The Value Flip Calculator

First the cause, then the price. The value of L&D, shown before the training starts.

The paper in one page

L&D tries to prove its value after the training, at the one moment it can no longer change anything. The Value Flip Calculator moves that value to the decision.

• The problem. Proving impact afterwards rarely works. Studies seldom isolate the effect of training, the effect thins out on the way to the organization, and results measured in individuals do not simply add up to results of the organization (Garavan et al., 2019; Kim et al., 2025; Kozlowski et al., 2000).
• The order. First the diagnosis, then the calculator. The Performance Consultant AI applies Human Performance Technology and establishes whether knowledge is the cause. A price without a cause decides nothing.
• The value. Spend that does not go ahead, hours that stay in the work, a faster route to the result, protection for the training that does go ahead, and causes that keep coming back made visible.
• Why it counts. These numbers sit at the level where finance keeps them: the budget, the team dashboard, the roster. Nothing has to be isolated from everything else that moves a business.
• The limits. Avoided cost is not cash in the bank. Hours are not money. The calculator does not replace the evaluation of a training that goes ahead.

Why proving impact afterwards so often fails

Attendance, satisfaction, test scores, sometimes a return on investment: L&D collects every one of these after the budget is spent. The effort to link them to business results fails for three reasons.

2 · The effect thins out on the way

A meta-analysis of 157 studies of safety training shows the pattern in one set of numbers: large effects on what people learn, moderate effects on what they do at work, small effects on safety in the organization (Bisbey et al., 2026). None of this says training does not work. It says that its value at the level of the organization is smaller, less certain and hard to show afterwards.

3 · The level error, and what follows from it

The third reason sits deeper than measurement. It is a reasoning error about levels.

Most ways of showing impact follow one chain inside the individual, from reaction and learning to behavior and results, and then add the outcome up to the organization. Organizational results often arise from how people work together and from the conditions of the work, not from the sum of individuals (Kozlowski & Klein, 2000). An effect in individuals then says little about the organization, and its absence says little either (Kozlowski et al., 2000).

The error is easy to overstate. Where people work side by side on independent tasks, adding up comes closer to the truth, and measuring individuals remains worth doing; the error lies in the inference, not in the measurement. No. 11 in this series sets out the level error in a figure, with the evaluation models it affects (Arets, 2026c).

Diagnosis first, then the calculator

The calculator prices a decision. Only a diagnosis shows which decision is right. That fixes the order.

The diagnosis uses Human Performance Technology (HPT). It analyzes the business need, the performance that falls short and the cause, and its main gain is that someone has investigated whether a lack of knowledge or skill is really the cause. HPT looks for causes above the individual. Gilbert (1978) placed three of his six conditions for performance in the environment: information, resources and incentives. Tulser’s version of the method is described in earlier work and in this series (Arets & Heijnen, 2008; Arets, 2026b).

The Performance Consultant AI applies HPT to every request, which most L&D teams cannot do by hand for the volume they receive. It names the cause, where it sits (the organization, the process or the work) and how certain it is, and it recommends training only when a gap in knowledge or skill is confirmed (Arets, 2023, 2026d). How an organization then designs the solution differs per organization. The calculator prices the decision; it does not choose the design.

The Value Flip Calculator, as on the website

One screen with two sides: what saying yes costs, and what the decision keeps in the business.

The left side adds up three lines: fees, venue and catering, and the working time of the participants. The right side shows what does not go ahead when the cause lies in the work: the cash, before the cost of the fix, and the hours that stay in the work. Two checks come first. Without an approved or planned budget, or when the law requires the training, nothing counts. Starting values come from national labor statistics (Eurostat, 2026; U.S. Bureau of Labor Statistics, 2026) and give way to the prices finance uses; each report carries its method version, VFC 1.0. The rest is on the website.

The value L&D adds before the training starts

L&D creates value every week that it hardly ever reports: the value of a better decision. The calculator makes five kinds of it visible.

Value without revenue

In care, education and public services there is often no revenue to point to, and sometimes no agreed number at all. There the hours weigh most: a nurse who stays on the ward, a teacher who stays in class, temporary staff who are not hired. Incidents, waiting times and absence serve as the number to watch. Social return on investment offers the logic to correct such outcomes for what would have happened anyway (Nicholls et al., 2012).

At least as valuable as proving impact afterwards

The two answer different questions. Side by side, the difference is plain.

The argument is a comparison, not a finding. Impact afterwards remains worth measuring where a stakeholder asks for it and the budget allows a sound design; even its sharpest critic says so (Thalheimer, 2025). But the value of a decision has three advantages a CFO recognizes. It is known when it matters. It needs no isolation from the rest of the business. And it is checked by finance rather than claimed by L&D.

The CFO-proof report: three numbers a CFO counts

L&D rarely goes to the CFO alone. The report is one page the manager, the budget holder and finance can all check.

The report carries the first three values of chapter 04 as three numbers: Decided at the decision, checked in the budget; Delivered after three months, checked in the team’s dashboard and priced with finance’s own unit cost; and Kept from day one, checked in the roster.

What the report shows per outcome of the diagnosis

The report claims only what it can show: completions alone support “trained”, not “competent”. Who signs it, each organization decides for itself.

What the calculator gives each business model

Learning functions run one of four business models, from taking orders to creating value with the business, set out in the destination piece of this series (Arets, 2026a). The calculator, used after the diagnosis, gives each of them something different.

Each model is set out in the destination piece of this series. The Learning Specialist is a full business model in its own right: where the task of L&D is a reliable supply of capable people, it is a legitimate end point. For learning specialists, a diagnosis before the design is part of good practice. In leadership training, programs built on a needs analysis showed larger effects on learning and transfer, though not on results (Lacerenza et al., 2017).

Everyone else at the table

Medicine took this step first. In 2012, the Choosing Wisely campaign asked medical societies to name tests and treatments that add little for patients (Cassel & Guest, 2012), and researchers called the practice evidence-based de-implementation: stopping practices that are contradicted or unproven (Prasad & Ioannidis, 2014). A doctor who stops a treatment still treats the patient. L&D does the same: it does not say no to the manager. It says no to what will not work.

Why the cause decides

The calculator only creates value if the cause is right. The research below shows why the work around the learners matters, and how far that evidence reaches.

For learning designers. Strong design models, such as four-component instructional design, build learning around whole tasks from real work. Across 12 studies, all in schools and higher education, that improves learning (d = 0.79), with large differences between studies (Costa et al., 2022). Its authors add that transfer involves more than good instructional design, because the work environment can range from favorable to obstructing (Frerejean et al., 2023). A diagnosis before the design does not replace the designer. It protects the design.

What the evidence does not show

A paper that asks L&D to drop claims without evidence has to hold itself to the same standard.

Two popular numbers do not help. The claim that only 10% of training transfers to the job goes back to a rhetorical question in a 1982 article, not to measurement (Ford et al., 2011). The claim that 80% or more of performance problems lie in the environment is an estimate from experience and from surveys of what employees believe. The calculator uses neither. It does not need a percentage: nobody knows per request where the cause lies, and that is exactly why the diagnosis comes first.

How this paper uses the research

Effects are given with their size, their basis and the caveat their authors give. A number from one meta-analysis is an average across studies, not a promise for the request on your desk. That is the work of the diagnosis.

Requests from practice

Tulser holds [dozens of] real, anonymous cases. Five of them, each with the main cause in the work:

One case per outcome

One case for each outcome of the diagnosis, including one where training was the answer.

Placeholder · Tulser fills these four cases from its case library

Built to stay on L&D’s side

A number that can be misread will be. Without rules, the flip turns into a budget cut, a case against training or a question about past waste. Six safeguards are built into the calculator and the report.

• Real budgets only. No approved or planned budget, no avoided cost.
• Required training stays. Training required by law or regulation is never flipped. The question becomes how to make it work in the job.
• Forward only. The calculator works on the request in front of you, never on last year’s decisions. Past decisions were made without a diagnosis; that is a gap in the process, not a failure of L&D.
• The controller’s prices. Every starting value shows its source and gives way to the prices finance sets. Each report carries its method version.
• The money stays in performance. Agree with the CFO in advance that avoided cost stays twelve months in a budget for fixes and for training that works.
• Shared credit. The decision belongs to the manager; L&D prepared it. The report says so.

What the calculator does not claim

The calculator is credible only if its limits are on the same page as its numbers.

• It does not diagnose. Used before the diagnosis, it shows only what is at stake, and the diagnosis has to follow at once.
• Starting values are averages. Labor costs come from national statistics; in countries that publish earnings only, the real cost is higher. Fees and venue costs are Tulser’s starting values and vary widely in practice.
• Avoided cost is not cash in the bank. It becomes money for something else only when the budget is released or redirected.
• Hours are not money. Only where they replace overtime, temporary staff or a vacancy.
• The cost of the fix and of the diagnosis count. The report subtracts the fix; the time spent on the diagnosis is not in the calculator and belongs in the conversation.
• The second number needs an agreement. Without a baseline and a price agreed in advance, the delivered value is an opinion.
• It does not prove what a training earns. It shows the value of a better decision, not the effect of a training on results, which remains hard to isolate (Garavan et al., 2019). A training that goes ahead still needs its own evaluation.
• Not all value is in the three numbers. Safety, quality and risk often weigh more than cost, and the calculator does not price them.
• The evidence has limits too. Much of the research on the work environment shows what goes together, not always what causes what, and several estimates rest on few studies.
• The cases illustrate. They come from Tulser’s practice and do not establish a rate.

Run it on your next request

Five questions, in this order, before anyone says yes.

Started with question 4? Then question 3 is next, the same week. The Value Flip Calculator and the free diagnosis are at tulser.com/en/value-flip.

Two problems, one method, sixteen papers

Two models carry the method. Gilbert’s six conditions ask what kind of cause; Rummler and Brache’s three levels ask where it sits. Both are sound and this series uses both, and it leads with Gilbert, because his unit is the accomplishment, which is what an organization already counts, and because his six conditions are the categories an organization can actually change.

Where the method comes from

Two further pieces close the set and carry no number, because each points at all the numbered papers. Before the guide sits Arets L&D Business Models, the destination piece for the people who receive the requests: four models a learning function can run, and the route from HR-driven to business-driven. Last sits Solutions that cannot work, the field guide, ordered by the sentence that lands on your desk: seventeen forms a performance problem arrives in, each with an address in one of the numbered papers. Three orders in all: the destination you choose, the question you are on, and the form that landed on your desk.

Sixteen papers, six parts

Number 01 is the map and number 02 is the foundation. From number 05 on, every number is depth on exactly one step of the method. The filled block is the paper you are holding; the two unnumbered pieces close the set.

Where every claim in this paper comes from

Sources, continued