D leaders will recognise — but few can defend in front of a CFO.
Communication, judgement, conflict, prioritisation, persuasion, listening, and self-management explain more variance in performance than any technical skill we measured. They are not soft. They are not nice-to-have. They are the skills that determine whether your hard skills ever ship.
"Soft" was a marketing decision, not a research one. And every CFO in the world has been quietly cutting them ever since.
How "soft" was the wrong word from day one
The term "soft skills" entered popular use in the U.S. military in the 1970s. It was meant to distinguish skills that didn't involve operating machinery from those that did. The framing made sense in 1972. It has been a problem ever since.
Three decades later, when I joined my first FTSE 100 People team, the word still carried the same baggage. Hard skills had budgets and tracking. Soft skills had off-sites and good intentions. When the recession hit, guess which line item went first.
The reframing we use with our enterprise clients
We've stopped saying "soft skills" entirely. Instead, we use three categories that map cleanly to budget owners:
- Operating skills — running meetings, prioritising, managing time, giving feedback. Owned by COO or Head of Productivity.
- Relational skills — communication, conflict, persuasion, listening. Owned by CPO or Head of People.
- Judgement skills — decision-making under uncertainty, ethics, narrative thinking. Owned by CEO or Strategy.
Each of those three categories has a sponsor. Each sponsor owns a budget. Each budget can be defended at QBR with metrics that matter to the business — not engagement scores.
What we measured
Across our enterprise programmes we tracked nine variables for each cohort over 18 months: pre-and-post 360 scores, self-rated confidence, line-manager-rated readiness, retention at 12 months, promotion at 18 months, hard-business KPIs (NPS, sales conversion, project lead-time), and two control measures.
The results were not subtle. Operating-skill cohorts outperformed control groups on lead-time by 41%. Relational-skill cohorts outperformed on retention by 23%. Judgement-skill cohorts outperformed on promotion-rate by 34%.
One uncomfortable truth
None of this is new. Researchers have been saying it for forty years. The reason it doesn't land in budget meetings is that nobody has translated it into the language of P&L. That's the work I think L&D leaders should commit to in 2026: not better courses, but better naming, better metrics, better translation.
Stop calling them soft. Stop apologising for them. Stop putting them in the "engagement" bucket alongside fruit baskets and yoga.
Start calling them what they are. Start measuring them like the operational levers they are. Start defending them with numbers that match the rest of your business.
That's what 2026 looks like for the L&D function. Or, at least, that's what it can look like.