All posts

Blog

Phillips ROI Methodology: Adding Level 5 to Your Evaluation

November 7, 2026 · By Zohaib Butt

Phillips ROI Methodology: Adding Level 5 to Your Evaluation

Your CFO just asked what the new supervisor program actually returned, someone offered the post-workshop survey scores, and the silence that followed is exactly the silence the Phillips ROI methodology was built to end. Kirkpatrick gave L&D four levels of measurement; Phillips added the fifth — return on investment, expressed as a percent, computed from data finance can audit. Getting to Level 5 is less about statistics and more about records you already keep.

The money at stake is not small. US organizations spent $102.8 billion on employee training in 2023 — an average of $874 per learner (Training magazine, 2023 Training Industry Report) — and 71% of organizations offer leadership development (Training magazine, 2023). Yet the commonly cited Ebbinghaus forgetting curve holds that people lose about 50% of new material within 24 hours without reinforcement — programs nobody measures are also programs nobody manages.

This guide walks the five levels, the ROI formula, the isolation and conversion steps that make the number credible, what a study costs, and a worked example to copy.

Key Highlights

  • The Phillips ROI Methodology keeps Kirkpatrick's four levels and adds a fifth — ROI in percent.
  • US training spend hit $102.8 billion in 2023, an average of $874 per learner (Training magazine, 2023).
  • 71% of organizations offer leadership development (Training magazine, 2023) — few can state what any of it returned.
  • Level 5 is arithmetic: converted benefits minus fully loaded costs, divided by costs — the example below reaches 150%.
  • Isolation and conversion are the two steps that separate a defensible ROI percent from a vanity number.
  • People lose about 50% of new material within 24 hours without reinforcement (Ebbinghaus, commonly cited) — measure at 90 days, not day one.
  • In our experience roughly 10–15% of programs justify a full Level-5 study (illustrative).

Phillips ROI Methodology: The Five Levels at a Glance

The Phillips ROI Methodology extends the Kirkpatrick model, and the levels build on each other. Level 1 captures reaction — did the session engage people — and Level 2 captures learning, whether knowledge or skill actually increased.

Level 3 captures application — is the behavior showing up weeks later. Level 4 captures business impact — did output, quality or retention move. Level 5 converts all of it into money and compares it to cost.

Most US teams stop at Levels 1 and 2 — training reviews become happy sheets and quiz scores. The ROI conversation begins at Level 3, the first level tied to a business measure. The table below is the version to print for your next planning meeting.

Level What it measures The question it answers Typical tool
Level 1 — Reaction How participants engaged with the session Did the program land? Post-session survey
Level 2 — Learning Knowledge or skill gained Did they learn it? Skill check or assessment
Level 3 — Application Behavior on the job Are they using it? Manager observation at 30–60 days
Level 4 — Impact Business measures Did outcomes move? Operational and HR records
Level 5 — ROI Benefits converted to money versus cost Was it worth it? The Level-5 ROI worksheet

Effort rises steeply with each level — the honest reason most teams stop early. For an illustrative 20-person program, expect about 3 hours of Level-1 survey work, 8 for Level-2 testing, 18 for Level-3 observation, 30 for Level-4 impact pulls and 45 for a full Level-5 study (illustrative planning figures).

Hbar chart of Phillips ROI Methodology evaluation hours per level for an illustrative 20-person program: Level 1 needs 3 hours, Level 2 needs 8, Level 3 needs 18, Level 4 needs 30 and Level 5 needs 45. Illustrative planning figures.

Key takeaway: the five levels stack — no credible Level-5 percent exists without Level-3 and Level-4 data underneath it.

What Level 5 Requires: The ROI Formula and the Data

The formula takes one line: ROI percent equals net benefits divided by fully loaded costs, times 100. The formula is the easy part; the hard inputs are dollars and full costs.

Fully loaded costs means facilitation fees, licenses, participant time, backfill, travel, venue and your team's hours. If finance finds a cost line you left out, the whole percentage loses its standing. Benefits means operational records — turnover, output, error rates, overtime — converted at rates finance recognizes.

Here is the worked example for the rest of this guide: a $18,000 fully loaded program for six supervisors at a 140-person specialty manufacturer. Converted benefits reach $45,000 in the first year: avoided turnover worth $20,000, productivity gains worth $15,000 and reduced errors worth $10,000. Net benefits are $27,000, so the Level-5 ROI computes to 150% (illustrative).

Worked example line Amount Notes
Program cost, fully loaded $18,000 Facilitation, time, backfill, materials
Avoided turnover $20,000 One avoided exit versus a $20,000 replacement cost
Productivity gains $15,000 Time saved per week, converted at a loaded rate
Reduced errors $10,000 Rework hours avoided, same loaded rate
Net benefit $27,000 $45,000 in benefits minus $18,000 in costs
Level-5 ROI 150% $27,000 divided by $18,000, times 100

Bar chart of converted benefits in the illustrative worked example: avoided turnover $20,000, productivity gains $15,000 and reduced errors $10,000, a $45,000 total against an $18,000 program cost. Illustrative midpoints.

Key takeaway: report the percent only when the cost side is fully loaded and every benefit line traces to a record finance can pull — half a model is worse than no model.

Isolating the Effects of Training

An ROI percent inherits every other influence on the business, which is why isolation earns the number its credibility. Output may have risen because of the program, a new hire, a price change or seasonality. Isolation techniques carve out training's share before it becomes a percentage; pick two and triangulate.

The five methods in everyday use are a comparison group, a trend line, confidence-adjusted participant estimates, manager estimates and a contributing-factors table. The table below is how teams choose.

Isolation method How it works Credibility Effort
Comparison group Compare a trained team with an untrained one on the same metric Highest Needs two teams
Trend-line analysis Extend the pre-program trend and attribute only the deviation High Needs clean history
Participant estimates Ask participants what share their behavior drove, adjusted down Moderate Fast, some bias
Manager estimates Front-line managers estimate the program's contribution Moderate Fast, some bias
Contributing factors List every influence and assign training a stated percentage Moderate Needs honesty

Trend-line work has a humbling habit: the pre-program trend already explains part of the improvement. In our work with a 140-person specialty manufacturer, extending the output trend forward cut the claimed benefit roughly in half — and the remaining half still cleared the investment (our observation). If Levels 1–4 are not instrumented yet, start with our Kirkpatrick model guide before attempting Level 5.

Key takeaway: choose your isolation method before results arrive and write it down — an isolation note drafted after the fact reads as advocacy, not measurement.

Converting Benefits to Dollars

Conversion turns Level-4 movement into money with records you already own. Quality converts at the cost of rework, scrap and callbacks; turnover converts at replacement cost — we plan about $20,000 for a frontline supervisor in US mid-market work (illustrative); productivity converts hours saved at a loaded hourly rate — salary times 1.3 divided by 2,080 hours (illustrative).

The three buckets came straight from those conversions: one avoided exit at $20,000, supervisor time saved worth $15,000 a year, and rework hours avoided worth $10,000. None required a survey; all required a record and a rate.

Benefits do not arrive at once — the payback picture matters as much as the final percent. In the illustrative model below, converted benefits accumulate to $6,000 by month 3, $16,000 by month 6, $30,000 by month 9 and $45,000 by month 12, crossing the flat $18,000 cost near month 7 (illustrative).

Line chart of cumulative converted benefits for the illustrative worked example reaching $6k at month 3, $16k at month 6, $30k at month 9 and $45k at month 12, crossing the flat $18k program cost near month 7. Illustrative model.

Timing also protects you from the forgetting curve: the commonly cited Ebbinghaus finding puts about 50% of new material at risk within 24 hours without reinforcement — measure at 90 days, not the day the workshop ends. The line-by-line arithmetic lives in our training ROI formula walkthrough.

Key takeaway: convert with rates finance already uses — a figure they would compute differently is a claim, not a benefit.

What a Level-5 Study Costs

Evaluation has a price, and pretending otherwise produces half-finished studies. Cost drivers are the number of programs studied, whether data is instrumented, how hard isolation is, and whether work is internal or bought. Internal runs cost time; external runs cost fees but bring method discipline.

Illustrative planning ranges: an internal Level-5 study runs about $1,200 of staff time for a 20-person program, $3,000 at 100 participants and $6,500 at 300, while an external consultant run costs roughly $3,500, $9,000 and $18,000 at the same sizes (illustrative midpoints, based on our practice).

Grouped bar chart of Level-5 evaluation cost by program size: internal teams spend about $1,200 for 20 participants, $3,000 for 100 and $6,500 for 300, while external consultants run $3,500, $9,000 and $18,000. Illustrative 2026 US ranges.

Against those fees, an $18,000 program with $45,000 in modeled benefits comfortably funds its own measurement. If the study would cost a third of the program's value, roll it up to portfolio level and sample within it.

Key takeaway: budget evaluation as a named line in the program budget — an unfunded study is a study that stops at the survey.

Phillips ROI Methodology in Practice: When to Add Level 5

Not every program earns a full Level-5 study, and saying so makes the ones that do get believed. Candidates are strategic, expensive, visible or recurring, with a sponsor who will actually read the result. In our experience about 10–15% of a portfolio fits (illustrative); the rest should stop at Level 4 with benefits described in units.

Evaluation spend follows the same logic: expect about 2% of program budget for Level 1–2 surveys, 4% for Level-3 checks, 6% for Level-4 pulls and 8% for a full Level-5 study (illustrative midpoints).

Bar chart of Phillips ROI Methodology evaluation spend as a share of program budget by level, illustrative midpoints: Levels 1–2 need about 2%, Level 3 about 4%, Level 4 about 6% and a full Level-5 study about 8%. Illustrative planning figures.

Two habits keep the practice honest: repeat the flagship study every two to three years, since annual percent swings are usually noise, and package the result the way finance reads — percent, payback month, isolation note and assumptions, as our guide on how to prove training value shows.

Key takeaway: run Level 5 on the few programs where the percent will change a real decision — everywhere else, Levels 3 and 4 tell you enough.

What Your Business Should Do

Days 1–30: Get the data plumbing ready.

  • List your five largest programs and mark which Kirkpatrick levels each measures today.
  • Pull two quarters of baseline data for the program you most need to defend.
  • Agree with finance on a fully loaded cost template, including participant time and backfill.

Days 31–60: Run one conversion exercise.

  • Pick the strategic program and convert two benefit streams to dollars using your own records.
  • Apply one isolation method — a comparison group or a trend line — and write down the assumptions.
  • Compute the ROI percent with a deliberately conservative estimate and finance's loaded rates.

Days 61–90: Report and decide.

  • Report the ROI percent with the isolation note and assumption list attached.
  • Present the payback month alongside the percent — let the crossing point argue.
  • Decide which programs earn a full Level-5 study and which stop at Level 4.

Key takeaway: one program measured properly beats ten programs surveyed — ninety days yields your first defensible ROI percent.

The Bottom Line

The Phillips ROI Methodology turns training from a cost story into a percent your CFO can audit: five levels, fully loaded costs, converted benefits and one isolation method honestly applied. In the worked example above, an $18,000 program produced $45,000 in converted benefits — a 150% ROI — and it survived because the isolation note came before the number, not after it. Start with one strategic program and let the records argue for you. If you want this mapped to your own team, book a free strategy call or start with the free resource library.

Work With Zohaib Butt

1. Free Resource Library — Zohaibbutt.com/store

World-class training and coaching resources, all in one place. Download the assessments, templates and playbooks you need today.

2. Online Courses — learning.zohaibbutt.com

Self-paced leadership and coaching programs for managers and owners who want structure, not scattered tips.

3. Free Strategy Session — funnelsthatsell.com

Book your complimentary strategy call and we will map your biggest people-development bottleneck to a 90-day fix.

Free assessment pack (PDF)

Score your readiness to prove training value at Level 5.

Download the PDF pack

FAQ

What is the Phillips ROI methodology in plain English?

It is a five-level evaluation model: Kirkpatrick's four levels — reaction, learning, application and business impact — plus a fifth that expresses the return as a percentage. That percent equals net program benefits divided by fully loaded program costs, times 100. Its two defining steps are isolating the effects of training and converting benefits to dollars, which makes the number auditable rather than rhetorical.

How is Phillips Level 5 different from Kirkpatrick Level 4?

Level 4 measures business impact in operational units — error rates down, output up, regrettable turnover down. Level 5 converts those movements to dollars, subtracts the fully loaded cost and reports the result as a percent. You cannot compute Level 5 without Level-4 data first — teams that skip Levels 3 and 4 produce percentages nobody believes.

What is a good ROI percentage for a training program?

Finance will compare your percent to the hurdle rate it applies to any investment, so there is no universal target. In our experience, a conservatively modeled ROI above 100% gets attention, but a credible 25% beats an inflated 400% (illustrative, based on our experience). If your model needs heroic assumptions to break even, present the payback month and the impact units instead.

How long does a Level 5 ROI study take?

Plan on 60–90 days after a program ends for a single-program study (illustrative, based on our experience). Benefits need time to appear in operational data, isolation takes real analysis, and finance will check your conversions. Collect the baseline before the program starts — retrofitting a baseline after the fact is the most common reason studies stall.