Last time, this series took a hundred individual skill trajectories and read them together, until a talent map appeared where a performance report used to be. It showed where capability concentrates, where it is thinning, and which critical skill hangs on a single person nobody had noticed. That is a genuinely useful picture. But a picture is not a decision, and the next obstacle is rarely analytical. It is a meeting, usually the one where the budget gets set, and it is where good evidence most often goes to be politely ignored.
L&D has a long habit of walking into that meeting with activity: courses delivered, completions logged, hours consumed, satisfaction scores. Finance does not read those as evidence of anything except spend. The talent map changes what can be brought through the door, but only if it is translated from the language of learning into the language of risk, capacity and cost. This post is about making that translation.
A completion rate answers a question the finance director never asked. They are not wondering whether people finished the course; they are wondering what the organisation gets for the money and what happens if the money is withheld. An activity report cannot answer either, because it describes what L&D did rather than what changed in the business. A request built on it sounds like a department asking to continue because it exists, and it is usually treated accordingly: flat, trimmed, or quietly cut when the quarter tightens.
Two cycles of evidenced skills data offer something structurally different. They describe the workforce, not the department: which capabilities the strategy depends on, how many people hold them, and which direction those numbers are moving. The question on the table stops being whether training is worth funding and becomes whether the organisation is comfortable with what its own data says about its position.
The strongest opening in a budget conversation is a specific, named exposure. Not “we need leadership development funding”, but “the platform migration depends on a skill held by two people, one of whom has been flat for two cycles, and there is no third.” That sentence can be checked, it has a number inside it, and it describes a cost of doing nothing in terms any executive recognises: delay, dependency and replacement risk.
This is where the
analytics and reporting view earns its keep. The aim is not a bigger dashboard. It is the two or three charts that isolate a concentration or a thinning in a skill the business has already said it cares about. Everything else, however interesting, belongs in an appendix. A budget meeting has room for one clear exposure, and it is better to land one than to gesture at ten.
Once the exposure is on the table, the temptation is to answer it with a grand programme. Resist it. Decision-makers trust requests that are proportionate to the problem and bounded in time, and distrust anything that reads as a framework in search of a budget. A better shape is an explicit goal, a named group of people, a defined window, and a measure that the next assessment cycle will settle one way or the other.
That is precisely the job of structured
training goals: a stated target for a specific skill and a specific population, tied to the evidence that justified it. Framed this way, the request is also falsifiable. If the next cycle shows the skill rising where funding was applied, the case for the following year is already written. If it does not, the organisation has learned something cheaply, and the credibility of the person who said so in advance goes up rather than down.
Executives rarely choose between a proposal and nothing. They choose between a proposal and the other ways of closing the same gap, usually hiring externally or accepting the risk. Treat those honestly in the paper. External hiring is sometimes the right answer for a skill the organisation has no foundation to grow, and saying so builds more trust than pretending development always wins. Where there is a rising internal base to build on, the comparison tends to favour development on speed and retention, but it should be allowed to favour the other side when it does.
Delivery cost belongs in the same honest ledger. Content that can be built and refreshed quickly changes the arithmetic, which is why the economics of
course creation and management are worth stating plainly instead of leaving finance to assume the slowest, most expensive version of the work. A proposal that has already priced its own delivery is much harder to dismiss than one that leaves that number open.
The first evidence-led budget conversation is the hardest, because the room is used to a different kind of paper. The second is easier, because there is now a prior cycle to point to: a goal that was set, funded and measured, with a result nobody had to take on trust. Over a few rounds the relationship with finance shifts from defending a cost line to reporting on a capability position, which is a much better place for L&D to be standing.
The practical instruction is short, and it follows from everything this series has built. Do not carry the talent map into the meeting as a finding to admire. Carry it as an exposure, a bounded proposal, an honest comparison, and a date on which the next cycle will say whether it worked. Evidence that comes with its own test is the kind a budget meeting can actually act on.
Last time, this series watched a second assessment cycle land beside the first and turn a static heat map into a trend — proof, skill by skill, of whether a development approach actually works. That was still a story about individuals: one person’s rating, compared with itself, six months apart. Almost every organisation stops there, because a single trajectory is a satisfying, legible thing. It answers one question well: is this person getting better at this skill? It was never built to answer the question sitting one level up.
That question belongs to the organisation, not the person: where does capability actually concentrate, where is it thinning without anyone deciding it should, and who is quietly carrying a critical skill that nobody has noticed depends on exactly one person? A single trajectory cannot see any of that. A hundred trajectories, read together rather than one at a time, can — and the view that appears looks less like a performance report and more like a talent map.
Line up every individual’s skill trend against the same axis and something appears that no single line can show: clustering. Three people rising together in the same critical skill is a pipeline. Three people flat in that same skill, all reporting to the same manager who retires next year, is a very different pattern — and it is invisible from inside any one of their individual reviews, because each manager only ever watches their own line move, or not, in isolation. Nobody in that chain of one-to-one conversations is lying or even withholding anything; the pattern simply does not exist at the scale any single conversation operates at.
This is the argument for building the skills analytics view at the level of the organisation rather than leaving it as a dozen separate spreadsheets belonging to a dozen separate managers. The aggregate is not simply a bigger report. It is a different instrument, built to answer a question the individual view was never asked — and, tellingly, it is usually the first time anyone has actually looked at the whole workforce’s trajectories side by side rather than one row at a time.
Succession planning has traditionally run on a name in a box: a critical role, and the two or three people theoretically ready to fill it. Two cycles of trend data quietly force that box open, because a name sitting inside it says nothing about whether the person’s relevant skills are rising, flat, or drifting the wrong way while everyone assumes the succession question is already handled. A bench that was true in March and never checked again is not a plan. It is a memory of one, and organisations tend to discover the difference at the worst possible moment — when the role suddenly needs filling and the assumed successor turns out to have been flat in the relevant skill for a year.
The uncomfortable finding, when organisations actually look, is usually concentration rather than absence: a critical capability sitting inside two or three trajectories instead of being distributed at all, each of those trajectories ordinary and unremarkable on its own, and none of them backed up by a fourth. Position asks whether the skill exists somewhere. Trend, read across the bench rather than down a single row, asks the harder question — whether it exists in more than one place, and whether that is getting better or worse.
Once trajectories are grouped rather than read one at a time, three shapes do almost all of the work. A rising line in someone who wants more of the skill is the easy case — succession and stretch, decided together in a genuine skill-will conversation rather than assumed from the chart. A flat line is not automatically a problem; flat in a skill the role no longer needs is simply accurate, and chasing it spends a quarter nobody had to spend. The mistake is treating all three shapes the same way, which is what happens by default when nobody has separated them out first.
The line that deserves the room’s full attention is the falling one in a skill the organisation still depends on, because it rarely announces itself as urgent. Nobody schedules a crisis meeting over a two that used to be a three; the number simply drifts while the work that once exercised it moves to someone else. Turning that finding into a dated training goal is exactly the mechanism this series described two posts ago — it only helps if something is still watching for the drift in the first place.
A succession conversation built on individual memory tends to produce the same three or four names every year, because those are the people leadership happens to remember clearly. A succession conversation built on aggregated trend data produces a different, occasionally uncomfortable list — the quietly rising person two levels down whom nobody has mentioned in a planning meeting, and the well-regarded name whose relevant trajectory has actually been flat for two cycles running. Neither of those findings is visible from memory alone, however good that memory is; both are sitting in data the organisation already collected and simply never grouped.
It also changes what shows up in the development review itself, because a manager holding that conversation with an aggregate view in hand is negotiating from evidence rather than impression. “I think they’re ready” becomes “these three skills have risen two cycles running, against evidence, and this one hasn’t moved” — a sentence a succession committee can actually act on, rather than one it has to take on faith.
The org chart shows who reports to whom. It has never shown where capability actually concentrates, or where it is thinning without anyone deciding that it should. Two cycles of honest, evidenced skills data, read across people rather than down them one at a time, comes closer — not a replacement for judgement about succession, but the layer of evidence that judgement has usually had to do without. It will not draw the org chart’s replacement on its own, but it will tell the room, in its own numbers, exactly where that chart is quietly lying.
The practical instruction is short, and it is the same one this series keeps arriving at from a different direction. Do not stop at the individual trajectory. Group it, look for concentration and thinness rather than just movement, and let the bench be redrawn by what the data actually shows — including the name nobody would have thought to put there first.
Last time, this series followed an honest skills rating into the working week — how a gap becomes a named goal, how a goal is built to fit a real calendar, and why the loop only counts once someone goes back to check whether the number moved. A closed loop is a satisfying thing: a gap found, worked on, and visibly narrowed. It is also, statistically speaking, an anecdote. This post is about what appears when the cycle runs again across a whole organisation — a second snapshot, and with it the first honest answer to the question leadership has been asking since the programme began: are we actually getting better?
Most organisations still read skills data the way they read a photograph. Where do we stand? What is covered? How much of the map is red? Those are position questions, and the first assessment answers them well enough. But a heat map is a photograph of a moving object. Two organisations can produce identical heat maps while travelling in opposite directions — one improving steadily into its gaps, the other quietly decaying out of its strengths. The snapshot cannot tell them apart. The second snapshot can, and nearly everything interesting about a skills programme lives in the difference between the two.
The first cycle tells an organisation where it stands, and that knowledge is genuinely valuable — expensive, hard-won, and ageing from the moment it arrives. The second cycle answers a different kind of question: what moved. The moment a second honest rating lands beside the first, the data acquires a shape rather than just a state. Gaps that were worked on and closed. Strengths that faded because nobody exercised them — capability decays where it is not used, a fact every snapshot politely conceals. Teams that convert development goals into movement, and teams that accumulate development plans the way a drawer accumulates cables.
This is also the moment the previous post’s discipline pays out. The training goals that were named and dated last quarter stop being commitments and become hypotheses — each one a small, testable claim that a particular action would move a particular number. The reassessment is the result coming back from the lab. An organisation that ran the loop properly does not need to wonder whether its development approach works; it can read the answer, skill by skill, in its own data.
A delta, however, is only as honest as the two ratings it stretches between — and there are several well-intentioned ways to break one. The comparison holds only if the second rating is made against the same skill, by the same standard, with the same kind of evidence as the first, ideally in the ordinary rhythm of a development review rather than as a special occasion. Rate a skill one way in March and another way in September and the resulting movement is fiction, whichever direction it points.
The most common breakage is the tidy-up. Between cycles, someone improves the framework — renames a skill, splits it in two, sharpens a level description — and in doing so quietly cuts the thread that made comparison possible. The framework post in this series argued for taxonomies that stay small and current, and that still holds; the refinement simply belongs at the cycle boundary, declared, so everyone knows where the line breaks. The other quiet threat is rater turnover: a new manager reads the scale differently, and a team appears to improve or collapse on the strength of a personnel change. The anchor for both is evidence — structured feedback captured close to the work gives the second rating the same footing as the first, whoever happens to be holding the pen.
Once trend is visible, the skills analytics stop being a league table and start being a time series, and decisions change accordingly. Development budget can be aimed where it demonstrably works: at the gaps that move when pushed. A gap that has absorbed two cycles of effort without moving is not proof that the people are hopeless — it is usually proof that the method is wrong, most often a course where practice was needed. The trend does not just show progress; it audits the development approach itself.
Direction also reads differently against ambition. A skill rising in a person who wants more of it is a succession plan writing itself; the same skill flat in someone whose energy points elsewhere is a signal to stop pushing and have the more honest conversation a skill-will view exists for. And the trend surfaces the one risk no snapshot can: the critical capability that is quietly sliding because the work that once exercised it has moved elsewhere. Position flags the gap you have. Direction flags the gap you are about to have — while there is still time to do something cheaper than a crisis.
The second snapshot also changes what L&D sends up the stairs. The traditional report is a record of activity — enrolments, completions, hours of learning consumed — numbers about effort, silent on results. A trend report reads differently: these gaps were found, these were prioritised, this is what moved, and here is the capability the organisation has this quarter that it did not have last quarter. In June this blog made the business case for skills-based L&D in prospective terms — investment and expected return. The trend line is the retrospective version: the case proving itself, quarter by quarter, in the organisation’s own numbers.
And it compounds the honesty dividend this series keeps encountering. People are candid with a system they have watched do something useful with the truth; executives are no different. A leadership team that has seen capability visibly move stops treating the skills conversation as an L&D enthusiasm and starts treating it as operating data — which is, in the end, what all the careful work on frameworks and ratings was for.
Each layer this series has examined hands its result to the next. The framework decides what can be seen. The rating decides whether what is seen is real. The loop decides whether any of it changes a working week. The trajectory is where all three face judgement — the layer at which the programme itself is rated, by its own data, with no room for appeal. A snapshot can flatter or alarm, depending on the lighting. Only the trend can say whether the machinery underneath is actually working.
The practical instruction is short. Run the cycle again. Compare like with like, and be honest about where the line breaks. Read direction, not just position — and report movement, not activity. The still photograph was never the point. The film is.
7 September 2026
Last time, this series looked at the rating moment — the small act of judgement that turns a vocabulary into data, and what it takes to make the resulting number honest enough to plan on. Suppose, for a moment, that the work has been done. The framework is current, the managers have compared notes, the ratings arrive with evidence attached and a distribution a statistician would not laugh at. The organisation now knows something true about its own capability. This post is about the week after that — the quiet, dangerous pause in which most skills programmes stall.
Because there is a version of success that looks exactly like failure with better graphics. The assessment closes, the dashboard fills, leadership nods appreciatively at the heat map — and nothing in anyone’s working week changes. Six months later the data has gone stale, the next cycle asks everyone to rate again, and people who watched nothing happen the last time answer accordingly. The distance between an honest number and a different Tuesday is where the whole endeavour is decided.
The first mistake is treating the gap analysis as a to-do list. It is not; it is a list of facts. An honest assessment of a workforce will surface far more gaps than any organisation can close, and the instinct to respond to every red cell with a development action produces four hundred initiatives, none of them resourced, most of them abandoned by March. A gap only becomes a priority in the light of what the business is actually trying to do next — which is why the most useful hour after an assessment is spent reading the skills analytics with the strategy document open beside it, not admiring the coverage.
Three questions do most of the sorting. Which gaps sit directly in the path of something the organisation has committed to — a new service line, a system migration, a succession risk with a date on it? Which are cheap to close because the appetite is already there and only the opportunity is missing? And which are simply tolerable — real, visible, and not worth anyone’s quarter? Deciding on purpose not to act on a gap is a perfectly respectable outcome. It is the undeclared version, where everything is nominally a priority, that quietly teaches the workforce the numbers lead nowhere.
Once the shortlist exists, each entry has to survive a conversion: from an observation about a person to a commitment with that person. This is the point where the data either becomes a training goal — specific, owned, dated — or evaporates into the general aspiration to ‘develop our people’. The difference is almost embarrassingly concrete. ‘Improve stakeholder management’ is a wish. ‘Run the next two client steering sessions, with a debrief afterwards, and reassess in November’ is a goal an employee can actually walk toward, and can tell whether they have reached.
The route matters as much as the destination, and the route depends on will as much as skill. A gap a person is eager to close calls for stretch and opportunity; the same gap in someone privately relieved to leave it alone calls for a different conversation entirely. This is what a skill-will conversation is for — the point where the organisation’s shortlist meets the employee’s own ambitions, and the goal gets chosen with the person rather than assigned to them. Goals arrived at that way have a habit of surviving contact with a busy quarter. Goals issued from a spreadsheet do not.
Most development plans are written for an imaginary calendar — one with slack in it. The real one has delivery deadlines, a team member on leave, and a manager with forty minutes of discretionary attention. Development that actually happens is built for that calendar: small, adjacent to real work, and cheap to start. A stretch task inside an existing project. A fortnightly half-hour where the skill is used in front of someone who can comment. Structured feedback captured near the moment it happened, while the example is still warm, rather than reconstructed at review time from memory and generosity.
Courses still have their place — some capabilities genuinely begin in a classroom — but a course is an ingredient, not a plan. The pattern that works is unglamorous: a little formal input, followed quickly by a place to apply it, followed by someone noticing. An organisation that sequences those three things inside the working week will outrun one with a better content library every time, for the same reason this series gave when it looked at practice: capability is built between courses, in the work, or it is not built at all.
The last step is the one most programmes skip: going back. A goal that was named and dated deserves a reassessment against the same skill, by the same standard, at the next development review — not as an audit, but as the moment the loop closes. Did the number move? If it did, the organisation has just produced its most persuasive artefact: a gap that was found, worked on, and visibly closed. If it did not, that is worth knowing too, while the plan is still warm enough to adjust.
The previous post in this series argued that ratings which go nowhere are never wrong, because nothing ever tests them. The loop is the test. And it pays for itself in an unexpected currency: honesty. When a two becomes a three because of something everyone can point to, the next assessment cycle changes character — people tell the truth about their gaps to a system they have watched do something useful with the truth. Candour, it turns out, is not a cultural trait. It is a learned response to being taken seriously.
The output of a skills programme is not a report. It is a workforce that can do more this quarter than it could last quarter, and knows it — and every layer this series has examined exists in service of that. The framework decides what can be seen. The rating decides whether what is seen is real. This is the layer where either of those comes to matter: the short, deliberate path from an honest number to a named goal, a changed week, and a closed loop.
The gap between knowing and doing has been this series’ quiet theme from the beginning, and it applies to organisations just as it applies to the people in them. A truthful dataset that changes nothing is knowledge without practice. The remedy is the same at every scale: pick something specific, start it soon, and go back to check.
26 August 2026
Last time, this series looked at the framework underneath — the taxonomy that quietly decides what an organisation is able to see about its own capability. But a vocabulary only becomes data when somebody applies it to a person. Someone sits down, reads a skill description, thinks about a colleague or about themselves, and picks a number. That moment — small, routine, repeated thousands of times a year — is where every dashboard, every gap analysis and every workforce plan ultimately comes from.
It is also the least examined part of the whole apparatus. Organisations will spend a quarter designing a framework and an afternoon deciding how it gets rated, then treat what comes out as measurement rather than judgement. This post is about that moment: what a skills rating actually is, the two directions it reliably drifts in, and what makes one honest enough to plan on.
A skills rating looks like a measurement. It arrives as a number on a five-point scale, it averages neatly, and it renders beautifully in a bar chart. But it is not a measurement in any sense a scientist would recognise. It is a claim — one person’s judgement about capability, made at a particular moment, on whatever evidence happened to be within reach that week. The number is a compression of that judgement, and like every compression it discards the part that mattered most: the reason.
This is not an argument against rating people. Judgement is unavoidable and often rather good, and organisations have run on it for a century. It is an argument against forgetting what the number is made of. The moment a rating is treated as a fact, it stops being questioned — and an unquestioned claim will sit happily in a skills analytics view for two years, shaping succession shortlists and hiring briefs, long after the person it described has moved on. Everything downstream depends on remembering that each cell in the grid was once a sentence somebody said out loud.
Ratings drift in two directions, and most organisations manage both at once. The first is generosity. Marking a colleague down is unpleasant, the conversation is awkward, and a three feels like a criticism where a four feels like encouragement. Nobody sets out to inflate anything; each individual four is defensible and kindly meant. Two cycles later, ninety per cent of the workforce is proficient or better in almost everything, the gaps have vanished from the report, and the only honest conclusion available from the data is that no development is required anywhere.
The second direction is fear, and it usually shows up in self-assessment. If people suspect a rating will be read as a verdict — that admitting to a two will cost them a project, a promotion or a place on the list — they will rate themselves defensively, and the safest answer is always the middle. What makes a skill-will conversation work is precisely the opposite reflex: the willingness to say plainly that you are not there yet and would like to be. That candour is a function of consequence, not of scale design. Nobody tells the truth about a gap to an audience they expect to be judged by.
The most reliable cure for drift is not a better scale. Five points, four points, behavioural anchors, elaborate rubrics — organisations rework these endlessly and the ratings move very little, because the problem was never the granularity. The problem is that a number arrived without anything attached to it. Ask for one piece of evidence alongside every rating, a specific thing the person did and roughly when, and the whole exercise changes character. It becomes much harder to award a casual four when a sentence has to follow it.
This is also the argument for keeping assessment close to work rather than to the calendar. Structured evaluations and feedback gathered near the moment something happened are worth more than a retrospective sweep in November, when the only evidence anyone can recall is whatever occurred in October. The same applies at the other end: a rating that produces a specific training goal is being used, and things that get used get corrected. Ratings that go nowhere are never wrong, because nothing ever tests them.
Different managers mean different things by the same number, and no amount of guidance in a handbook fixes that. What does fix it, slowly, is managers comparing notes out loud. Half an hour with three or four peers, each explaining what a four in one specific skill looks like in their team, does more for consistency than a page of definitions — mostly because it surfaces the disagreement rather than averaging it away. The point of the session is not to force agreement on every rating. It is to ensure that when two people say four, they are describing roughly the same thing.
It is worth being clear about what this is not. Calibration in the skills sense is a shared-language exercise, not a forced distribution and not a ranking; the moment it starts allocating a fixed quota of fives it has become a performance mechanism and the honesty drains out of it within one cycle. Done well, it is a quiet, recurring habit attached to the development review cycle, and its main output is not a corrected spreadsheet but a group of managers who now describe capability in compatible terms.
There is a useful diagnostic available to anyone with access to the reporting. Look at the distribution rather than the average. A healthy skills dataset has spread — ones and twos in places, genuine variation between teams, some skills where the organisation is visibly weak. A dataset where almost everything clusters at three-and-a-half is not describing a uniformly competent workforce; it is describing a population of people who have worked out what the safe answer is. The shape of the data is usually more informative than anything in it.
Which leads somewhere slightly uncomfortable. A partial dataset that people believe is worth considerably more than a complete one they quietly discount, and coverage is the wrong first target. Better to have forty skills rated candidly, with evidence, by managers who have compared notes, than four hundred rated to fill in a form. The framework tells you what can be seen. The rating tells you whether what you are seeing is real — and no amount of downstream sophistication will rescue a number that nobody meant.
17 August 2026
Last time, this series looked at the line manager — the person in the middle who decides, week by week, whether a skills-based approach becomes a habit or an announcement. Every post before it has leaned on the same unexamined noun. The taxonomy. The framework. The skills list. It has been named in passing in almost every piece and never given a post of its own, which is odd, because it is the thing all the rest of it sits on.
A skills framework is the vocabulary an organisation uses to talk about capability. Get it right and everything downstream — the reviews, the analytics, the development plans — has something solid to point at. Get it wrong and you have built a very expensive machine for producing confident numbers about nothing in particular. This post is about the framework underneath: why it decays, what makes it decay faster, and what it takes to keep one honest.
Nobody gets excited about a skills taxonomy. It is a list. It has no launch event, no dashboard and no obvious owner, and in most organisations it is written once, early, by whoever had capacity that quarter and needed something in place before the platform went live. Then it disappears into the background — and quietly becomes the dependency for everything else the programme does.
This matters because the framework determines what can be seen. A skill that isn’t named cannot be assessed, developed, reported on, or hired for — it simply falls outside the field of view. When leaders complain that the skills analytics don’t reflect what the business actually needs, the problem is rarely the reporting layer. It is that the vocabulary underneath was written to describe an organisation that has since moved on, and the numbers are faithfully measuring the wrong things.
The most common failure is not a framework that is wrong. It is a framework that is enormous. The instinct when building one is completeness: consult every function, capture every specialism, make sure nobody’s work is missing from the list. It feels like rigour, and every individual addition is easy to justify. The result is eight hundred skills, each defensible on its own, and collectively unusable by a manager with forty minutes and a team of nine.
Size compounds in a way that is easy to miss at the design stage. A large framework takes longer to assess against, so assessments get rushed. Rushed assessments produce noisy data, and noisy data erodes the trust that makes anyone bother assessing carefully in the first place. Within two cycles you have a comprehensive taxonomy that nobody believes. A smaller framework — the fifty or so capabilities that genuinely differentiate performance in this business — will be assessed properly, which makes it more accurate than the exhaustive version it replaced. Completeness and usefulness pull in opposite directions here, and usefulness should win.
There is no prize for writing a skills framework from a blank page. Industry libraries, role standards and off-the-shelf taxonomies are a reasonable starting structure, and in 2026 an AI-assisted first draft will get you a workable skeleton in an afternoon rather than a quarter. Borrowing the scaffolding is sensible, and the organisations that insist on originality here mostly spend six months arriving at something close to the library they could have started from. The mistake is a different one: believing the borrowed version is finished.
Generic frameworks describe a generic organisation. They cannot know that ‘stakeholder management’ in your business mostly means holding a difficult conversation with a client’s finance team, or that ‘data literacy’ has a specific and demanding meaning in one division and a much lighter one in another. The work that cannot be outsourced is the local translation: writing the descriptions in the language people at your organisation actually use, with examples they recognise. That is also what makes a skill-will conversation land — an employee can only place themselves honestly against a description they can picture themselves doing.
Most taxonomies rot for an unglamorous reason: after launch, no one owns them. Roles change, tools change, a new service line appears, and the framework stays exactly as it was written eighteen months ago. It doesn’t break loudly, which is precisely the problem — there is no moment where the framework announces that it has fallen behind. It just drifts, a role at a time, until the gap between the words on the page and the work happening in the building is wide enough that people quietly stop taking it seriously.
The fix is boring and effective. Name an owner — a person, not a committee, with the authority to change the thing. Set a review cycle, twice a year is usually enough, where the framework is checked against what the business now actually does, with additions and, more importantly, retirements. The best signal for that review is friction reported from the front line: skills managers keep having to explain, ratings that cluster suspiciously at one level, capabilities that never once appear in a training goal. Those are the entries doing no work, and a framework improves as much by deleting as by adding.
A skills framework is not a reference document, and it is not an asset to be admired for its coverage. It is a working vocabulary, and the only real test of a working vocabulary is whether people reach for it when nobody is making them. If a manager opens it during a development review because it helps them say something true and specific, and the employee across the table recognises themselves in the description, it is doing its job — regardless of how many capabilities it leaves out.
Which suggests the right ambition is smaller than it first appears. Not a complete map of everything the organisation can do, but a short, current, locally-worded list of the things that matter, maintained by someone whose job it is to keep it true. Every layer this series has covered — the practice, the feedback, the analytics, the manager conversations — inherits its credibility from that list. It is worth the unglamorous effort of keeping it honest.
11 August 2026
Last time, this series looked at the first twelve months — the sequencing that keeps a skills-based L&D programme from quietly dying in the year it is stood up. That piece, like the ones before it, kept its eye on the design: the taxonomy, the evidence, the order of operations. But designs don’t run themselves. Somewhere between the strategy deck and the employee’s desk sits a person the series has mentioned in passing and never given its own post — the line manager. And that is an oversight worth correcting, because the manager is the point at which almost every good intention either becomes a habit or evaporates.
It is tempting to treat the manager as a channel — someone you brief once, hand a dashboard, and expect to relay the programme downward. But a skills-based approach asks the manager to do something far harder than relay: to change how they read their own team, week after week, in the ordinary run of work. This post is about that person in the middle — what the model actually demands of them, why their resistance is usually a signal rather than an obstacle, and what it takes to make the role survivable.
Every skills-based L&D programme has an org chart, and on that chart the line manager sits at the one node no strategy can route around. The platform can be excellent, the taxonomy sound, the executive sponsorship real — and none of it reaches an employee except through the person they report to. When a manager treats a skills conversation as a box to tick, the employee learns, correctly, that it doesn’t matter. When a manager takes it seriously, the same conversation becomes the most useful half-hour of the month.
This is why programmes that look identical on paper produce such different results in practice. The difference is rarely the tooling; it is whether the managers running the day-to-day chose to carry it. A skills model is not a system employees interact with directly so much as a set of expectations their manager either upholds or lets slide. Which means the manager is not one stakeholder among many — they are the single point of failure, and the single point of leverage.
It helps to be honest about the size of the ask. A manager already carries delivery targets, hiring, the endless triage of a real team. A skills-based approach now adds a second lens: alongside ‘did the work get done,’ they are asked to watch ‘what did this person get better at, and what should they build next.’ That is not a lighter version of the old job — it is a genuinely different discipline, closer to coaching than to supervision, and most managers were promoted for the supervision.
Concretely, it means holding a regular conversation about capability rather than only output; setting clear training goals an employee can actually walk toward; and giving honest, specific feedback about where someone stands against a skill instead of a vague reassurance at review time. Tools like a skill-will matrix make the thinking easier, but they don’t do the noticing. That still falls to a human who has to pay a kind of attention the previous version of the job never required.
When managers push back on a skills programme — ‘I don’t have time,’ ‘the ratings feel made up,’ ‘this is HR’s project, not mine’ — the reflex is to treat the resistance as a change-management problem to be overcome with communication. It is usually more accurate to treat it as feedback about the design. A manager who says they have no time is often telling you the programme has asked for the ceremony without removing anything to make room for it.
The manager who distrusts the ratings is frequently the most valuable person in the room, because they are the one who will actually be held to the numbers when a staffing decision goes sideways. Their scepticism is a demand for the evidence to be real before it is acted on — the same discipline that keeps skills analytics honest at the organisational level. Listening to the objection, rather than steamrolling it, is how a programme finds the places it was going to break anyway.
The fastest way to lose managers is to hand them a new obligation with no visible payoff to the team they are trying to run. If the skills work shows up as extra admin in service of a central dashboard, it will be done grudgingly and badly, and the data underneath it will be worth exactly what was put into it. The managers who lean in are the ones who can see the programme making their own week easier — a clearer read on who is ready for what, fewer surprises at review time, a defensible basis for the promotion they wanted to make anyway.
So the sequencing that matters most in year one is not technical, it is motivational: give the manager an early, concrete win before you give them the full apparatus. A development review that finally has real substance behind it, a case where the skills view caught a gap before it became a problem — these are what convert a compliant manager into a committed one. Reason first, task second. Do it the other way round and you get the task without the reason, which is another name for a programme nobody trusts.
There is a comforting myth that some managers are simply ‘people developers’ by nature and the rest are not, and that the programme’s job is to find the naturals. It is a myth worth dropping. The managers who run skills conversations well are, overwhelmingly, the ones who were shown how — given a simple rhythm to follow, a little language for talking about capability, and the cover to spend time on it without feeling they were stealing it from ‘real’ work.
That is the quiet implication of everything above: if the manager in the middle decides the programme’s fate, then equipping that manager is not a side-project of the rollout — it is the rollout. The taxonomy and the analytics and the platform all matter, but they are scaffolding around a single human act repeated across a workforce: a manager paying attention to how their people are growing, and doing something about it. Get that person right and the rest compounds. Get everything else right and neglect them, and you have built a very sophisticated machine with no one to turn it on.
03 August 2026
Last time, this series took on the machinery — what artificial intelligence genuinely does for skills-based L&D, and which decisions have to stay human. That post, and the ten before it, shared a quiet assumption: that the programme already exists. The series has spent four months explaining why skills-based L&D matters and what it looks like when it works. It has said almost nothing about the part that keeps people up at night — how you actually stand the thing up from a standing start.
That silence is telling, because the first year is where most of these programmes quietly die. Not from a lack of conviction — the business case usually clears — but from a handful of avoidable mistakes made in the opening months, when the temptation is to do everything at once and the reality is that almost nothing is ready. This week, the series changes register: from why and what to how, and specifically to what the first twelve months actually demand of the team standing at the starting line.
The instinct, once the business case lands, is to treat the first year as a build: choose a platform, define the skills, roll it out, and start reporting. Framed that way, the work looks like a checklist, and checklists get done in parallel. But skills-based L&D is not a system you install; it is a set of habits a workforce has to adopt, and habits cannot be installed in parallel. They have to be sequenced, because each one depends on the trust built by the last.
That reframing is the whole game. The organisations that stall in year one are almost never the ones that lacked resources or executive air cover. They are the ones that tried to do everything the model eventually requires — the full taxonomy, the analytics, the performance linkage — before anyone had a reason to believe the first small piece. Sequencing is not a project-management nicety here. It is the difference between a programme that compounds and one that is quietly shelved by March.
The first and most common way the year goes wrong is the taxonomy. It is intellectually satisfying to map every skill in the organisation before doing anything with it, and it is a trap. A comprehensive framework built in a vacuum takes months, ages before it ships, and lands as a monument nobody asked for — thousands of skills, no evidence behind any of them, and a workforce that experiences the whole thing as a survey.
The teams that get through the first year start narrow and deliberately incomplete. One function, one set of roles, the fifteen or twenty skills that actually decide whether someone is good at the job. A smaller map that is used beats a complete one that is admired, because the point of the taxonomy was never accuracy for its own sake — it was to give people clear training goals they can actually walk toward. A framework earns the right to grow by being useful first.
The second failure mode is the mirror image of the first: rushing to measurement before there is anything trustworthy to measure. A skills dashboard stood up in month two, populated by self-assessments nobody has any incentive to make honest, produces a confident-looking picture that is quietly wrong. Worse, it teaches leaders to make decisions on it — and the first time a staffing call goes visibly sideways because the data was fiction, the whole programme loses the room.
The discipline here is to let analytics follow evidence rather than lead it. Numbers earn authority slowly, by being grounded in something real — a completed assessment, an observed piece of work, a manager and employee who looked at the same picture and agreed it was true. Reporting that arrives before that groundwork is not insight; it is decoration that will eventually be believed, which is more dangerous than being ignored.
If the first year is a sequencing problem, the first move is not a platform decision or a taxonomy workshop. It is a question, asked of the people whose skills are about to be measured: where do you actually want to go? Capturing career aspirations alongside skill and will before deciding what anyone should learn does something no framework can — it makes the whole exercise feel done with people rather than to them, which is the difference between honest data and polished data.
Starting with the person also quietly solves the adoption problem the rest of the year will otherwise fight. When the first thing someone experiences is a conversation about their own direction, the profile that follows reads as a tool for getting there, not a file being kept on them. Get that opening beat right and the later steps — assessment, feedback, measurement — inherit a reservoir of goodwill. Get it wrong, and every one of them is uphill.
The single most useful thing a team can do in the first year is to close one complete loop, small, before scaling anything. Pick a handful of roles. Capture aspirations, agree a couple of priority skills, run one round of real practice, and let evaluations and feedback put honest signal back in front of the person — all the way through to a development review that both sides can actually have because they are looking at the same evidence. One loop that people believe is worth more than a full rollout they merely tolerate.
A closed loop is proof, and proof is what buys the second year. It gives you a real story to tell the next function, a handful of people who will vouch for the thing because it demonstrably worked for them, and — crucially — a set of assumptions tested against reality before they were baked into a platform-wide rollout. Scale the loop, not the ambition. The ambition was never in doubt; the loop is what makes it survivable.
There is a simple test for whether a first year has gone well, and it has nothing to do with how many skills were mapped or how polished the dashboard looks. It is this: at the twelve-month mark, do the people who went through it want more of it? If a manager is asking to bring the next team in, if an employee is keeping their profile current without being chased, the sequence worked — trust was built in the right order, and the model has somewhere to grow.
If instead the honest answer is that profiles were filled in once and abandoned, that the dashboard is impressive and unbelieved, that the taxonomy is complete and unused — the failure was almost never ambition. It was sequence: too much attempted, too fast, before anyone had a reason to trust the first small piece. Skills-based L&D is not won in the boardroom that approves it or the platform that hosts it. It is won, or lost, in the patience of the first year. Start small enough to be believed, and let being believed earn you the rest.
16 July 2026
Last time, this series changed seats and looked at skills-based L&D from the employee’s side of the profile. Sitting there, something became hard to ignore: almost everything the series has promised — profiles that stay current, feedback that compounds, career paths that assemble themselves from skills data — quietly assumes a layer of machinery doing work no L&D team could do by hand. That machinery has a name everyone is either overusing or avoiding: AI.
So this week the series asks the question directly. Not “is AI coming for L&D” — that debate has been stale for two years — but something more useful: in a skills-based operating model, what is AI actually good for, what can it not carry, and how should a sensible L&D team divide the labour in 2026?
Go back through this series and count the moving parts a skills-based organisation depends on: a taxonomy kept current as roles evolve, thousands of profiles refreshed as people learn, gaps mapped against strategy, content matched to gaps, practice generated at the right difficulty, feedback captured in the flow of work. Now imagine maintaining all of that manually. The honest answer is that nobody does — and nobody ever intended to.
This is the assumption hiding inside every skills-based pitch deck: the model only scales because inference, matching, and generation are automated. That is not a criticism. But it does mean the AI question is not an optional add-on to a skills strategy. It is the load-bearing wall, and it deserves the same scrutiny the series has given taxonomies, measurement, and the business case.
Three jobs stand out, and they happen to be the three that were breaking L&D teams before automation arrived. The first is inference — reading the exhaust of everyday work and learning activity and suggesting what it says about capability, so profiles drift towards accuracy instead of towards fiction. Paired with analytics that surface the patterns, this is what turns a skills database from a survey snapshot into something closer to a living picture.
The second is matching: person to gap, gap to content, content to moment. This is pure pattern-work across more variables than any human curator can hold, and it is where the “course catalogue” posts of this series quietly get their answer. The third is generation. AI-assisted course creation has collapsed the cost of producing a first draft of learning content from weeks to hours — which is precisely what makes the skills-first content library, built against the taxonomy rather than the calendar, economically possible for teams that are not enterprise-sized.
Now the other column of the ledger. AI cannot decide what capability the business actually needs next year — that is strategy, and outsourcing it to a model means optimising towards last year’s patterns. It cannot supply the trust this series wrote about from the employee’s seat; an inference engine that quietly rescores people’s profiles without explanation is the fastest way to lose the honesty that makes the data worth having.
And it cannot replace judgement at the moments that matter. A model can flag that someone’s evidence looks thin; only a human conversation — the kind that structured evaluations and feedback exist to hold — can establish what is actually true and what should happen next. The pattern across every failure story of the past two years is the same: the organisation automated the judgement and kept the admin, when it should have done the opposite.
Put the two columns together and a clean rule emerges: let AI propose, let people decide. AI drafts the profile update; the person confirms it. AI suggests the learning path against a training goal; the employee and manager commit to it. AI surfaces who is ready for a stretch move; the career conversation decides whether the person wants it. Every “propose” saves hours; every “decide” protects trust.
The teams getting this right in 2026 are noticeably unglamorous about it. They talk less about AI strategy and more about which specific decisions stay human. That list — written down, shared with employees, honoured in practice — is doing more for adoption than any capability the technology itself has shipped this year.
So when the next AI-for-skills demo lands in your inbox, skip the feature tour and ask one question: for each thing this automates, who was doing it before, and who checks it now? If the answer to the first half is “nobody — it simply wasn’t being done,” that is genuine capacity you are buying. If the answer to the second half is “nobody,” walk away.
Everything this series has argued — the taxonomy, the measurement, the practice loops, the feedback rhythm, the employee’s ownership of their own picture — depends on machinery that works and people who trust it. In 2026 you can, at last, have both. But only in that order.
06 July 2026
Last time, this series watched L&D reinvent itself as the steward of the capability picture the whole business now runs on. That closed a loop this series has been drawing since April — managers, taxonomies, measurement, content, practice, feedback, the business case, the organisation, the function itself. Read the list back and something stands out: every one of those posts looked at skills-based L&D from above. The one perspective the series has never taken is the one belonging to the person whose skills are actually in the data.
That gap matters more than it looks, because in 2026 the success of a skills-based organisation is not decided in the boardroom that approved it. It is decided at thousands of individual desks, where people quietly choose whether to keep their profile honest, whether to engage with the feedback, and whether to believe that any of this is being done for them rather than to them. This week, the series changes seats.
To the organisation, a skills profile is infrastructure — a row in the dataset that feeds planning, staffing, and development decisions. To the employee, it is something far more personal: a standing claim about who they are and what they are worth, held and read by their employer. Whether that claim feels like an asset or an exposure is not a soft question about sentiment. It is the single biggest determinant of whether the data underneath the whole strategy can be trusted.
The mechanism is simple. Skills data is only as good as people’s willingness to be honest in it, and honesty follows consequences. In an organisation where admitting a gap reliably produces support — a course, a stretch assignment, a coaching conversation — people tell the truth. In one where a visible gap quietly costs you the interesting project or the promotion shortlist, people polish. No taxonomy, platform, or dashboard survives a workforce that has learned to polish.
The old deal was passive: learning was something the organisation did to you. You were enrolled, you completed, you were certified, and the record of it belonged to a system you never looked at. The skills-based deal is different in kind, not just degree. A living profile shows you where you stand; visible gaps show you what is between you and the role you want; and clear training goals turn that distance into something you can actually walk.
The best implementations go a step further and ask the person where they want to go before deciding what they should learn, capturing career aspirations alongside skill and will so that development is negotiated rather than assigned. That single change — from consumer of a curriculum to owner of a direction — is what employees mean when they say skills-based working finally feels different from the annual training calendar it replaced.
There is a harder edge to all this, and pretending otherwise is how programmes lose the room. The same data that powers development can, from the employee’s seat, look like surveillance. Who sees my gaps? Does this number follow me into pay conversations? Was I passed over because of something in a system I have never been shown? These questions rarely appear on an implementation roadmap, and they are the ones the workforce is actually asking.
The organisations that answer them well do it with transparency rather than reassurance. They put plain rules around who sees what and what it is used for, and — crucially — they make the data flow back to the person, not just upward. When evaluations and feedback reach the employee as insight they can act on, and the analytics visibly serve their development rather than only management reporting, measurement starts to read as investment. Opacity, not measurement, is what people distrust.
Get the trust question right and the payoff lands squarely on the employee’s side of the table. Career conversations stop being exercises in advocacy — who noticed you, who will vouch for you — and start being grounded in shared evidence. A development review built on a skills profile is a conversation two people can actually have, because both are looking at the same picture rather than trading impressions.
There is a fairness dividend here that deserves more airtime than it gets. When capability is the currency, the quiet performer in the regional office competes on the same terms as the confident voice in head office. Internal moves stop depending on being in the right meeting. That is not a soft benefit; it is the employee-side answer to why any of this is worth the discomfort of being measured at all.
The honest catch is that handing employees ownership means asking them to do something with it. A profile only stays true if the person keeps it true; feedback only compounds if it is engaged with; a visible gap only closes if somebody chooses to close it. Organisations that frame skills-based working as pure empowerment set themselves up for the quiet disappointment of profiles that were filled in once, enthusiastically, and never touched again.
Nor can the organisation use ownership as an exit. “It’s your career” is a partnership when it comes with time, support, and honest data — and an abdication when it does not. The test worth applying in 2026 is disarmingly simple: would the person being measured choose this system for themselves if it were optional? Everything this series has covered — the taxonomy, the measurement, the manager conversations, the business case — ultimately exists to make the answer yes. Build it so that it is.