Phillips ROI Level 5: How to Calculate Training Benefits & ROI in Indonesia (10 Steps, Isolation Methods, Monetization)

Short answer: The Phillips ROI Methodology is a training-impact measurement framework from Jack J. Phillips (formally since 1973–1983; ROI Institute co-founded with Patricia Phillips in 1992) that adds Level 5 — Return on Investment on top of the four Kirkpatrick levels. The process is 10 steps: set objectives across five levels, plan evaluation, collect L1–L2 data, collect L3–L4 data, isolate the training effect (control group / trend line / participant estimation), convert benefits to monetary, fully-load costs (including participant opportunity cost), compute BCR = Benefits ÷ Costs and ROI% = ((Benefits − Costs) ÷ Costs) × 100, report intangibles separately, then communicate honestly — including assumption limitations. For Indonesia: tie Level 4 to RKAP (BUMN) or performance contracts (agencies), use SBM PMK as cost-component ceilings, and pair with the cost of inaction.

Most "training ROI" articles hand over a formula without the discipline behind it — a pretty number that collapses the moment a CFO questions its assumptions. This guide closes that gap: the history and five levels of Phillips, the 10 operational steps, the five isolation methods with when to use each, the five monetization routes, the full list of fully-loaded cost components, a complete worked example from objective to ROI%, the treatment of intangibles, when ROI is overkill, and how to defend it to the CFO/RUPS/BPK audit.

Intended readers: HR / HC / L&D / SDM, Finance/budgeting, Procurement, and unit leaders who must justify training investment — in private companies, BUMN/BUMD, government agencies, institutions, associations, and non-profits.

Quick navigation

1. Phillips ROI: origin & position vs Kirkpatrick
2. Five Phillips levels & chain of impact
3. 10-step ROI Process Model
4. Set objectives at five levels (step 1)
5. Collecting L1–L4 data (steps 2–4)
6. Five methods to isolate training effect (step 5)
7. Convert benefits to monetary (step 6)
8. Fully-loaded cost (step 7)
9. BCR & ROI%: formula + interpretation (step 8)
10. Intangibles: why not forced to monetary (step 9)
11. Report honestly to stakeholders (step 10)
12. Full worked example (illustrative)
13. When ROI fits & when it is overkill
14. Indonesia adaptation: RKAP, DIPA, SBM, cost of inaction
15. Nine common mistakes & how to avoid them
16. Phillips ROI checklist
17. FAQ
18. Next step

Phillips ROI: origin & position vs Kirkpatrick

Jack J. Phillips began experimenting with training-ROI measurement in 1973 as a training manager at a US industrial firm. His first formal book on the method, Handbook of Training Evaluation and Measurement Methods, was published in 1983. With his wife Patricia Pulliam Phillips, PhD, he co-founded the ROI Institute in 1992 to standardize the methodology.

Phillips does not replace Kirkpatrick — it extends it. Kirkpatrick names what to measure (Reaction, Learning, Behavior, Results). Phillips adds how to monetize it (Level 5 ROI) plus process discipline (10 steps, isolation, fully-loaded cost). The academic critique that Kirkpatrick "does not explain cost" is answered by Level 5; the critique that "inter-level causality is not guaranteed" is answered by isolation discipline.

Rule of thumb: Use Kirkpatrick to measure impact. Use Phillips to prove that impact exceeds cost. They synergize; do not pick one.

Five Phillips levels & chain of impact

Level	What is measured	Key question
L1 Reaction & Planned Action	Participant reaction + intent to apply	Are participants satisfied AND intending to apply?
L2 Learning	Knowledge, skill, attitude, confidence, commitment	What do participants know/can do now?
L3 Application & Implementation	On-the-job application + required drivers	Are participants actually applying it?
L4 Business Impact	Effect on business indicators	Are revenue / cost / quality / time / satisfaction moving?
L5 Return on Investment	Monetary benefits vs program cost	Are financial benefits > costs?
(+) Intangibles	Benefits not monetized	What other valuable benefits are reported without conversion?

Chain of impact

The five levels form a causal chain: each level requires the prior to be valid for itself to be valid. A strong ROI comes from strong L4; strong L4 from strong L3; L3 from required drivers + strong L2; L2 from strong L1 design; L1 from a strong TNA. A broken link = brittle ROI. Experienced CFOs pull on one link to test the whole chain.

10-step ROI Process Model

This is Phillips' operational discipline. Each step has technique and artifact.

#	Step	Output
1	Set program objectives at L1–L5	List of measurable objectives per level
2	Design evaluation & data-collection plan	Instrument × level × timing matrix
3	Collect L1 & L2 data during program	Documented reaction & learning scores
4	Collect L3 & L4 data after program	Behavior + business indicator data at 30/60/90/180 days
5	Isolate program effect	% of impact attributable to training
6	Convert L4 data to monetary	Total benefits in IDR
7	Tabulate fully-loaded program cost	Total cost in IDR
8	Compute BCR & ROI%	Ratio + percentage
9	Identify & report intangibles	Structured list with evidence
10	Report results to stakeholders	Final report with honest narrative

All 10 steps are mandatory for credible ROI. The most-shortcut (and most-weakening): step 5 (isolation) and step 7 (fully-loaded).

Set objectives at five levels (step 1)

Five-level objectives must be consistent and built via backward design (see the Kirkpatrick deep guide). Example:

Level	Objective
L5	ROI ≥ 30% in 12 months
L4	Sales conversion rate from 18% to 24%; revenue per sales up 15%
L3	Every salesperson runs ≥ 1 diagnostic conversation per client meeting; documents 3 hidden needs per proposal
L2	100% of participants pass the SPIN-selling framework assessment ≥ 80%; role-play rubric ≥ 4/5
L1	Session NPS ≥ 60; intent-to-apply ≥ 4/5

These L1–L5 objectives become the evaluation benchmark — and become a commitment known to sponsor, CFO, participant, and vendor from the start.

Collecting L1–L4 data (steps 2–4)

Level	Instrument	When
L1	Smile sheet (relevance, intent-to-apply, barriers), session NPS	End of session / within 24 hours
L2	Pre/post test, role-play rubric, simulation	Pre: before session. Post: immediately after
L3	180/360 survey, manager/peer observation, work-product sampling, system data (CRM/HRIS)	30/60/90 days + 6 months
L4	Business KPIs from existing systems (revenue, conversion, turnover, NPS, time-to-X)	6–12 months

Per-level details: see Kirkpatrick deep guide §Instruments per level.

Five methods to isolate training effect (step 5)

This is Phillips' most important step — and the one most often shortcut. Without isolation, you claim 100% of a KPI gain comes from training, while market, product, or new incentives also contributed. Isolation makes ROI defensible.

(1) Control group arrangement

How: a trained group vs a similar untrained group. Compare performance in the same period. Difference = training effect.

Strengths: statistically strongest; bias controlled; easy to present to CFO.

Limits: not always possible (ethics — why "punish" the control group with no training?); needs sufficient population; difficult for org-wide mandatory training.

Fits: staggered training rollout (batch by batch), large populations (e.g. 200+ distributed salespeople), random assignment to early-batch vs late-batch as natural control group.

(2) Trend-line analysis

How: project pre-training performance trend into the future (as the prediction without training). Compare with actual post-training performance. Difference = training effect.

Strengths: no control group needed; fits when historical data is clean and stable.

Limits: the assumption that the trend continues without training is not always valid (market change, new products); needs ≥ 6 months of historical data.

Fits: KPIs with long, clear trend (e.g. error rate, NPS, retention).

(3) Forecasting methods

How: predictive math model (multivariate regression) predicts performance based on input variables (price, marketing, etc.). Compare prediction vs actual post-training.

Strengths: most precise when the model is mature.

Limits: high analytical capacity required; overfitting risk; hard to explain to a non-technical CFO.

Fits: organizations with strong analytical teams (banking, telco, e-commerce).

(4) Participant estimation

How: participants estimate the % of performance improvement that came from training (vs other factors), plus their confidence level. Applied factor = % × confidence.

Example: a participant estimates training contributed 40% of the conversion gain, with 70% confidence. Training contribution = 40% × 70% = 28%. If total conversion gain = IDR 500 million, training contribution = IDR 140 million.

Strengths: most practical; usable in any organization.

Limits: self-bias (can over/underestimate); needs confidence-factor discipline to be credible.

Fits: Phillips' default fallback when control group / trend line are not available.

(5) Supervisor / management estimation

How: same as participant estimation, but managers estimate. Often more conservative than participants.

Strengths: different perspective (further from training); calibration with participant estimation produces a range.

Limits: managers often have less detail on day-to-day implementation.

Fits: paired with participant estimation for triangulation.

Summary table

Method	Statistical power	Ease	When to use
Control group	Highest	Hard to set up	Large population, ethics OK
Trend line	High	Moderate	≥ 6 months stable historical data
Forecasting	Very high	Hard (needs analytics)	Data-driven organizations
Participant estimation	Moderate (needs confidence factor)	Easy	Default fallback
Supervisor estimation	Moderate	Easy	Triangulation with participant

Rule of thumb: Pick one primary method + one supporting (triangulation). State limitations in the report — experienced CFOs respect honesty more than a tidy formula.

Convert benefits to monetary (step 6)

After L4 (business indicators) is documented and isolated, convert to IDR. Phillips defines five conversion routes:

Route	How	Example
(1) Standard values	Values agreed by the organization	HR cost-per-hire IDR 25M; QA cost-per-defect IDR 5M
(2) Historical cost	Cost recorded from similar past cases	Customer churn cost = IDR 50M per account (CRM data)
(3) Internal expert input	Finance/Operations analyst computes	Analyst computes cost of poor onboarding = IDR 15M per FTE
(4) External data / benchmark	Public industry research	Benchmark turnover cost ≈ 50% of annual salary
(5) Participant/supervisor estimate	Internal estimate with confidence factor	Sales estimates value of 1 won deal due to diagnostic = IDR 200M

Commonly monetized indicator categories

Category	Indicator	Monetary value
Productivity	Output/hour, transactions/person	Revenue per unit
Quality	Defect rate, rework	Cost of defect
Time	Cycle time, time-to-X	Hourly cost saved × volume
Retention	Turnover rate	Replacement cost (recruit + train + ramp-up)
Customer satisfaction	NPS, CSAT, retention	LTV × retention gain %
Compliance	Fine avoidance, audit pass	Cost of non-compliance

Rule: use the strongest route available. Standard values > historical > expert > external > estimate. When conversion requires brittle assumptions, move it to intangibles — honesty protects the whole report.

Fully-loaded cost (step 7)

Phillips treats this discipline as the differentiator of credible ROI. Mandatory components:

#	Component	Note
1	Needs assessment / TNA	Internal + vendor (if any) cost
2	Module design & development	Often a large fixed cost for custom in-house
3	Participant materials	Books, job aids, simulations, etc.
4	Facilitator fee + travel/accommodation	Per vendor contract
5	Venue & equipment rental	In-house: internal allocation
6	F&B	During sessions
7	Participant opportunity cost (salary + benefits)	Often the largest, most-forgotten component
8	Participant travel/accommodation	When in-house at another location
9	Program management (L&D team)	Time allocation of L&D running it
10	Administrative overhead	Proportional allocation
11	Evaluation & measurement	Instruments, analysis, report
12	Tax	VAT, PPh 23 (see PO/VAT/tax-invoice guide)

How to compute participant opportunity cost (#7)

Formula: salary + benefit per hour × number of training hours × number of participants.

Example: 24 manager-level participants with total compensation IDR 30M/month = ~IDR 175,000/hour. 5 days × 7 hours = 35 hours. Total opportunity cost = 24 × 35 × IDR 175,000 = IDR 147 million.

For many programs, opportunity cost equals or exceeds facilitator fees. Ignoring it makes ROI look positive when it is actually negative.

Rule of thumb: Fully-loaded cost is the first integrity test of an ROI report. If one large component (especially opportunity cost) is missing, the CFO reading it knows you are selling a narrative while skipping impact measurement.

BCR & ROI%: formula + interpretation (step 8)

Formula

• BCR (Benefit-Cost Ratio) = Total Program Benefits ÷ Total Program Costs
• ROI% = ((Total Benefits − Total Costs) ÷ Total Costs) × 100

Interpretation

BCR	ROI%	Interpretation
< 1.0	Negative	Program loses money; review
1.0 – 1.5	0–50%	Marginal; context matters (urgency, large intangibles)
1.5 – 2.0	50–100%	Good; worth continuing
2.0 – 4.0	100–300%	Very good for behavior programs
> 4.0	> 300%	Re-test assumptions; may overestimate

Communication rules

• Report both (BCR & ROI%) — some stakeholders understand ratios, others percentages.
• Include a range of scenarios (conservative/moderate/optimistic) instead of a single number.
• Include confidence factor from the isolation method.
• Pair with cost of inaction: ROI 50% looks small until placed next to the cost of leaving the gap open.

Intangibles: why not forced to monetary (step 9)

Phillips explicitly advises against forced weak monetization. Common intangibles:

• Employee engagement — pulse survey scores
• Job satisfaction — annual engagement score
• Employer brand — external surveys, applicant ratio
• Corporate image — reputation index
• Cross-functional collaboration — collaboration metrics
• Team confidence — self-efficacy survey
• Qualitative customer satisfaction — documented testimonials

How to report intangibles:

1. Structured list with measurement evidence (scores, percentage change, data source).
2. State why not monetized (conversion assumption not strong enough).
3. Show trend where present (e.g. engagement up 7 points post-program).
4. Link to business context (high engagement → correlates with retention).

When forced to monetary with brittle assumptions: the CFO checks one assumption, finds it weak, and rejects the whole report. Intangibles stand as a separate report — equally valid, equally valuable.

Report honestly to stakeholders (step 10)

A healthy Phillips ROI final-report format:

1. Executive summary — 1 page: objective, L1–L5 results, BCR/ROI%, primary intangibles, recommendations.
2. Background & TNA — gap, baseline, participants, sponsor.
3. Program design — modules, methods, facilitators, schedule.
4. L1–L4 results — data, instruments, timing, triangulation.
5. Isolation method & justification — what was used, limits, confidence factor.
6. Benefit-to-monetary conversion — per-indicator route, assumptions, data source.
7. Fully-loaded cost — per-component detail.
8. BCR & ROI% — conservative/moderate/optimistic range.
9. Intangibles — list with evidence.
10. Lessons & recommendations — what worked, what failed, for the next cycle.

Honesty notes:

• Report negative ROI when that is the result — a more valuable signal than a fabricated number.
• State assumption limits openly.
• Avoid over-attribution to training (other factors must be acknowledged).

Rule of thumb: Experienced CFOs trust a report with range + confidence + limits more than one with a single tidy number. Reporting discipline = long-term credibility strategy.

Full worked example (illustrative)

Illustrative scenario.

Program

Diagnostic-sales training for 30 B2B salespeople, 4 days in-class + 6 months reinforcement. Sponsor: VP Sales.

L5 objective

ROI ≥ 50% in 12 months; BCR ≥ 1.5.

L1–L4 results

Level	Result
L1	Smile sheet 4.5/5; intent-to-apply 4.7/5
L2	Pre-test 62% → post-test 89%; role-play rubric 4.3/5
L3	90% of salespeople run diagnostic conversations in ≥ 80% of client meetings (CRM data); manager observation: 4.1/5
L4	Sales conversion rate up from 18% to 25% (over 6 months of measurement)

Effect isolation (step 5)

Method: participant estimation + supervisor estimation (triangulation).

• Sales estimate training contributed 65% of the gain, confidence 75% → contribution = 65% × 75% = 48.75%
• Supervisors estimate 55%, confidence 80% → contribution = 55% × 80% = 44%
• Average: (48.75% + 44%) ÷ 2 = 46.4%

Scenarios: conservative 40%; moderate 46%; optimistic 55%.

Convert benefits to monetary (step 6)

7 pp conversion gain × lead volume = ~120 additional deals per 6 months. Avg deal value IDR 80M (Finance data, standard value). Additional 6-month revenue = 120 × IDR 80M = IDR 9.6 billion.

Training contribution (moderate 46.4%): IDR 9.6B × 46.4% = IDR 4.46 billion.

Net margin on additional revenue (Finance: 30%): IDR 4.46B × 30% = IDR 1.34 billion = monetary benefit over 6 months.

Fully-loaded cost (step 7)

Component	IDR
TNA	25 million
Module design & customization	80 million
Participant materials	18 million
Facilitator fee + accommodation	240 million
Venue & equipment rental	40 million
F&B	30 million
Participant opportunity cost (30 × 28 hours × IDR 150,000/hour)	126 million
Participant travel	25 million
Program management (L&D)	50 million
Administrative overhead	20 million
Evaluation & measurement	60 million
11% VAT on vendor components	~45 million
Total fully-loaded cost	~759 million

BCR & ROI% (step 8)

Moderate scenario:

• BCR = IDR 1,340M ÷ IDR 759M = 1.77 → for every IDR 1 invested, IDR 1.77 in benefits
• ROI% = ((1,340 − 759) ÷ 759) × 100 = 76.5%

Conservative scenario (40% attribution, 25% margin): benefits IDR 960M × 40% × 25% / same assumptions, ROI ~ 25%.

Optimistic scenario (55% attribution, 35% margin): ROI ~ 143%.

Reported range: ROI 25–143%, moderate target 76%. The initial 50% target is exceeded.

Intangibles

• Sales team engagement score up 8 points (from 62 to 70).
• 4 salespeople received internal promotions (retention proxy).
• 12 positive client testimonials on diagnostic conversation quality.
• Not monetized (engagement→retention conversion assumption not strong enough for this number).

Communication (step 10)

The final report to VP Sales + CFO presents: target ROI 50%, moderate result 76.5% (range 25–143%), participant+supervisor estimation triangulated isolation, fully-loaded cost IDR 759M including IDR 126M opportunity cost, and intangibles separate. Recommendation: continue to batch 2 with required-drivers adjustment (data shows 90% adoption but 10% lag — extra coaching needed).

When ROI fits & when it is overkill

Fits

• Large programs (budget > 5% of annual L&D).
• Multi-year programs (academies, retainers).
• High-risk programs that must be defended to RUPS/CFO.
• Programs that become a strategic business case.
• Scaled BUMN/government procurement with audit-grade budget justification.

Overkill

• Small single classes (< 1 day).
• Routine annual refreshers.
• Regulation-mandated certification (must run regardless of ROI).
• Standard onboarding training.

The 5–10% rule

Phillips himself recommends applying ROI to 5–10% of the program portfolio — the largest/highest-risk. The rest are fine with Kirkpatrick L1–L3. Reason: ROI collection & analysis itself costs ~5–10% of program budget. Do not spend IDR 50M to prove an IDR 30M program returned IDR 60M.

Indonesia adaptation: RKAP, DIPA, SBM, cost of inaction

(1) Tie Level 4 to RKAP / performance-contract KPIs

BUMN: use KPIs already in RKAP (turnover, customer NPS, productivity per FTE). Government: use indicators already in Echelon I/II performance contracts. Do not invent new KPIs only for training.

(2) Fully-loaded cost per BAS + SBM

For DIPA-funded training, component costs are subject to the Standard Input Cost (SBM) set annually by PMK (FY 2025: PMK 39/2024; FY 2026: PMK 32/2025). SBM caps honoraria (speaker, instructor, module developer, organizer), per-diem, and F&B per province. Common expenditure accounts: Belanja Jasa Profesi (external speaker/instructor fees), Belanja Bahan, Belanja Perjalanan Dinas — per the Standard Chart of Accounts (BAS). For budgeting detail, see the training budget (RAB) guide.

(3) Cost of inaction (cost of doing nothing)

Pair ROI with cost of inaction: what is the loss if the gap is left open? Examples:

• High employee turnover from weak managers: replacement cost = ~50% annual salary × number of resignations.
• Customer churn from poor service: churn cost = customer LTV × churn count.
• Compliance gap: regulator fine potential, remediation cost.
• BUMN: BPK findings from non-compliant procurement, remediation cost & sanctions.

ROI that touches BPK/BPKP audit indicators is far more persuasive than generic ROI for BUMN/government.

(4) Monetary reporting status

Tax in fully-loaded: 11% effective VAT (HPP Law No. 7/2021) on vendor components; 2% PPh 23 on service fees (SE-35/PJ/2010). See the PO/VAT/tax-invoice guide.

Nine common mistakes & how to avoid them

#	Mistake	How to avoid
1	Ignoring participant opportunity cost	Compute salary + benefit per hour × hours × participants
2	No pre-training baseline	TNA mandatory + pre-test
3	Claiming 100% impact to training without isolation	Use ≥ 1 isolation method; state confidence factor
4	Forcing intangibles to monetary	Report as separate intangibles
5	Confidence factor ignored	Mandatory in participant/supervisor estimation
6	Measuring L4 too early	Wait 6–12 months
7	Standard values not Finance-agreed	Get Finance approval before conversion
8	Reporting ROI without valid L1–L4	Chain of impact must be complete
9	ROI reported a year after program	Agree the timeline in the proposal

Phillips ROI checklist

Before executing an ROI program:

• L1–L5 objectives set with sponsor before the program.
• L2–L4 baseline recorded (pre-test + initial business indicators).
• 10-step evaluation plan written.
• Isolation method selected + Finance-agreed.
• Standard values confirmed by Finance for monetary conversion.
• Confidence factor agreed for estimation methods.
• Fully-loaded cost components complete (including participant opportunity cost).
• Expenditure accounts & SBM confirmed (for BUMN/government).
• Vendor engaged in L3–L5 measurement (in contract).
• Final-report format agreed (10-section Phillips).
• Reporting timeline clear (6 months + 12 months).
• Intangibles will be reported separately with evidence.

FAQ

What is the Phillips ROI Methodology and how does it relate to Kirkpatrick?

The Phillips ROI Methodology is a training-impact measurement framework developed by Jack J. Phillips (starting 1973; formalized in Handbook of Training Evaluation and Measurement Methods, 1983; the ROI Institute was co-founded with Patricia Phillips in 1992). Phillips extends Kirkpatrick's four levels (Reaction, Learning, Behavior, Results) with Level 5: Return on Investment — monetizing Level 4 benefits and comparing them with the program's fully-loaded cost, producing BCR (Benefit-Cost Ratio) and ROI% ((Net Benefits ÷ Costs) × 100). The five levels are accompanied by an 'intangibles' classification (benefits intentionally not monetized). Phillips does not replace Kirkpatrick — it answers the critique that Kirkpatrick does not address cost.

What are the five Phillips ROI levels?

Level 1 — Reaction & Planned Action: participant reaction + intent to apply. Level 2 — Learning: knowledge/skill/attitude/confidence gain. Level 3 — Application & Implementation: on-the-job application + required drivers. Level 4 — Business Impact: effect on business indicators (revenue, cost, quality, time, satisfaction). Level 5 — Return on Investment: converting Level 4 to monetary value and comparing with program cost. Plus 'Intangible Benefits' that are reported but not monetized (job satisfaction, engagement, brand). The five levels form a chain of impact: every level requires the prior level to be valid for itself to be valid.

What is the Phillips ROI 10-Step Process Model?

Ten sequential steps. (1) Set program objectives at all five levels. (2) Design the evaluation & data-collection plan. (3) Collect Level 1 & 2 data during the program. (4) Collect Level 3 & 4 data after the program. (5) Isolate the effects of the program (five main methods). (6) Convert Level 4 data to monetary value. (7) Tabulate fully-loaded program costs. (8) Compute ROI = ((Net Benefits ÷ Costs) × 100). (9) Identify & report intangible benefits. (10) Report results to stakeholders with an honest narrative. The method's core is not the final formula but the discipline of each step — especially isolation (5) and fully-loaded cost (7), because these are the most-shortcut steps and the ones that most weaken ROI credibility.

What are the methods to isolate the effects of training?

Five main methods. (1) Control group — a trained group vs a similar untrained group; performance difference = training effect. Statistically strongest. (2) Trend-line analysis — project pre-training performance trend forward, compare with actual post-training performance; difference = training effect. (3) Forecasting methods — predictive math model of performance, compare predicted vs actual. (4) Participant estimation — participants estimate the % improvement attributable to training, calibrated with a confidence level (the 'discount' factor). (5) Supervisor/management estimation — managers estimate training's contribution. Plus: expert input, customer input. Pick the most feasible per context — control group is ideal but often impossible (ethics, small populations); trend line works when historical data is clean; participant estimation is the most practical with discount-factor discipline.

What is fully-loaded cost and why is it mandatory in training ROI?

Fully-loaded cost is the total training cost covering EVERY component, going well beyond facilitator fees and venue. Mandatory components: (a) needs assessment / TNA; (b) module design & development; (c) participant materials; (d) facilitator fees + travel/accommodation; (e) venue & equipment rental; (f) F&B; (g) participants — salary + benefits while in class (opportunity cost, often the largest); (h) participant travel/accommodation; (i) program management (L&D team running it); (j) administrative overhead; (k) evaluation & measurement; (l) tax (VAT, PPh). Omitting just component (g) flips many programs' ROI from positive to negative. Phillips treats this discipline as the differentiator of credible ROI.

How do you compute BCR and ROI%?

Two classic Phillips formulas: BCR (Benefit-Cost Ratio) = Total Program Benefits ÷ Total Program Costs. Example: benefits IDR 1.2 billion / costs IDR 800 million = BCR 1.5 — meaning for every IDR 1 invested, the organization gets IDR 1.50 back. ROI% = ((Total Benefits − Total Costs) ÷ Total Costs) × 100. For the same numbers: ((1,200 − 800) ÷ 800) × 100 = 50%. Rule of thumb: ROI < 25% weak; 25–100% good; > 100% very good for behavior programs; > 200% retest assumptions. BCR below 1 means financial loss; between 1 and 2 = marginal; above 2 = strong. Use both: BCR for ratio context, ROI% for percentage context in CFO reports.

How do you convert benefits (Level 4) to monetary value?

Five Phillips conversion routes. (1) Standard values — when the organization has agreed standard values (e.g. HR cost-per-hire, QA cost-per-defect). (2) Historical cost — use recorded cost from prior cases (e.g. customer churn cost). (3) Internal expert input — Finance/Operations analysts compute the financial impact. (4) External data / benchmark — industry research or public benchmark. (5) Participant/supervisor estimate — internal estimate with confidence factor. Commonly monetized indicator categories: productivity/output (revenue per unit), quality (cost of defect, rework), time (cost-per-hour saved), retention (replacement cost), customer satisfaction (LTV impact). Benefits hard to monetize credibly are still reported as intangibles — honesty protects the whole report's credibility.

What are intangible benefits and why not force monetization?

Intangible benefits are valuable program effects that are hard to monetize credibly: improved employee engagement, job satisfaction, employer brand, team confidence, cross-functional collaboration, corporate image. Phillips explicitly advises AGAINST forced weak monetization — data-collection cost is high and assumptions will be challenged. Instead, report them as a structured list with evidence (survey scores, 360 results, documented testimonials). This discipline protects ROI% credibility — if intangibles are forced in with weak assumptions, the CFO checks one assumption, finds it brittle, and rejects the whole report. Intangibles stand as a separate report, equally valid.

When is Phillips ROI a fit and when is it overkill?

Level-5 ROI fits: large programs (budget > 5% of annual L&D), multi-year programs (academies, retainers), high-risk programs that must be defended to RUPS/CFO, programs that become a strategic business case. Level-5 ROI is overkill for: small single classes, annual refreshers, regulation-mandated certifications (which must run regardless of ROI). Phillips himself recommends applying ROI selectively to 5–10% of programs (the largest/highest-risk); the broader portfolio is served by Kirkpatrick L1–L3. The rest are sufficient with Kirkpatrick L1–L3. Reason: ROI collection & analysis itself costs ~5–10% of program budget. Do not spend IDR 50 million to prove an IDR 30 million program returned IDR 60 million.

How do you defend training ROI to a CFO or the board?

Seven practices. (1) Agree Level 4 objectives with the sponsor BEFORE the program — not searched for afterward. (2) Show a pre-training baseline with written data (TNA). (3) Use Finance-agreed standard values (not vendor assumptions). (4) Declare the isolation method you used, including its limitations. (5) Include a confidence factor on estimates (e.g. 'participants estimated 30% improvement from training, 75% confidence → training contribution = 30% × 75% = 22.5%'). (6) Report intangibles separately, never forced to monetary. (7) Place results next to the cost of inaction (cost of doing nothing). CFOs respect a report that honestly names its limitations more than a too-tidy report — the latter triggers skepticism.

What are the most common mistakes in calculating training ROI?

Nine mistakes. (1) Ignoring participant opportunity cost — often the largest cost; (2) No pre-training baseline; (3) Claiming 100% impact to training without isolation (other factors ignored); (4) Forcing intangibles to monetary with brittle assumptions; (5) Confidence factor ignored in participant estimation; (6) Measuring Level 4 too early (before behavior settles); (7) Standard values not agreed with Finance; (8) Reporting only ROI without valid L1–L4; (9) ROI reported a year after program — decision momentum lost. Root cause: ROI chased as 'sales pitch' instead of 'test'. A credible ROI is sometimes negative — and that is a valuable signal that improves the next program.

How is Phillips ROI applied in BUMN and Indonesian government agencies?

Four adjustments. (1) Tie Level 4 to RKAP KPIs (BUMN) or Echelon performance-contract KPIs (government) — do not invent new indicators only for training. (2) Fully-loaded cost includes all components per the Standard Chart of Accounts (Belanja Jasa Profesi, Belanja Bahan, Belanja Perjalanan Dinas) + the Standard Input Cost (SBM, PMK 39/2024 for FY 2025, PMK 32/2025 for FY 2026) as the per-component honorarium ceiling. (3) For LKPP procurement, ROI becomes part of budget justification in the KAK and can serve as a tool at RUPS (BUMN) or budget-performance evaluation (government). (4) Pair with relevant cost of inaction: ASN/BUMN employee turnover, procurement errors, failed tax/regulatory compliance. ROI that touches BPK/BPKP audit indicators is far more persuasive than generic ROI.

Next step

You now have a complete Phillips ROI framework: the five levels, 10 steps, five isolation methods, five monetization routes, fully-loaded cost, BCR & ROI% formulas, intangibles, a complete worked example, and BUMN/government adaptation. The sensible next step is to run a TNA + L4 objective with the sponsor before executing any program whose ROI will be measured.

Neksus designs scaled programs with the Kirkpatrick + Phillips framework from day one: L1–L5 objectives agreed with the sponsor in the proposal, pre-training baseline run via TNA, isolation method selected (control group / trend line / participant estimation triangulated), monetization using Finance-agreed standard values, and a final report following the 10-section Phillips format with assumption honesty. Discuss your team's ROI program needs via the Neksus contact page — no obligation.

Also see the companion guides:

• Kirkpatrick 4-Level Deep — measurement foundation for L1–L4
• Training Needs Analysis (TNA) — baseline & L4 objective
• Building a Training Budget (RAB) — fully-loaded cost components
• How to Choose a Corporate Training Vendor — pick a vendor capable of L4–L5
• Corporate Training RFP: Template & Criteria — ROI questions to include
• Vendor Scoring Rubric — "methodology & measurement" weight
• Training PO, VAT & Tax Invoice Procedure — tax in fully-loaded cost
• See the full training catalog →

---

Last updated: 18 May 2026. The frameworks cited (Jack J. Phillips, Handbook of Training Evaluation and Measurement Methods, 1983; Jack & Patricia Phillips, ROI Institute since 1992; ROI Methodology 10-Step Process Model; Phillips Five Levels of Evaluation; Donald Kirkpatrick four levels; Jim & Wendy Kirkpatrick New World Kirkpatrick Model 2016; SBM PMK 39/2024 for FY 2025 and PMK 32/2025 for FY 2026; HPP Law No. 7/2021; SE-35/PJ/2010; Standard Chart of Accounts) are attributed to their original sources. Worked calculations are illustrative for method demonstration; the numbers exist to show the math. Neksus does not publish client names or success statistics; a credible ROI is sometimes negative — and that is a valuable signal worth reporting openly.