Your Training Partner
Techniques Toolbox
Decision diagram for make-or-buy analysis. A required deliverable enters the analysis, which weighs capacity, skills, expertise, permanent positions, discounted cost and risks. The Make branch leads to internal resource planning; the Buy branch leads to the procurement strategy, the call for tenders, vendor assessment and then the contract.

Make-or-Buy Analysis

Make-or-buy analysis determines whether the organisation produces a piece of work or a deliverable itself or buys it from an external source. It compares the two options on the same factors: available capacity, skills, the need for specialised or independent expertise, the permanent positions to be created, the lifetime cost and the risks of each. It comes before the choice of a supplier: until it concludes in favour of buying, there is no tender to run and no supplier to compare.

Goal

Make-or-buy analysis compares two options for a specific deliverable, producing it in-house or buying it, on the same factors and over the same cost horizon. The PMBOK Guide lists it among the tools of the Plan Sourcing Strategy process.

It settles who produces the deliverable. The choice of supplier comes afterwards, through vendor assessment. Allocating people already on staff belongs to resource planning, which the PMBOK Guide distinguishes from the sourcing strategy: the latter covers work acquired from outside the organisation.

The deliverable is a decision record for each item examined: the option chosen, the factors that prevailed, the costed comparison, the risks accepted and the person who decided. The PMBOK Guide files these insourcing or outsourcing decisions in the sourcing strategy plan, next to the source selection criteria that will apply if buying wins.

Usage

When to use

  • Deliverable achievable either way: a module, service or component that the market offers and that the team knows how to produce.
  • Before a call for tenders: establish that buying is justified.
  • Study of a HERMES project: compare an IT development option with an IT adaptation option.
  • Missing capacity or skills: team at full capacity, tight deadline, speciality absent from the organisation.
  • Independent expertise required: an audit, penetration test or second opinion that the team cannot conduct on its own work.
  • End of a contract or of maintenance: renew, change provider or bring the work back in-house.

When not to use

  • Option imposed: by law, a framework contract or a group decision; move on to vendor assessment or to internal planning.
  • Routine low-value purchase: the analysis costs more than the possible difference; apply the organisation's purchasing procedure.

Description

Scoping the decision

The analysis covers a specific item: a deliverable, a work package or a service, with the requirements it must meet. "The information system" is too broad, because an organisation buys part of it and develops another part. Break it down to the level where both options are real: the invoicing module, the hosting, the data migration, the penetration test.

Both options must deliver the same thing. An off-the-shelf product often covers part of the requirements. The "buy" option then includes the adaptation, configuration or additional development that closes the gap. Failing that, the organisation drops those requirements and says so in writing. Comparing a full development with a product that covers three quarters of it skews the result before any cost is entered.

Mixed options exist: buy then adapt, have a provider develop on an internal platform, buy for three years then take the work back in-house. Each enters the comparison as a separate option, with its own costs and risks.

The decision factors

The PMBOK Guide names five factors and adds the risks of each option; its Table 2-3 sets the advantages of insourcing against those of outsourcing. Each factor comes down to a question whose answer tips the balance towards one option.

The factors of make-or-buy analysis, based on §X4.3 and Table 2-3 of the PMBOK Guide.
FactorQuestion to settleFavours make whenFavours buy when
CapacityAre the competent people free during the required period?The team has time availableThe team is assigned elsewhere and the deadline cannot move
SkillsDoes the organisation know how to produce this deliverable?The expertise exists in-houseThe expertise is lacking in-house
Specialised expertiseDoes the work require a rare speciality, needed only once?The speciality will serve again in other projectsThe need is one-off
Permanent positionsDoes making require long-term hires?The positions created will have work after the projectThe need ends with the project
IndependenceMust the result come from someone outside the team?No independence requirementAudit, second opinion, penetration test
NoveltyDoes the deliverable already exist on the market?It requires substantial innovationIt is a standard good or service
Link to strategic advantageDoes the deliverable set the organisation apart from its competitors?It carries the value propositionIt does not; buying frees capacity and capital for what does
ControlHow much control must the organisation keep over the deliverable and its evolution?Direct control and oversight requiredA contract and a service level agreement are enough
CostWhich option costs least over the lifetime, in present value?The lifetime is long enough to recover the initial investmentThe horizon is short or the product is standard
RisksWhat risks does each option add, and who carries them?The dominant risk is dependence on a single supplier or data hosted abroadA market of sound suppliers can take on the delivery risk

A single factor can decide: an independence requirement rules out the in-house option, whatever its cost. The others are weighted. When several options remain after the eliminating factors, a multi-criteria decision analysis makes the weighting explicit and open to challenge.

The resource plan answers the capacity question. A competent team assigned to another project until March cannot deliver before March. The cost of the in-house option then includes what the other project loses.

Costing both options over the same horizon

Cost is compared over the expected lifetime of the deliverable. The first year misleads. A subscription looks cheap in year one and a development expensive, because one spreads its spending and the other concentrates it at the start. Total each option over the same horizon, with the same cost categories, as total cost of ownership does.

The cost categories to estimate for each option, over the same horizon.
CategoryMakeBuy
Set-upDesign, development and testing, at the fully loaded cost of the peopleLicence or one-off set-up fees, configuration, adaptation, integration with existing systems
Recurring costsCorrective maintenance and enhancements, operations, hostingSubscription or licence maintenance, price increases set in the contract
ManagementTeam supervisionContract and supplier management, service level monitoring
CapacityWork the team does not do elsewhere meanwhileInternal time to specify, run acceptance tests and administer
ExitKnowledge transfer if key people leaveData export, migration to another product, termination penalties

The first row hides a trap: the comparison sets a supplier's invoice, which contains all its costs, against the bare salary of internal staff. A person's fully loaded cost adds to the salary the employer's social contributions (OASI, occupational pension, insurance), the workstation, supervision and non-productive time. Counted at salary alone, the internal day makes the "make" option win by mistake.

The PMBOK Guide lists the financial techniques the analysis can use: payback period, return on investment, internal rate of return, discounted cash flow, net present value and cost-benefit analysis. The list is not exhaustive. The choice among them depends on the shape of the cash flows.

When both options deliver the same result, their benefits cancel out in the comparison and only the costs differ. The analysis then computes the difference between the two options, year by year: what making costs extra at the start, what it saves afterwards. The net present value of this difference shows which option costs less in today's francs. The payback period of the difference shows in how many years the extra investment pays for itself.

Discounting has a large effect on the result, because the two options have opposite profiles: making invests early, buying pays for a long time. An undiscounted total favours the option whose spending comes first. The discount rate is the return the organisation requires on its capital; the finance department supplies it. The result can flip from one horizon to another. A sensitivity analysis on the horizon, the rate and the subscription price shows at what value the decision tips.

Assessing the risks of each option

Making exposes the organisation to schedule and budget overruns, to dependence on the few people who know the code and to maintenance it carries alone for the deliverable's whole life. Buying exposes it to the supplier's failure or takeover, to dependence on the supplier's product and prices, to loss of control over the product's evolution and to data hosted outside the organisation, sometimes outside Switzerland.

If the product processes personal data, the Federal Act on Data Protection (FADP, art. 16) governs its disclosure abroad. It allows disclosure to a state whose legislation the Federal Council has recognised as adequate. Elsewhere, it requires appropriate safeguards, such as standard contractual clauses. The bid must therefore name the hosting country.

Table 2-3 of the PMBOK Guide counts the transfer of delivery risk to the supplier among the advantages of buying. The contract transfers that risk; the consequence stays with the project. If the supplier delivers late, the project misses its date and the penalty compensates only in money. The risk register therefore records what remains with the organisation under each option, with its response: exit clause, source code escrow, data export, trained internal backup.

Reversibility is costed with the rest. Reversing a purchase costs a migration and sometimes a termination fee. Reversing a development costs abandoning what was built. When the need is uncertain, prefer the more reversible option.

The decision and what follows

The decision belongs to whoever carries the budget and the consequences: the sponsor or the steering committee for a project deliverable, senior management for a lasting service. The PMBOK Guide places this choice in the Governance performance domain, because it decides who does the work and who answers for it. The analyst prepares the comparison and leaves the decision to them.

The decision record also states the assumptions behind the calculation and the event that would force a review of the decision: a price increase beyond a threshold, the end of maintenance, a change in requirements. A decision that is still readable two years later spares the team from redoing the analysis from memory.

"Make" leads to internal resource planning. "Buy" leads to the procurement strategy: the delivery method, the contract type and the procurement steps, which the PMBOK Guide describes in §X4.4. Then come the call for tenders, vendor assessment against the selection criteria and the contract.

For a public body, the "buy" branch follows the procedure that public procurement law sets according to the contract value: the Federal Act on Public Procurement (LMP/PPA) for the Confederation, the Intercantonal Agreement on Public Procurement (AIMP/IVöB) for the cantons and municipalities. The thresholds determine whether the contract is awarded directly, by invitation or through an open or selective procedure.

The trap reverses the order: a team that has already chosen a product, after a demonstration or a recommendation, writes the analysis afterwards to justify the purchase. Those who do not want the in-house option then estimate its costs. To avoid this trap, the team that would do the work estimates the effort of the in-house option. A party with no stake in the outcome consolidates both options and questions their assumptions: the analyst, the PMO or the finance department.

In a HERMES project

HERMES 2022 determines where the technique applies in a project, without naming it. In the initiation phase, the study describes solution options in enough detail to evaluate them, defines the evaluation criteria and their weighting, then evaluates the options. An option that develops the solution leads to the IT development scenario. An option that buys a market solution, adapts it and integrates it leads to the IT adaptation scenario.

If the project plans a purchase, the procurement analysis runs in parallel with the study. It describes what is to be procured, by whom and when, the state of the market and the applicable procurement procedure. The project coordinates it with the controlling and compliance bodies responsible for public procurement. Together, the study and this analysis form the basis of the decision on next steps. Make-or-buy analysis supplies the study with the comparison of options. According to HERMES, the documented evaluation shows the state of knowledge at the time of the decision: the decision record captures it.

AI considerations

A language model speeds up preparation. It drafts a checklist of cost categories for the type of deliverable. It extracts the prices, indexation clauses, commitment periods and exit conditions from a bid or from general terms and conditions. It spots a missing data-export clause in a contract. It writes the formulas for the discounted difference table and the variants of the sensitivity analysis.

It replaces neither the estimate nor judgement. The effort of the in-house option comes from the team that would do the work. A price comes from a bid, not from a market average a model has reconstructed. Senior management judges whether a deliverable carries the organisation's strategic advantage. Internal rates, salaries and supplier bids are confidential: a prompt to a hosted service takes them outside the organisation.

Examples

A technical services SME in the canton of Fribourg wants a customer portal for service requests. It compares an in-house development with a subscription-based online application that covers the same requirements after configuration. The discounted difference table is the financial deliverable of the analysis.

Discounted difference between make and buy, customer portal. Amounts in CHF, excluding VAT. Make: 200 person-days at a fully loaded cost of CHF 900, then CHF 26'000 a year, i.e. 20 person-days of maintenance (CHF 18'000) and CHF 8'000 for hosting and operations. Buy: configuration and ERP integration at a fixed price of CHF 40'000, then a subscription of CHF 50'000 and internal administration of CHF 6'000 a year. Internal time for specification and acceptance testing, exit costs: judged equivalent in both options and excluded. Discount rate 6%, five-year horizon.
YearMakeBuyDifference (make − buy)Factor at 6%Discounted differenceCumulative discounted
0180'00040'000140'0001.0000140'000140'000
126'00056'000−30'0000.9434−28'302111'698
226'00056'000−30'0000.8900−26'70084'998
326'00056'000−30'0000.8396−25'18859'810
426'00056'000−30'0000.7921−23'76336'047
526'00056'000−30'0000.7473−22'41913'628
Total310'000320'000−10'00013'628

The undiscounted total favours making, by CHF 10'000. Discounting reverses it: over five years, making costs CHF 13'628 more in present value. The simple payback period of the difference is 4.7 years (140'000 / 30'000); discounted, it exceeds five years. With a sixth year at the same annual difference, the cumulative total falls to about CHF −7'500 and making becomes the cheaper option. The decision therefore hinges on the horizon: the SME buys if it expects the portal to last five years at most.

Visualisations

The decision diagram places the analysis before the choice of supplier and shows where each branch leads. Extended beyond the horizon, the cumulative discounted difference shows the year in which it drops below zero. The decision therefore depends on the deliverable's expected lifetime. A sensitivity analysis then tests this tipping point against the rate and the subscription price.

Cost

PhaseLevelJustification
PreparationMediumScope the item, obtain an estimate from the internal team and at least one bid or verifiable market price.
ExecutionLowA cost table, a discounted difference calculation and a review of the factors, in one or two sessions.
DocumentationLowOne decision record per item, with the assumptions and the condition that triggers a review.

Tooling

A spreadsheet is enough for most analyses. The NPV and IRR functions of Excel or LibreOffice Calc compute the net present value and the internal rate of return of the difference. NPV discounts the first value in the range by one period: the year-0 amount is therefore added outside the function, otherwise the calculation is shifted by one year. A data table or the spreadsheet's scenario manager produces the sensitivity analysis on the rate, the horizon and the price.

The organisation's total cost of ownership model, if it has one, provides the cost categories and the internal rates. A weighted matrix in the spreadsheet breaks the tie between the options when several non-financial factors carry weight.

Sources

Lift chart and cumulative gains chart
All techniques
Market Analysis