Talent on Demand

1. 1Frame every talent decision as risk management: quantify the cost of overshooting (surplus, layoffs, turnover of developed people to competitors) against the cost of undershooting (lost business opportunity, reactive outside hiring at premium prices).
2. 2Reject the two dysfunctional default camps: the do-nothing/reactive outside-hiring approach, and the 1950s bureaucratic forecast-and-succession-planning approach. Both fail because you cannot forecast away uncertainty.
3. 3Adopt 'make and buy,' not 'make versus buy': set internal development to deliberately undershoot your best demand estimate, then close the gap with outside hires so you never build an inventory of surplus talent.
4. 4Cost the retention leak explicitly. Use cost-accounting to show what a developed employee walking to a rival takes with them — that loss usually dwarfs the premium you pay for an outside hire, which is why overshooting now costs more than undershooting.
5. 5Treat the internal pipeline like a supply chain: attack bottlenecks that block advancement, speed processing time, and improve demand forecasts to shrink the mismatch on both ends.
6. 6Recalculate the balance per organization: the make/buy split is a strategy judgment that varies by the nature of the jobs, the work performed, and internal capabilities — do not apply one ratio everywhere.
7. 7Batch and stagger development intake to cut fixed costs (as in hiring recent grads in spring, fall, and winter), so you carry less pre-built inventory against an uncertain forecast.

Internal development collapsed in the 1970s as a live case of overshooting: business forecasting failed to predict the economic downturn, but talent pipelines kept turning out managers under the old assumption of booming corporate growth. Combined with no-layoff policies for white-collar workers, this produced a bloated surplus — an inventory of talent the business no longer needed. The steep 1981 recession then forced white-collar layoffs, ended lifetime employment, and gutted the development staff and programs that had created the talent. Firms swung to the opposite failure: outside hiring worked through the early 1990s only because it drew down the pool of already-laid-off talent. Once that pool emptied, hiring pulled talent from competitors, creating a retention problem — and employers cut development further as they watched developed people walk out the door. The lesson: forecasting to full demand under uncertainty guarantees you lurch between surplus and shortfall.

• If your demand forecast is uncertain (which it now always is), then undershoot the estimate on internal supply and plan to buy the remainder outside.
• If the cost of losing a developed employee to a competitor exceeds the premium on an outside hire, then bias toward buying rather than over-building internal talent.
• If a role requires competencies that don't exist internally and you are changing strategy or direction, then hire outside for it rather than trying to develop it.
• If you cannot demonstrate that a development program pays off against its cost, then treat that cost as real exposure — costly programs that develop talent for rivals are still failures.
• If you are relying solely on either outside hiring or 1950s succession planning, then you are in a dysfunctional camp; balance the two.

• Forecasting to full expected demand and building internal supply to match it. Tell: pipelines keep producing under old growth assumptions while the business shrinks, and you end up with a talent inventory and eventual layoffs.
• Defaulting to 'we'll hire someone' for every vacancy. Tell: outside hiring lags the actual need, costs escalate, and results are unreliable — the CA pattern where the reflex answer to any opening was to buy.
• Cutting development because employees leave, then discovering you must hire everyone outside. Tell: retention problems and rising reactive-hiring bills feed each other in a loop.
• Ignoring retention costs because you can't account for them. Tell: 'people are our most important asset' rhetoric coexists with development treated purely as a cost center to be squeezed.
• Assuming one make/buy ratio fits the whole organization. Tell: you apply the same balance to easily-bought roles and hard-to-develop, company-specific roles alike, and mismatch persists in pockets.

• ▸Role/skill: the position or competency being planned for
• ▸Demand forecast (best estimate): expected headcount need and its confidence level
• ▸Undershoot cost: lost business opportunity + reactive outside-hiring premium if short
• ▸Overshoot cost: surplus salary carry + layoff/restructuring cost + turnover of underutilized people
• ▸Retention leak cost: value of development investment lost if this developed person leaves for a rival
• ▸Outside-hire availability: how readily this skill can be bought on the open market
• ▸Make/buy split: planned % developed internally (set below forecast) vs. % bought to fill the gap
• ▸Pipeline bottleneck check: where advancement is blocked or processing time is slow for this role
• ▸Review trigger: the demand-change signal that forces recalculation of the split

1. 1Measure your actual managerial turnover rate as the entering input. At 10 percent annual turnover, model the compounding: you lose roughly half your development pipeline within five years. Treat any multi-year program against this decay curve, not against the enrollment headcount.
2. 2Reframe every development investment as a recoupable expense with a payback horizon. Ask whether the employee stays long enough to return the cost. If the average tenure post-program is shorter than the payback period, you are subsidizing your competitors.
3. 3Shorten the time employees spend in development so shorter, more accurate forecasts govern the investment and less time is exposed to attrition risk. Split long curricula into a common-skills block followed by a function-specific block, so the second forecast reaches only half as far.
4. 4Split intake cohorts across time (June and September rather than all in June) so fewer developmental roles are needed at once and exposure to departure is staggered rather than concentrated.
5. 5Choose a make/buy mix deliberately: build internally the skills where retention is achievable and payback is realistic; buy from the outside where turnover would destroy the investment before it pays. Do not treat outside hiring as the default just because retention feels hard.
6. 6Test the aggregate logic before adopting a 'retain-and-recruit' goal. Getting better at both poaching competitors' talent and holding your own is impossible across the whole market at once; decide which side of that trade you actually occupy.
7. 7Track the 'back door': monitor whether experienced employees are leaving for jobs that look like the ones you are simultaneously filling with expensive outside hires. That symmetry is the signal that your retention problem is inflating your hiring cost.

A CEO in the medical device industry is handed a management development program proposal by his head of HR and rejects it with, 'Why should we develop people when our competitors are willing to do it for us?' Read through the retention-risk lens, his reasoning is coherent given the environment: in an open labor market, an employee cultivated at his cost can walk out the back door to a rival, so his development spend becomes a subsidy to competitors. But the CEO's escape hatch—buying trained talent from outside—hits the same wall this book documents: by the late 1990s the pool of available talent was depleted, outside hires failed at high rates especially at the executive level, and firms lost their own experienced people even as they hired experienced replacements. So the choice is not develop-or-poach; both sides face the retention leak. The disciplined answer is to develop but shorten and stagger the investment so the payback lands before the average departure, converting an unrecoverable subsidy into a recoverable expense.

• If your program's payback period exceeds average post-program tenure, then shorten the program or buy the skill instead of building it.
• If turnover in the pipeline exceeds roughly 10 percent annually, then treat multi-year succession programs as sieves, not pipelines, and reforecast against the decay.
• If you can defer a forecast by splitting a curriculum into common then function-specific blocks, then do so—the later forecast is shorter and more accurate and cuts exposure time.
• If you are losing experienced people out the back door to jobs resembling those you fill with outside hires, then invest in retention on those roles rather than escalating external hiring spend.
• If your stated strategy is to both out-recruit rivals and out-retain your own people across the board, then reject it as aggregately impossible and pick a defensible position.

• Treating development spend as a fixed cost rather than a recoupable investment; the tell is nobody can state the payback period or the tenure needed to break even.
• Counting your pipeline by enrollment headcount instead of survival rate; the tell is surprise when only half the class remains after five years.
• Defaulting to outside hiring as the retention-proof solution; the tell is escalating recruiting costs alongside continued experienced departures and high executive-hire failure rates.
• Cutting development staff during downturns to save cost, then having no capacity to build talent when the market tightens; the tell is a firm that 'lurches from surpluses to shortfalls and back.'
• Setting a goal to simultaneously poach better and retain better without acknowledging the market can't let everyone win both; the tell is the same slogan appearing at every competitor while none actually improves.

• ▸Role/program name: the development investment being evaluated.
• ▸Annual voluntary turnover rate for this population: the departure percentage driving pipeline decay.
• ▸Program length (months): total time from intake to completion.
• ▸Total per-candidate development cost: full spend to develop one person.
• ▸Estimated payback period (months): time on the job needed to recover that cost.
• ▸Average post-program tenure (months): how long developed employees typically stay.
• ▸Recoup verdict (payback vs tenure): 'covered' if tenure exceeds payback, 'exposed' if not.
• ▸Shortening/splitting action: how you'll cut or stagger the program to reduce exposure.
• ▸Make-vs-buy call: build internally, buy externally, or blend—with the reason tied to the verdict.

1. 1Replace the chess-master model with an internal job board where employees bid on posted openings. This is where 96 percent of surveyed companies have moved, and it makes internal moves the default path for changing jobs without leaving.
2. 2Post every internal vacancy openly and let virtually all internal moves flow through the board, so employees see viable advancement without scanning the external market. Visible options are what convert a would-be leaver into an internal applicant.
3. 3Give employees information and choice over their own moves. When you turn career management over to employees, supply the data (open roles, required experience, progression paths) they need to make matches for themselves.
4. 4Offer voluntary developmental projects with the leadership team, sometimes restricted to work outside the employee's current function to broaden experience. These give employees access to company leaders and advancement they otherwise wouldn't get.
5. 5Spot talent and potential early and give those employees the chance to advance faster than they otherwise might. Speeding a capable person to a higher-value job is both a retention and a value move; identifying who is ready is the line manager's primary task.
6. 6Balance employee interests against organizational needs rather than dictating moves. In tight labor markets, forcing assignments backfires because good candidates who dislike the placement can secure an offer more to their liking and leave.
7. 7Let employees share the cost of their own development by taking on learning projects voluntarily in addition to regular work, since they are now the main beneficiaries able to cash in on the open market.

Consider the shift the passages describe away from the AT&T-era chess master. In the Organization Man period, the company decided which candidates got which experience to meet its own longer-term needs, and employees had little or no choice; refusing a new position meant career advancement stopped. That worked only while labor markets were loose. Once markets tightened, good candidates who could easily secure a job offer more to their liking walked when handed placements they disliked. In response, virtually all companies, 96 percent in a recent survey, replaced the chess master with internal job boards where employees bid on posted openings. The employer's conclusion was explicit: if employees want a new job, at least make it easier for them to find one inside the organization. That single change hands career control to the employee and keeps the move, and the prior development investment, inside the firm.

• If good employees are leaving to take opportunities better than what they can find internally, then make internal opportunities visible and biddable before they look outside.
• If labor markets are tight for a role, then do not dictate assignments; forcing placements on candidates who can secure a preferable offer elsewhere drives them out.
• If an employee shows early potential, then create a faster advancement path and give access to leadership projects rather than making them wait their turn.
• If you want employees to fund their own growth, then offer voluntary projects outside their current function that carry no pay increase but broaden experience and give leadership exposure.
• If you have given up on centralized career planning, then you must equip employees with the information and internal job market to manage their own careers, or agency collapses into attrition.

• Keeping the chess-master model where the company decides who moves where and employees have no choice. The tell: capable people quit for outside offers the moment markets tighten.
• Abandoning career planning without providing the internal market to replace it, so you turn the problem over to employees but give them no viable internal paths. The tell: employees find their next role only by leaving.
• Treating internal job boards as a filing formality rather than the real channel for moves. The tell: postings exist but promotions still happen through back-channel sponsorship.
• Assuming retention is about pay when the main reason good employees leave is to take opportunities better than what they can find in their current organization. The tell: raises fail to stop exits.
• Investing heavily to develop talent then losing the investment because there is no internal opportunity to retain the person. The tell: your best-trained people cash in on the open market with rivals.

• ▸Open roles posted internally: list every current vacancy visible to employees on the internal board.
• ▸Bid channel: state whether employees can actually apply for these openings themselves.
• ▸Share of moves via board: estimate what percent of internal moves run through open bidding vs. manager assignment.
• ▸Voluntary development projects available: name leadership-team or cross-functional projects employees can volunteer for.
• ▸Early-potential flags: identify employees ready to advance faster and the next role each could reach.
• ▸Information provided to employees: list what path, requirement, and opening data employees can access unaided.
• ▸Forced-assignment risk: note roles where the company still dictates moves and could trigger exits.
• ▸Recent regretted exits: record who left for a better external opportunity and whether an internal option existed.
• ▸Cost-sharing options: state which developmental experiences employees can take on voluntarily without a pay increase.

1. 1Frame talent as a matching problem: quantify not just how much talent you need, but the cost of getting it wrong on both sides — the cost of a shortfall (missed projects, forced promotion of unready people, emergency outside hiring at a premium) versus the cost of a surplus (idle bench, layoffs, restructuring charges).
2. 2For each critical role or skill, compute the combined cost stack: development cost, external hiring cost, cost of vacancy time, cost of holding surplus, and cost of unwanted turnover. Total these rather than optimizing any single line.
3. 3Choose a make/buy mix that minimizes the combined cost under uncertainty — don't overbuild internal pipelines to guarantee zero shortfall, because overshooting the forecast is itself a cost driver (this is what bloated organizations in the 1970s).
4. 4Stop treating every plan as free: because commitments carry costs and long-horizon forecasts are usually wrong, shorten planning horizons and keep talent commitments flexible so wrong bets stay cheap.
5. 5Make internal development justify its cost against the outside-hiring alternative, exactly as you would evaluate any other investment — if you cultivate talent that walks to rivals, you have paid the development cost and lost the asset.
6. 6Measure the failure rate and quality of outside hires (especially at executive level) and price that unreliability into any buy decision rather than assuming external hiring is the cheap default.
7. 7Reduce unwanted turnover cost by recognizing you cannot simultaneously recruit rivals' talent and retain your own at the aggregate level — plug the back-door leak before scaling front-door hiring.
8. 8Borrow supply-chain logic: manage uncertainty instead of pretending to forecast it away, using shorter cycles, buffers sized to cost-of-error, and just-in-time replenishment where the role allows it.

Consider the medical-device CEO in the passage who killed a proposed management-development program by asking, 'Why should we develop people when our competitors are willing to do it for us?' His logic is cost-effective on its face: development is expensive and uncertain, so let rivals absorb it and buy the finished talent. But the passage shows where this breaks: by the late 1990s the pool of available talent was depleted, outside hires had high failure rates and hard-to-measure quality, and firms lost their own experienced people 'out the back door' to jobs identical to those they were filling externally. So the CEO's buy-only strategy quietly accumulates vacancy costs, premium hiring costs, and unwanted-turnover costs that a blended make/buy approach would have lowered. The cost-effective answer is not pure buy or pure make but the mix that minimizes the combined cost stack under his actual uncertainty.

• If the cost of a talent shortfall (lost projects, forced under-qualified promotions) exceeds the cost of a modest surplus, then build a small buffer internally — but only up to the point where surplus-holding cost equals shortfall cost.
• If you cannot forecast the business demand accurately, then do not commit to a long-horizon internal pipeline, because talent forecasts are only as good as the business forecasts beneath them.
• If a skill is common in the external market and quality is verifiable, then buy; if it is unique to your organization or quality is hard to assess in outsiders, then make.
• If your own experienced employees are leaving to rivals for roles like the ones you're hiring for, then fix retention before increasing external recruiting spend — recruiting and retaining at the aggregate level cancel each other out.
• If a development practice cannot demonstrate value against the outside-hiring alternative, then treat it as an unjustified cost like any other failed investment.

• Doing nothing — no anticipation, relying entirely on reactive outside hiring. The tell: every quit is treated as a calamity and every new skill need as a crisis (true of roughly two-thirds of U.S. employers).
• Reviving 1950s bureaucratic forecasting and succession planning. The tell: elaborate multi-year plans that are both inaccurate and costly because the predictable business environment they assume no longer exists.
• Overshooting talent estimates to guarantee zero shortfall. The tell: a deep bench plus no-layoff policies that bloat the organization until a recession forces layoffs (the 1970s-1981 pattern).
• Assuming outside hiring is the cheap default. The tell: high executive-hire failure rates and losing your own cultivated talent out the back door while paying premiums for replacements.
• Optimizing one cost line in isolation. The tell: cutting development to save money while vacancy, emergency-hiring, and turnover costs quietly rise elsewhere — the total cost stack is never summed.

• ▸Role/skill: name the hard-to-fill position under review
• ▸Demand uncertainty: high / medium / low confidence in the business forecast driving this need
• ▸Shortfall cost: dollar or impact estimate of not having the talent when needed
• ▸Surplus cost: cost of holding this talent idle on the bench if demand doesn't materialize
• ▸Make cost: development, assignment, and program cost to grow it internally
• ▸Buy cost: external hiring premium plus expected failure/rework rate for outside hires
• ▸Turnover risk: likelihood and replacement cost of this talent leaving to a rival
• ▸Chosen make/buy mix: the ratio that minimizes the summed costs above under stated uncertainty
• ▸Planning horizon: shortest commitment window that still meets the need, to keep wrong bets cheap

1. 1Reframe the goal: stop treating talent management as an end (development, succession benchmarks, 5% turnover targets) and tie every talent decision to the organization's objective of making money, weighing costs against benefits.
2. 2Diagnose your current camp: determine whether you 'do nothing' (two-thirds of U.S. employers do no talent planning) or over-invest in internal pipelines built on long-range forecasts—both were tailored to a predictable environment that no longer exists.
3. 3Shorten and decentralize the forecast: follow EDS's model—have each business leader project demand in terms of competencies, roles, and regions over a rolling twelve-month window rather than attempting multi-year plans.
4. 4Model options, not just gaps: like Citibank calculating the human capital cost of opening a new branch versus expanding existing ones, quantify the talent implications of each strategic move so talent feeds strategy formation instead of lagging it.
5. 5Make the make-versus-buy call explicitly: decide which skills to develop internally and which to source externally, recognizing outside hiring is reactive, costly, lags the need, and leaks talent to rivals in an open labor market.
6. 6Manage talent as business risk: because people are the biggest cost component and source of core competencies, treat supply/demand mismatches (surpluses forcing layoffs, shortfalls forcing crises) as risks to hedge rather than crises to react to.
7. 7Build an internal market to move existing talent to where new strategies need it, so you can adjust and respond in the strategy arena quickly without waiting on external hires.

EDS, an IT consulting and outsourcing firm, illustrates the shift from paralysis to responsive planning. Before 2004 it had no central talent planning; business grew so fast the company simply hired as quickly as it could, and executives believed information-systems technology changed too fast to bother forecasting talent needs. In 2006 EDS installed a deliberately simple forecast extending only twelve months, in which each business leader outlines expected demand in competencies, roles, and regions. Because at EDS people are the product, demand translates cleanly—knowing a project requires 100,000 programming hours per month maps directly to how many people across job titles are needed. The real work is understanding current supply, which the staff can assess internally. This short-horizon, option-aware method lets EDS match talent to shifting project demand without pretending it can predict years ahead.

• If you cannot forecast business needs accurately in an uncertain market, then shorten the planning horizon (e.g., twelve months, per EDS) rather than extending detailed multi-year pipelines.
• If a skill is core to your competencies and predictable in demand, then develop it internally; if demand is volatile or the skill is commodity, then buy or outsource to convert fixed labor cost to variable.
• If you are growing through acquisitions, then defer heavy internal development—you don't yet know which employees you'll inherit (as the Wall Street HR executive observed).
• If competitors are poaching your developed talent, then don't respond by abandoning development wholesale; instead price the leak and hedge, because 'let competitors develop them' fails at the aggregate level.
• If a talent decision cannot be linked to the organization's money-making objective, then it is a tactical benchmark, not strategic—drop or redesign it.

• Doing nothing: no planning, so every quit is a calamity and every new skill demand is a crisis—the tell is 'panic' over predictable events like baby-boomer retirements.
• Forecasting on stale boom assumptions: pipelines keep producing talent after the market turns (as in the 1970s), bloating the organization until a recession forces white-collar layoffs—the tell is excess supply plus no-layoff policies.
• Treating strategic planning as a calendar exercise: Rivkin's point that plans become consensus-negotiation 'time to generate a plan' rituals rather than direction—the tell is plans that go out the window the moment a real opportunity or shock appears.
• Defaulting to outside hiring: reactive, kicks in after problems occur, disappoints on cost and lags the need—the tell is rising retention problems as you poach and get poached simultaneously.
• Cutting development because rivals will train for you: the CEO who asked 'why develop people when competitors will do it for us?'—the tell is a purely cost-driven view that ignores that this is impossible for everyone at once.

• ▸Business move/strategy: the specific decision or opportunity driving talent demand
• ▸Competencies needed: skills/roles this move requires
• ▸Roles & regions: where and in what jobs the talent is needed
• ▸Demand quantity: headcount/hours by title over the next 12 months
• ▸Current internal supply: people/skills already available to redeploy
• ▸Gap: demand minus supply, by role
• ▸Make vs. buy call: develop internally / hire outside / outsource, with reason
• ▸Cost & risk: cost of each option plus leak/lag risk if it fails
• ▸Feedback to strategy: what this talent picture tells decision-makers about feasibility

1. 1Frame every talent need as a make-versus-buy decision: for each role, ask whether internal development ('make') or external hiring ('buy') is the lower-risk, lower-cost path given your inability to forecast demand accurately.
2. 2Quantify the two error costs separately for each role: the cost of a talent shortfall (having too little talent, forcing expensive, slow, unreliable outside hiring that lags the need) versus the cost of a talent surplus (over-developing employees who then leave for rivals, wasting the investment, or triggering layoffs and broken career paths).
3. 3Set internal development targets deliberately below expected demand for roles where surplus is costly, then plan to close the residual gap with external hires — treat the shortfall as intentional, not a failure.
4. 4Consolidate decentralized development programs into a single pooled program so that unanticipated demand in one division and an unanticipated shortfall in another cancel out, reducing overall volatility the way a stock portfolio reduces the risk of holding one asset.
5. 5Build talent pools of broad, general competencies rather than developing people for narrow specialized jobs; allocate them to actual vacancies as they appear and close residual fit gaps with just-in-time training and coaching.
6. 6Shorten every forecast: break long development programs into discrete parts, each with its own updated forecast, because shorter forecasts always beat longer ones.
7. 7Convert fixed labor costs to variable where surplus risk is high — use outsourcing, staffing agencies, or leased employees (even at the executive level) for roles you cannot reliably plan around.

Take a large decentralized organization where each division runs its own program to develop managers for expected demand over the next five years. The odds of any single division hitting its five-year target accurately are very poor, so each program misses — and because the divisions face different business demands, they miss in different ways: some end up with a surplus of talent, others with a shortfall. Operated separately, one division scrambles to hire externally at prohibitive cost while another carries excess managers it may eventually have to lay off. Consolidate those programs into one pooled program, and the unanticipated demand in one part cancels the unanticipated shortfall in another, just as a diversified portfolio smooths the volatility of individual stocks. You deliberately develop somewhat fewer managers than total projected demand, accepting a planned shortfall, and fill the residual gap through targeted outside hiring — carrying less surplus and wasting less development investment. The remaining imperfect fits between pooled candidates and specific jobs get closed with just-in-time training rather than years of advance specialization.

• If a role is hard to fill and the cost of a shortfall is high while the workforce is likely to leave for rivals, then blend make and buy rather than relying on either alone.
• If external hiring for a role is both prohibitively expensive and lags the need for talent, then shift the balance toward internal development for that role.
• If the cost of over-developing (surplus) exceeds the cost of hiring externally under shortfall, then deliberately undershoot your internal development target and plan to buy the difference.
• If you run multiple separate development programs facing different business demands, then consolidate them into one pool so surpluses and shortfalls offset.
• If a forecast must be extended over many years, then break the program into shorter, separately forecast stages instead — the longer the forecast, the worse it is.

• Doing nothing — making no attempt to anticipate needs and defaulting to reactive outside hiring; the tell is that every quit becomes a calamity and every new skill demand becomes a crisis (roughly two-thirds of U.S. employers plan nothing).
• Relying on 1950s-style bureaucratic succession planning and long-term forecasting; the tell is a plan that must be rewritten every year because jobs change and people leave, so you keep asking what problem it actually solves.
• Developing employees for narrow, specialized jobs based on a best guess about future vacancies; the tell is stranded, wasted investment when the anticipated match never happens.
• Aiming to hit development targets exactly instead of deliberately undershooting; the tell is chronic surplus in some units and painful layoffs, because you treated an unavoidable forecast miss as a target to be met.
• Keeping development programs decentralized by division; the tell is duplicated infrastructure plus simultaneous surplus in one division and expensive emergency hiring in another that could have offset each other.

• ▸Role / job title: name the hard-to-fill position under analysis.
• ▸Expected demand (units, next horizon): your best current estimate of how many you'll need.
• ▸Shortfall cost: estimate the cost/delay/unreliability of filling the gap by outside hiring.
• ▸Surplus cost: estimate the wasted development investment plus layoff/attrition cost if you over-develop.
• ▸Internal development target: set this deliberately below expected demand where surplus cost is high.
• ▸Planned external hires: the residual gap you intend to buy rather than make.
• ▸Forecast horizon: shorten and split long programs into discrete, separately forecast stages.
• ▸Pooling scope: list which divisions/programs to consolidate so surpluses and shortfalls offset.
• ▸Just-in-time gap closers: training/coaching or leased/outsourced labor to close residual fit gaps.

1. 1Start by naming the actual source of your risk: business demand you cannot forecast. Accept that you will not forecast it away, so shift the goal from prediction to responsiveness — build practices that let you reassess and redeploy rather than commit years ahead.
2. 2Replace job-specific succession plans with talent pools built around broad competencies. Instead of grooming named individuals for named slots, develop a bench of people with transferable skills that can fill several possible future roles, so a wrong forecast costs less.
3. 3Break long development programs into shorter, modular stages. After each stage, stop and re-forecast: recheck how many people you actually need and in which directions, rather than committing to a full multi-year pipeline set on today's assumptions.
4. 4Run frequent reassessment cycles using real-time data, not one large detailed plan. Eisenhardt's finding applies: detailed long-range plans took time to build, were almost always wrong, and slowed firms down — so favor lighter, more frequent recalibration.
5. 5Deliberately avoid over-detailing your workforce plans. The downsizing study showed the most-detailed plans produced the worst outcomes because they were wrong in bigger ways; keep plans coarse enough to bend.
6. 6Pull demand-side uncertainty out of the individual employee by asking about their intentions. Confidential career coaching and simple questions about willingness to move let people clarify their own plans, which shrinks the range of outcomes you must plan around.
7. 7Time your development investments to shorten the payback window. Since 'how quickly employers can get a return counts,' size and stage programs so competencies are used soon after they are built, before strategy shifts or turnover erodes the return.

Deloitte's Deloitte Career Connections (DCC) shows uncertainty reduction through employee intentions rather than forecasting. Deloitte funds confidential coaches — including Deloitte employees — who advise on all career issues, even when it might be time to leave. More employees have used DCC than are currently employed there, and over four thousand of thirty-five thousand received coaching. The logic is not retention for its own sake: helping individuals figure out where they are going makes it easier for the company to plan around them. Deloitte estimates each premature departure costs $150,000, that at least 550 employees stayed because of a DCC intervention, and that this alone saved about $83 million. Corning's Matt Brush illustrates the same demand-clarifying move with a simple question: 'We ask if they are willing to move. If not, we take them off the candidate list.'

• If you cannot forecast business demand accurately, then do not build talent plans that assume you can — plan for responsiveness and reassessment instead.
• If a role's future need is uncertain, then develop broad, transferable competencies in a pool rather than grooming a named successor for a fixed slot.
• If a development program runs long, then split it into modular stages and re-forecast need at each break instead of committing to the whole pipeline upfront.
• If detailing a plan further does not improve your ability to adapt, then stop — the downsizing evidence shows more detail can produce worse, more disruptive errors.
• If an employee tells you they will not move, then take them off the candidate list — knowing their intentions removes uncertainty you would otherwise carry.

• Relying on 1950s-style forecasting and succession planning. The tell: your plans assume years of predictable growth and break the moment strategy shifts — as internal development collapsed in the 1970s when forecasts missed the downturn while pipelines kept producing talent.
• Defaulting to do-nothing outside hiring. The tell: you fill most vacancies externally, then watch cultivated talent walk to rivals, driving up cost and creating retention leaks in an open labor market.
• Over-engineering the plan. The tell: your most sophisticated, most-detailed workforce plans generate the biggest surprises and the most annoyance among staff and management when they turn out wrong.
• Building narrow, job-specific pipelines. The tell: a single strategy change strands people trained for a role that no longer exists, and you cannot redeploy them.
• Investing in development with too long a payback. The tell: strategy changes or declining tenure mean competencies aren't used long enough to recover the cost, so you quietly back away from training altogether.

• ▸Demand driver — name the outside factor (customer, competitor, regulation) creating this talent risk
• ▸Forecast confidence — rate how predictable this need is (low/med/high); low means plan for responsiveness
• ▸Broad competencies — list transferable skills for the talent pool instead of one job title
• ▸Modular stages — break development into stages with a reassessment point after each
• ▸Reassessment cadence — set how often you recheck need using real-time data
• ▸Plan detail cap — state the coarsest plan that still lets you act; avoid over-detailing
• ▸Employee intentions — record willingness to move/stay to remove candidates and shrink uncertainty
• ▸Payback window — estimate how soon the competency is used; flag if return is too slow
• ▸Mismatch cost — note cost of surplus (layoffs) or shortage (unfilled vacancy) to size the bet

1. 1Treat every development dollar like a supply-chain investment: before funding a program, name the specific risk it offsets — mismatch cost (too few/too many skilled people) or the retention loss of walking talent — because the biggest cost you're managing is unpredictable demand, not the training itself.
2. 2Share development cost with the employee when the skill is portable. Use tuition aid for self-study rather than fully-funded classroom time so the employee co-invests time and effort, converting a pure company outlay into a shared one and screening for motivation.
3. 3Compress development time so people reach productive roles faster. Model the pipeline as a supply chain: identify bottlenecks that block advancement and shorten processing time, because every extra month in a development job is unrecovered investment carrying retention risk.
4. 4Build low-cost, low-risk 'try-out' opportunities — the 'popcorn stands' model — small assignments or projects where you observe who performs before committing to expensive high-potential development.
5. 5Let self-selection do the filtering. Offer development as an opportunity people opt into (posting, application, self-study) rather than an entitlement pushed onto forecast-selected candidates, so you invest in those who reveal both interest and capacity.
6. 6Shorten the payback horizon. Structure development so you recoup value quickly — through productive project work during training, training wages, or immediate deployment — rather than betting on a multiyear return that retention leaks can erase.
7. 7Measure the return, not just the cost: track productivity gained, time-to-productive-role, and how much of your investment leaves via turnover, so you can compare make-versus-buy on actual dollars rather than the reflex 'people are our most important asset.'

Consider the CEO in the medical device industry who dismissed his HR head's management development proposal with, 'Why should we develop people when our competitors are willing to do it for us?' His logic was that outside hiring let rivals absorb development cost — rational in the 1980s–90s when the laid-off talent pool was deep. But the passages show the flaw: by the late 1990s the longest U.S. expansion depleted that pool, so employers were losing their own experienced people 'out the back door' to jobs identical to the ones they were filling with costly, unreliable outside hires. To maximize ROI he should not have chosen do-nothing OR full internal funding, but a balanced approach: co-invest via tuition aid, run try-outs like Goodrich's Manufacturing Leadership Program to spot performers cheaply, and shorten time-to-role so the investment pays back before any leak. That turns development from a gift to competitors into a hedged, retention-aware bet.

• If the skill is portable and marketable, then share the cost (tuition aid, self-study, training wages) so the employee co-invests and retention risk is partly priced in.
• If the skill is proprietary and firm-specific, then the company funds more of it, because the employee can't easily take that value to a rival.
• If you can't forecast demand for a role with confidence, then use short try-out assignments and self-selection instead of long, pre-committed high-potential tracks.
• If a development program's payback horizon runs years out, then redesign it so value accrues during the program (project work, immediate productive deployment) rather than only at the end.
• If the same experienced people are leaving as you're hiring in, then stop expanding outside hiring and shift investment to faster, cost-shared internal development.

• Choosing do-nothing and relying on outside hiring — the tell: you're paying a market premium for outside hires while your own experienced employees leave for near-identical jobs elsewhere.
• Running 1950s-style long-horizon succession and high-potential programs unchanged — the tell: costs are high, forecasts miss, and the pipeline keeps producing talent for demand that didn't materialize.
• Funding fully portable skills 100% yourself — the tell: you 'watch employees walk out the door' having absorbed all the development cost, subsidizing competitors.
• Pushing development onto forecast-picked candidates instead of letting people self-select — the tell: low motivation, wasted investment on people who don't want or use the development.
• Ignoring cost entirely because 'people are our most important asset' — the tell: no one can state the dollar return, so development gets cut first in the next downturn.

• ▸Role/skill targeted: name the specific hard-to-fill role.
• ▸Skill portability (portable / proprietary): drives who pays.
• ▸Cost-share method: tuition aid, self-study, training wage, or full-fund — pick per portability.
• ▸Try-out design: the low-cost assignment ('popcorn stand') used to spot performers first.
• ▸Self-selection gate: how employees opt in (posting, application) rather than being assigned.
• ▸Time-to-productive-role (target vs. current): the bottleneck you'll shorten.
• ▸Payback timing: when value accrues — during program or only after.
• ▸Retention-leak estimate: expected % who leave before payback, priced into the decision.
• ▸Make-vs-buy verdict: develop internally, hire outside, or hybrid — with the dollar rationale.

1. 1Treat your internal talent pipeline as a supply chain: the goal is to reduce bottlenecks that block advancement, speed processing time, and improve forecasts of need to avoid mismatches. Map where employees currently stall before roles open.
2. 2Build a transparent internal job board that posts vacancies internally before or alongside outside search, so you stop the pattern of losing experienced people 'out the back door' to jobs elsewhere that look like the ones you're filling with outside hires.
3. 3Publish clear career-path and skill-requirement information for each posted role, so employees can self-assess and pursue moves without relying on hidden succession lists inherited from 1950s-style bureaucratic planning.
4. 4Remove bureaucratic barriers to internal mobility — chiefly the requirement for current-manager approval — because these are the bottlenecks that make internal moves slower and less predictable than outside hires.
5. 5Let employees, not just central planners, drive the matching: allow them to apply against posted requirements, since in an open labor market where employees have a market for their skills, giving them internal options is a retention tool.
6. 6Balance make-versus-buy explicitly: use the internal market to fill what you can develop cost-effectively, and reserve outside hiring for genuine gaps, rather than defaulting to reactive outside hiring that lags the need for talent.
7. 7Measure the return: track internal fill rate, time-to-fill, and retention of developed talent against the cost of outside hires, whose quality is 'especially hard to measure' and whose failure rate is high at the executive level.
8. 8Acknowledge uncertainty in demand rather than forecasting it away — post roles and let the market clear as needs emerge, instead of planning years ahead on assumptions that collapse in a downturn.

Consider the medical device CEO Cappelli describes, who dismissed a management development program with 'Why should we develop people when our competitors are willing to do it for us?' The internal-market answer to his objection is not to abandon development but to retain the developed people so the competitor's poaching doesn't drain you. By the mid-1990s virtually every major employer wanted to recruit talent from competitors and retain its own — 'an impossibility at the aggregate level' — so the firms that kept some internal practices became 'academy companies.' An internal talent market changes the CEO's math: when a developed employee sees a posted internal role matching their skills and can move to it without their manager blocking them, they take the internal job instead of walking out the door. That converts the development spend from a gift to competitors into a retained investment. The tell that it's working is that the experienced employees you were losing 'out the back door' now fill your open requisitions.

• If most of your vacancies are currently filled from outside while your own experienced people leave for similar jobs elsewhere, then open an internal job board before external search.
• If manager approval is required to apply for an internal role, then remove it — that approval is the bottleneck slowing your internal supply chain.
• If a role can be developed internally at a cost below the price and failure risk of an outside hire, then make (fill internally); if not, then buy.
• If you cannot forecast the demand accurately, then post roles as needs emerge and let employees self-match, rather than committing to long-horizon succession plans.
• If a skill or path requirement is undocumented, then publish it — employees can't pursue roles whose requirements are hidden.

• Reverting to 1950s bureaucratic succession planning to run the market — the tell is a costly central process that is inaccurate because it assumes a predictable business environment that no longer exists.
• Letting managers veto internal applicants to hoard talent — the tell is high internal-application friction and people leaving for outside jobs that resemble your open roles.
• Doing nothing and defaulting to outside hiring — the tell is staffing that consistently lags the need and disappoints on both cost and quality.
• Posting roles without publishing skill requirements or career paths — the tell is low internal applicant volume because employees can't see how to qualify.
• Cutting development because employees walk out the door — the tell is you become dependent on poaching from competitors while they poach from you, an impossibility to win at the aggregate level.

• ▸Open role & title: name the posted vacancy
• ▸Skill/competency requirements: list what the role demands, published to all
• ▸Career path in: state which internal roles feed this one
• ▸Manager-approval barrier removed? yes/no — if no, name who must sign off and why
• ▸Make-vs-buy call: can we fill internally below outside cost + failure risk?
• ▸Internal fill rate: % of this role type filled from inside vs outside
• ▸Time-to-fill: days from posting to internal placement
• ▸Retention check: are developed people in this path staying or leaving to similar outside jobs?
• ▸Bottleneck note: where do candidates currently stall before this role opens?

1. 1Reframe the core problem: stop treating talent management as a forecasting exercise and start treating it as risk management. Name the two risks you are managing — mismatch cost (too many or too few skilled people) and retention cost (losing development investments to rivals).
2. 2Test your current model against the environment. Ask whether your product markets, technology, and competitors are as predictable as the post-WWII era. If strategies, structures, and operations change quickly in response to customer demands and competitor innovations, your long-term succession planning and multi-year pipeline assumptions are invalid.
3. 3Quantify the cost of the do-nothing default (relying on outside hiring). Pull recruiting-spend data: Deloitte's figure that the average U.S. employer spends ~50x more to recruit a $100,000 middle manager than it spends training its average employee is a benchmark. Add the retention risk of poaching from open labor markets.
4. 4Reject the two failed poles: the bureaucratic 1950s planning models (accurate only in a predictable world, and expensive) and the pure free-agency buy-everything model (unreliable, drives up recruiting cost, worsens retention). Position your choice as the third way — responsive and lower-risk.
5. 5Shorten your planning horizon to match uncertainty. Do not forecast years or decades ahead. Plan in shorter increments and build in the ability to adapt, because mistakes — waiting too long to change or planning for circumstances that never materialize — are the costly errors.
6. 6Balance make-versus-buy deliberately for each hard-to-fill role rather than defaulting entirely to one. Use internal development where you can absorb the lead time, and outside hiring where demand is genuinely unforecastable.
7. 7Share development risk and cost with employees (tuition assistance, cost-sharing, pre-hiring training) so that when talent walks out the door — an inevitable leak in an open labor market — your sunk investment is smaller.

Internal development collapsed in the 1970s as a direct case of environmental uncertainty defeating forecasting. Companies built talent pipelines on forecast assumptions of booming corporate growth, but business forecasting failed to predict the economic downturn. The pipelines kept producing talent under the old assumptions while demand fell, and no-layoff policies for white-collar workers meant that excess talent could not be released — producing a bloating of corporate organizations. When the steep 1981 recession hit, firms responded with white-collar layoffs, the end of lifetime employment, and reengineering that cut the very development staff and programs that had created talent. From that point employers turned to outside hiring, which worked only while it drew down the pool of laid-off talent; once the economy grew, outside hiring pulled from competitors, creating the retention problem. The whole sequence shows that a single unpredictable demand shock made an entire forecasting-based model uneconomic.

• If your business needs cannot be forecast accurately, then do not attempt to forecast your talent needs accurately either — plan for a range and build adaptive capacity instead.
• If a role's demand is highly unpredictable and the lead time to develop is long, then lean toward buying (outside hiring); if demand is more stable and internal cost advantages hold, then make (develop internally).
• If your development investment will likely walk out the door in an open labor market, then share the cost with the employee or shorten the payback period rather than fund it fully yourself.
• If you are relying on long-term succession planning built in the 1950s mold, then treat it as a legacy system that is inaccurate and costly, and replace it with shorter-horizon, risk-based planning.
• If outside hiring is your only staffing strategy, then price in the ~50x-training recruiting cost and the retention churn before defaulting to it.

• Forecasting away the uncertainty — assuming you can plan years or decades ahead. Tell: your talent plan rests on a single growth scenario with no contingency for a downturn or a competitor's move.
• Overbuilding the pipeline, then bloating. Tell: talent supply keeps flowing to forecast assumptions after demand has fallen, and you carry excess staff you can only shed through layoffs.
• Defaulting entirely to outside hiring to dodge planning. Tell: recruiting spend balloons, results are unreliable, and you find yourself poaching from — and being poached by — direct competitors.
• Reinventing 1950s practices (forced ranking, 360 feedback, assessment centers, succession planning) as if new, without noticing they assume a predictable environment. Tell: you adopt the technique but skip the question of whether your environment still supports its planning assumptions.
• Funding full internal development while ignoring retention risk. Tell: you invest heavily in employees, then watch them leave for rivals and cut development in reaction — losing both the talent and the capability to make more.

• ▸Role / skill assessed: name the hard-to-fill position.
• ▸Demand predictability (high/medium/low): how confidently can you forecast the need over your planning horizon?
• ▸Uncertainty drivers: which of customer demand, competitor innovation, technology change, regulation apply?
• ▸Development lead time: how long to grow this skill internally?
• ▸Mismatch cost estimate: cost of having too many or too few of this skill.
• ▸Retention risk: likelihood this talent leaves in an open labor market, and what that voids in sunk investment.
• ▸Outside-hiring cost: recruiting spend to buy the skill (benchmark against ~50x training spend).
• ▸Make/buy/mix decision: chosen sourcing given predictability and lead time.
• ▸Cost-share / hedge: how employees share development cost, or how you shorten the payback horizon.