In-House vs Public Training: A Complete Decision Guide — When to Choose Which

Short answer: Choose public training when only 1–3 people need training, the need is urgent, content is general, and cross-industry exposure has value. Choose in-house when the target is behavior change for one team, content must be customized, sensitive data is involved, or participant numbers cross the cost break-even. The decision starts from a Training Needs Analysis (TNA) and is scored on six axes — participants, customization, confidentiality, networking, schedule, depth of change — then locked with a break-even calculation using real quoted figures.

Most "difference between in-house and public training" articles stop at a pros-and-cons list and one rough price example. That is enough to understand the terms, and not enough to make a decision accountable to finance and leadership. This guide closes that gap: six decision axes, the real break-even math, the hidden costs of each model, a decision tree, tax and procurement implications, the hybrid path, and when public is genuinely the right choice.

The intended readers: HR / HC / L&D / SDM, Procurement, and unit leadership teams deciding the training format for their teams — in private companies, BUMN/BUMD, government agencies, institutions, associations, and non-profits.

Quick navigation

1. Three models, one decision: precise definitions
2. The decision starts from the TNA
3. Six decision axes
4. Break-even math: when in-house is cheaper
5. The hidden costs of both models
6. Format decision tree
7. Full comparison table
8. Tax & procurement implications per model
9. Certification: public schedule vs in-house arrangement
10. The hybrid path & public → in-house migration
11. When public training genuinely wins
12. When in-house genuinely wins
13. Format decision checklist
14. Common mistakes & how to avoid them
15. FAQ
16. Next steps

Three models, one decision: precise definitions

This decision is often simplified into two options, when there are three models and all three are legitimate candidates:

• In-house (private). A program designed and run for a single organization. Participants from the same company, content customizable to internal processes and confidentiality, schedule and location set by the client, cost based on program/batch.
• Public (open class). An open class filled by participants from different companies. Standard content, a fixed schedule from the provider, per-participant cost, and the value of cross-industry networking.
• Blended/hybrid. Self-paced online modules (prework, theory, microlearning) combined with facilitator sessions (practice, simulation, coaching). Spreads the participant time load and protects long-term retention.

The correct decision picks the model most able to close the capability gap from the TNA — sometimes a combination of all three for different roles within one plan.

The decision starts from the TNA

The most common mistake: organizations choose the format by habit ("we usually send people to public") or by lowest price, before the need is clear. Training then runs, participants are satisfied, and nothing changes at work.

The Training Needs Analysis (TNA) framework from McGehee & Thayer dissects need at three levels, and these three levels guide the format:

• Organizational level — business objectives and the capability gaps blocking them.
• Task/role level — the specific competencies that must rise for a given role.
• Individual level — who the participants are, their starting level, and barriers to application.

The single filter that determines format: how deep is the change required, and for how many people in the same context? Knowledge change for a handful of individuals leans public; behavior change for one team in an internal context leans in-house.

Rule of thumb: Format follows the gap. The TNA sets the gap; the format is simply the most efficient way to close it for the given participants and context.

Six decision axes

Cost per participant is one of six axes. Score all six before looking at price.

Axis	Leans public	Leans in-house
Participant count	1–3 people from one role	One team/department, many participants
Customization need	General content suffices	Must use internal processes, policies, case studies
Confidentiality/context	No sensitive data	Internal strategy/data is discussed
Cross-industry networking value	Cross-company exposure is valuable	Internal context matters more than networking
Speed & schedule	Needed fast, no batch to form	Can schedule a batch, needs time/location control
Depth of behavior change	Goal is individual knowledge gain	Goal is changing how a team works (Kirkpatrick L3)

The axis that most often decides is the last one. Knowledge can be entrusted to a public class; behavior change for one team demands the cohort, context, and reinforcement that only in-house (or blended) can design.

Break-even math: when in-house is cheaper

This is the section similar articles almost never calculate, yet it is what makes the decision defensible to finance.

The two models' cost structures differ fundamentally:

• In-house has a large fixed cost that does not rise as participants increase: facilitator fees, module development & customization, program management. Plus a small variable cost per participant: materials, catering, certificates.
• Public is purely variable: a per-participant fee, multiplied by the number of people sent.

So there is a break-even point — the participant count at which in-house total equals public total:

Break-even (people) = In-house fixed cost ÷ (Public fee per participant − In-house variable cost per participant)

The mechanics:

• Below break-even → public is cheaper (you pay for only a few heads, without carrying the in-house fixed cost).
• Above break-even → in-house is cheaper, and its per-head cost keeps falling as participants increase because the fixed cost is divided among more people.

International industry sources generally place the break-even at a relatively small participant count — but the exact number is computed from real quotes because it depends on the facilitator fee, customization depth, and the public-class fee being compared. Do not stop at the ticket price: add opportunity and travel cost (next section) so the comparison is honest. There is no honest fixed price list for corporate training — figures derive from a program's cost drivers; how to build them is covered fully in the budget guide.

The hidden costs of both models

A ticket-price comparison misleads because the two models hide costs in different places.

Hidden cost	On public	On in-house
Participant time opportunity cost	Present (classroom time)	Present (classroom time), often more controlled
Travel & commute	High if the class is in another city	Low (facilitator comes to the client)
Risk of an under-filled class	None (pay per head)	Present — per-head cost spikes if quorum is missed
Loss of context	High — general content, no bridge to internal process	Low — content uses company context
Standard consistency across teams	Low — each person attends a different class	High — one standard for the whole cohort

Participant time opportunity cost is often the largest component and appears on no invoice. Recording it keeps the format decision made with the right numbers; ignoring it makes a public class look cheap when it is expensive once travel and lost context are counted.

Format decision tree

Skip the habit debate. Run this sequence:

1. TNA first. Set the gap, target participants, and behavior target. Without this, the format decision is a guess.
2. How many participants in the same context? ≤ 3 people across roles → leans public. One team/many people → continue.
3. Customization needed or sensitive data present? Yes → in-house. General content is safe → public still viable.
4. Target team behavior change (Kirkpatrick L3)? Yes → in-house or blended. Individual knowledge only → public.
5. Team spread across locations or participant time very expensive? Yes → blended.
6. Compute break-even with real quoted figures + opportunity & travel cost. Set the format from the computed result.
7. Install measurement. Whatever the format, agree on at least Kirkpatrick L1–L3 evaluation with a baseline.

Full comparison table

Dimension	In-house (private)	Public (open class)	Blended/hybrid
Best for	One team, high context & confidentiality	1–3 people, cross-industry exposure	Spread teams, retention & scale
Customization	High	Low (standard)	Medium–high
Confidentiality	High	Low	Medium–high
Cross-industry networking	Low	High	Low–medium
Cost structure	Fixed + small variable	Purely variable per participant	Content license + sessions
Per-head cost	Falls as participants rise	Fixed per participant	Depends on license + sessions
Speed to start	Needs design & scheduling	Fast (open class)	Medium
Team standard consistency	High	Low	Medium–high
Main risk	Quorum not filled	Context loss + travel	Technology complexity
Depth of behavior change	High (L3 possible)	Medium (tends L1–L2)	High (phased reinforcement)

Tax & procurement implications per model

The two models carry different administrative consequences — often missed until an invoice is rejected by finance.

• VAT. Both are training services; a VAT-registered (PKP) provider issues a tax invoice. Nominal rate 12% from 1 January 2025 (HPP Law No. 7/2021), effective 11% for non-luxury services via the DPP Nilai Lain mechanism.
• PPh 23 & transaction form. Public is usually billed per participant with an invoice per transaction. In-house is usually a single-program service contract — the buyer withholds Article 23 income tax (PPh 23) on the fee (2% if the vendor has an NPWP, 4% if not; technical service per SE-35/PJ/2010, e-Bupot withholding slip).
• Procurement. For BUMN/government, a standard public class is sometimes available via a lighter procurement route; a custom in-house program demands a TOR/RFP and the full supplier process (the Presidential Regulation 16/2018 jo. 12/2021 framework & LKPP rules).

Neksus context: Neksus designs both customized in-house classes and the blended path, with a RAB itemized on cost drivers and a clear tax position, and documents ready for procurement verification. The full mechanics are covered in the vendor and budget guides.

Certification: public schedule vs in-house arrangement

When the goal is an official competency certificate, the scheduling factor decides:

• Public often follows a fixed assessment calendar from a BNSP-licensed Professional Certification Body (LSP) referenced to the SKKNI — fast to access for individuals, tied to pre-set dates.
• In-house demands a separate assessment arrangement: scheduling LSP assessors for your cohort, sometimes preparing an assessment venue. More flexible for one large team, needs earlier coordination.

If you need an official competency certificate, ensure a licensed LSP stands behind the program in any format; a plain workshop yields only an attendance certificate. Neksus frames readiness toward a standard/certification honestly as "readiness toward", with no claim to accreditations it lacks.

The hybrid path & public → in-house migration

This decision is not always one-and-done. Two advanced patterns are often the wisest:

• Blended as the middle path. Self-paced online modules lower classroom hours (cutting opportunity cost) while facilitator sessions keep practice and coaching. Suited to spread teams or when participant time is very expensive; aligned with the 70-20-10 principle that places most learning outside the formal classroom.
• Public → in-house migration. Send 1–2 people to public training to validate topic relevance and approach quality at low risk. If the gap proves to span a whole team, follow with customized in-house measured to Kirkpatrick Level 3–4. Public acts as a low-cost trial; in-house becomes the scaled intervention once the evidence is strong.

When public training genuinely wins

Honesty (§4) demands naming when public is the best choice, treated as a first-class option:

• 1–3 people from one role need training.
• Urgent need with no time to form an in-house batch.
• General content that touches no sensitive data or strategy.
• Cross-industry exposure is itself valuable (benchmarking, networking, external perspective).
• Testing a topic before investing in a full in-house program.

When in-house genuinely wins

• Behavior change for one team/department (Kirkpatrick Level 3) is the target.
• Customization is mandatory — internal processes, policies, case studies, and industry language.
• Sensitive data/strategy cannot leave the room (aligned with PDP Law obligations when participant data is processed).
• Participant numbers cross the break-even so per-head cost falls.
• Standard consistency across teams is an explicit goal.

Format decision checklist

• TNA done; gap, participants, and behavior target are clear.
• Six axes scored before looking at price.
• Break-even computed with real quoted figures.
• Time opportunity cost + travel included in the comparison.
• Model risks mapped (in-house quorum / public context loss) + mitigation.
• Confidentiality need assessed; data clauses prepared if in-house touches sensitive data.
• Certification need checked; LSP/BNSP availability & schedule confirmed.
• Blended option considered if team is spread / participant time is expensive.
• At least Kirkpatrick L1–L3 measurement with a baseline agreed for any format.
• Tax position (VAT incl./excl., PPh 23) and procurement route clear per model.

Common mistakes & how to avoid them

Key points to remember:

• Choosing the format by habit → start from the TNA, score the six axes.
• Comparing only the ticket price → compute break-even + opportunity & travel cost.
• Sending 30 people to public → above break-even, in-house is cheaper and more consistent.
• In-house for 2 people on a general topic → below break-even, public is more efficient.
• Forgetting confidentiality → sensitive data/strategy demands in-house + data clauses.
• Stopping at a satisfaction score → agree on Kirkpatrick L3 so the format is judged by behavior change.

FAQ

What is the difference between in-house and public training?

In-house (private) is a program designed and run for a single organization: participants from the same company, content customizable to internal context and confidentiality, schedule and location set by the client, cost based on program/batch. Public (open class) is an open class filled by participants from different companies: standard content, a fixed schedule from the provider, per-participant cost, and the value of cross-industry networking. Blended combines self-paced online modules with facilitator sessions to spread the time load and protect retention.

At how many participants does in-house become cheaper than public?

There is no single honest number; the break-even is always computed per case. In-house has fixed costs (facilitator fees, module development) that do not rise as participants increase, while public is purely variable (a per-participant fee). Break-even = in-house fixed cost ÷ (public fee per participant − in-house variable cost per participant). Below that point public is cheaper; above it in-house wins and its per-head cost keeps falling. Compute it with real quoted figures, then add the opportunity cost of time and travel so the comparison is honest.

When is public training the right choice?

Public wins when only 1–3 people need training, the need is urgent with no time to form a batch, the content is general and touches no sensitive data, participants actually benefit from cross-industry exposure and networking, or the goal is individual knowledge gain. Public is also a fast, low-risk way to test a topic's relevance before investing in a full-scale in-house program.

When is in-house training the right choice?

In-house wins when the target is behavior change for one team/department (Kirkpatrick Level 3), content must be tailored to internal processes, policies, and case studies, sensitive data or strategy cannot leave the room, participant numbers are high enough that per-head cost falls, or consistency of standard across teams is the goal. The deeper the behavior change required, the stronger the case for in-house.

Is cost per participant the only consideration?

It is not the only one. Beyond cost, weigh five other axes: customization need, confidentiality/internal context, the value of cross-industry networking, speed & schedule flexibility, and the depth of behavior change targeted. A format decision driven only by lowest price often produces training that is cheap on paper and has no impact at work. Start from a Training Needs Analysis (TNA), then choose the format most able to close the gap.

How do the tax implications of in-house differ from public?

Both are training services and attract VAT if the provider is VAT-registered (nominally 12% from 1 January 2025, effective 11% for non-luxury services). Public is usually billed per participant with a tax invoice per transaction; in-house is usually a single-program service contract, so the buyer withholds Article 23 income tax (PPh 23) on the fee (2% if the vendor has an NPWP, 4% if not; technical service per SE-35/PJ/2010). For BUMN/government procurement, a standard public class is sometimes available via a lighter route, whereas in-house demands a custom TOR/RFP. The full mechanics are covered in the vendor and budget guides.

What is blended training and when does it beat both?

Blended combines self-paced online modules (theory, prework, microlearning) with facilitator sessions (practice, simulation, coaching). It beats both in-house and public when the team is spread across locations, when participant time is expensive so the load must be spread, or when long-term retention demands phased reinforcement aligned with the 70-20-10 principle. Blended trades some classroom density for reach and sustainability.

What is the biggest risk of each model and how do I mitigate it?

In-house risk: the class is not filled, so per-head cost spikes — mitigate with a minimum quorum in the contract and realistic batch scheduling. Public risk: participants return with general knowledge and no bridge to company context, plus travel and commute cost — mitigate with post-class application assignments and choosing a class near the location. For both, at least Kirkpatrick Level 3 evaluation keeps the format judged by behavior change.

Can I start with public and then move to in-house?

Yes, and it is often the wisest path. A healthy migration pattern: send 1–2 people to public training to validate the topic's relevance and the approach's quality at low risk, then if the gap proves to span a whole team, follow with customized in-house measured to Level 3–4. Public acts as a low-cost trial; in-house becomes the scaled intervention once the evidence of need is strong.

How do I decide objectively using a framework?

Use a decision tree and axis scoring. Start from a TNA to set the gap and behavior target, then score the six axes (participants, customization, confidentiality, networking, schedule, depth of change), compute the cost break-even with real quoted figures including opportunity cost, and only then decide. A framework-based decision shifts the choice from 'we usually go public' or 'cheapest' to the format most likely to close the gap and be measurable.

Next steps

You now have the full framework: start from the TNA, score the six axes, compute the break-even with opportunity cost, map the risks, and lock measurement. The sensible next step is to run a short TNA to set the gap and behavior target — the format decision falls into place once the need is clear.

Neksus works exactly along this path: every program starts from a training needs analysis, designed in the format most efficient at closing the gap (customized in-house, blended, or a combination), with a RAB itemized on cost drivers and a clear tax/procurement position. Discuss your team's needs and request an initial TNA via the Neksus contact page — no obligation, as the right starting point.

Explore the guides and topics relevant to your decision:

• How to Choose a Corporate Training Vendor / Provider — scoring rubric, RFP, legality & tax
• Building a Training Budget (RAB) & Annual Plan — cost components, tax, budget break-even
• Leadership for First-Line Managers
• Organizational Change Management
• See the full training catalog →

---

Last updated: 18 May 2026. The tax & procurement sections explain generally applicable mechanisms confirmed per contract following the latest regulations (incl. Law No. 7/2021, SE-35/PJ/2010); validate with your tax and legal teams. The break-even is a calculation framework whose result derives from real quotes. Neksus does not display client names or success statistics; every cost figure derives from a program's specific cost drivers, and there is no fixed price list.