Should I build, buy, or hire?
Your checkout flow drops 12% of customers at the payment step. Your engineer estimates 6 weeks to build a custom integration. Stripe quotes $0.30 per transaction plus 2.9%. A payments specialist on LinkedIn wants $185K/year. Same problem, three paths - and each one hits your P&L on a different line, on a different timeline.
Build, Buy, or Hire is a Capital Allocation decision: should you spend engineering time to create a capability, purchase it from a vendor, or bring in a person who already has it? The right answer depends on your Time Horizon, the opportunity cost of your existing team, and whether the capability creates differentiation or is a Commodity.
Every capability gap in your business can be closed three ways:
The decision is not "which is cheapest." It is "which generates the most ROI given my Time Horizon, risk appetite, and strategic needs."
Each path hits your P&L differently, and the wrong choice compounds:
Build front-loads cost. If the build succeeds, you own an Asset that can become a competitive moat. If it fails, you burned Cash Flow and time with nothing to show.
Buy creates ongoing Fixed Obligations. A SaaS tool at $2K/month is predictable, but it is a permanent line item. Over a 3-year Time Horizon, that is $72K - possibly more than building it. And you do not own the Asset. The vendor controls the roadmap, the Pricing, and part of your Cost Structure.
Hire adds permanent overhead to your Operating Statement. A $150K hire might cost $195K once you include benefits and overhead. That is a Fixed Obligation regardless of output. But a strong hire brings Throughput that a tool cannot - judgment, adaptation, and the ability to solve problems you have not identified yet.
Operators who default to one path without running the comparison destroy value. Builders default to build. Finance people default to buy. Neither is thinking about Capital Allocation.
Evaluate each option on four dimensions:
How fast does each path deliver a working capability?
Do not compare upfront costs. Compare the Expected Total Cost over the period you need the capability. When computing costs that include people, always use the full annual cost: salary plus benefits plus overhead (payroll taxes, equipment, office space). This is the real number your Operating Statement reflects.
This is the most important dimension.
What else could those resources be doing? If your engineering team is at capacity and the Bottleneck is shipping Revenue features, pulling engineers off to build an internal tool has a measurable cost: the features they would have shipped, the Revenue those features would have generated, and the Time to Value delay on everything else in the Pipeline. Each engineering hour has a Shadow Price equal to the best alternative use of that capacity.
Run this analysis every time you identify a capability gap. The decision tree:
You do not always have to pick one path. When a capability is a source of differentiation but you need it immediately, Buy to close the gap, then Build or Hire for long-term ownership. This is investment sequencing applied to capabilities: the vendor bridges the gap while you invest in the permanent solution. Set Exit Criteria upfront - define when the custom solution must be ready and what triggers the switch. The best Operators treat Build, Buy, and Hire as stages in a sequence, not a single choice.
Your team needs a reporting dashboard to track operational metrics. You sell 10,000 units/month. Option A: Two engineers build it over 8 weeks. Each engineer costs $180K/year (salary, benefits, and overhead), or $3,460/week. Option B: Buy a SaaS analytics tool at $1,500/month. Time Horizon: 3 years. Your engineering team would otherwise be shipping a feature expected to generate $8,000/month in Expansion Revenue.
Build cost: 2 engineers × 8 weeks × $3,460/week = $55,360 upfront Implementation Cost. Annual maintenance at ~25% of build cost = $13,840/year. 3-year Expected Total Cost: $55,360 + (3 × $13,840) = $96,880.
Buy cost: $1,500/month × 36 months = $54,000 baseline. With 10% annual rate increases: Year 1 = $18,000, Year 2 = $19,800, Year 3 = $21,780. Adjusted total: $59,580.
Opportunity cost of Build: Those 2 engineers would otherwise ship a feature generating $8,000/month in Expansion Revenue. 8 weeks of delay = $16,000 in delayed Revenue. Adjusted build cost: $96,880 + $16,000 = $112,880.
Decision: Buy at $59,580 vs Build at $112,880. An analytics dashboard is a Commodity - your competitive moat is not your dashboards. Buying saves ~$53K over 3 years and delivers 6 weeks faster Time to Value.
Insight: When the capability is a Commodity, building it does not just cost more in dollars - the opportunity cost of your engineers' time often doubles the real price. The break-even point where Build wins here would require a Time Horizon beyond 5 years, and by then the dashboard needs a full rebuild anyway (Obsolescence).
You process 50,000 completed transactions/month at $80 average order value ($4M/month Revenue). Your checkout drops 12% of customers at the payment step - meaning roughly 56,800 customers attempt payment each month but only 50,000 complete it. Option A: Build a custom payment flow (1 senior engineer, 6 weeks, $210K/year including salary, benefits, and overhead). Option B: Switch to Stripe's optimized checkout at $0.30 + 2.9% per transaction (current processor: $0.25 + 2.5%). Option C: Hire a payments specialist at $185K/year (salary, benefits, and overhead). Time Horizon: 2 years.
Current processing cost: 50,000 × ($0.25 + $80 × 0.025) = 50,000 × $2.25 = $112,500/month.
Buy (Stripe): Per-transaction cost: $0.30 + ($80 × 0.029) = $2.62. If Stripe's optimized checkout recovers half the 12% drop (6 percentage points), the completion rate rises from 88% to 94% of the 56,800 attempted transactions. New completions: 56,800 × 0.94 ≈ 53,400 - roughly 3,400 extra transactions. Recovered Revenue: 3,400 × $80 = $272,000/month. New processing cost: 53,400 × $2.62 = $139,900/month, an increase of $27,400 over current. Net monthly gain: $272,000 - $27,400 = $244,600/month. Time to Value: ~2 weeks.
Build: 6 weeks of a $210K/year engineer = ~$24,200 Implementation Cost. Assume the same recovery and $272,000/month gain - but Time to Value is 6 weeks vs 2 weeks. Convert the monthly recovery to weekly: $272,000 / 4.33 ≈ $62,800/week. The 4 extra weeks of delay cost ~$251,000 in Revenue you would have captured under Buy. With Execution Risk (payment integrations commonly run 2-3x over estimate), a 12-week build means 10 extra weeks of delay: ~$628,000.
Hire: $185K/year = ~$15,400/month ongoing. Time-to-Fill of 6-8 weeks plus 4 weeks to reach full Throughput = 10-12 weeks before any impact. Uncertain recovery rate. Longest Time to Value of all three options.
Decision: Buy immediately. The $27,400/month processing cost increase is small against $272,000/month in recovered Revenue. The Build option's delay cost alone (~$251,000) exceeds its Implementation Cost by 10x. Re-evaluate Hire in 6 months if you need ongoing payment optimization beyond what the vendor provides. Payment processing is a Commodity - buy the capability, free your engineer to work on differentiation.
Insight: When the Revenue impact of delay is large, convert all figures to the same time unit before comparing. The weekly delay cost ($62,800) makes Time to Value the dominant factor - not Implementation Cost or Cost Per Unit. Operators who fixate on cost increases miss the Revenue side of the P&L.
The default for builders is to build. As an Operator, your job is to override that instinct with Capital Allocation math - sometimes buying a Commodity at a higher Cost Per Unit is the highest-ROI move because it frees your team to work on differentiation.
Always compare Expected Total Cost over your Time Horizon, not upfront price. Include the opportunity cost of engineering time - those hours have a Shadow Price equal to the best alternative use of that capacity.
The differentiation vs Commodity question is the tiebreaker. If it is your competitive moat, build it and own the Asset. If everyone needs it, buy it and ship faster.
Building Commodities: Engineers love building. But building your own email system, payment processor, or analytics platform when mature vendors exist is a Capital Allocation failure - you are converting scarce engineering capacity into an Asset that does not differentiate you, while your actual competitive moat starves for attention.
Ignoring the opportunity cost of engineering time: Comparing the dollar cost of Build vs Buy without accounting for what those engineers would otherwise produce. An engineer costs $180K/year whether they are building an internal tool or shipping Revenue-generating features. The question is which use generates more ROI - and the answer requires knowing what else is in the Pipeline.
Your 5-person engineering team is fully utilized. A critical internal workflow takes 4 hours/week of manual work from your Operations team (1 person at $65K/year). You can: (A) build an automation tool - 3 weeks of 1 engineer at $160K/year (salary, benefits, and overhead), (B) buy a workflow tool at $400/month, or (C) keep the manual process. Your Time Horizon is 2 years. The engineer you would pull off is currently building a feature expected to generate $5,000/month in Revenue. Which do you choose and why?
Hint: Calculate the cost of the manual process over 2 years first. Then compare Build and Buy on Expected Total Cost. Do not forget the opportunity cost of pulling an engineer off their current project - convert the monthly Revenue impact to weeks before multiplying by delay duration.
Manual cost: 4 hours/week at $65K/year ≈ $31.25/hour × 4 = $125/week = $6,500/year. Over 2 years: $13,000.
Buy: $400/month × 24 months = $9,600, plus ~1 week of engineer time for integration = $3,077. Total: $12,677. Time to Value: ~2 weeks.
Build: 3 weeks of engineer at $160K/year = $9,231 Implementation Cost. Maintenance at 25%/year = $2,308/year. 2-year total: $9,231 + (2 × $2,308) = $13,847. Plus opportunity cost: 3-week delay on a $5,000/month feature = $5,000 / 4.33 × 3 ≈ $3,460 in delayed Revenue. Adjusted: $17,307.
Buy wins at $12,677 vs Build at $17,307 vs status quo at $13,000. The real insight: status quo ($13,000) is close to Buy ($12,677). If the workflow tool also reduces defect rate or improves Throughput beyond pure time savings, Buy is clearly best. If it is purely time savings, even doing nothing is defensible - this is not worth pulling an engineer off Revenue work to build.
You run a marketplace with 200,000 monthly active users. Your recommendation algorithm is basic (rule-based) and your competitor just launched ML-powered recommendations. You can: (A) build an ML recommendation engine - 2 engineers, 4 months, $180K/year each (salary, benefits, and overhead), (B) buy a recommendation API at $0.002 per request (each user makes ~20 requests/month), or (C) hire an ML engineer at $220K/year (salary, benefits, and overhead). Your Time Horizon is 2 years. Which path and why?
Hint: This is a differentiation question first, a cost question second. Ask: is the recommendation engine a Commodity or your competitive moat? Then consider what happens to your Market Share if you do not act. Finally, consider investment sequencing - can you stage the options instead of picking just one?
Buy (API): 200,000 users × 20 requests × $0.002 = $8,000/month = $96,000/year. 2-year total: $192,000. Time to Value: 2-4 weeks. But every competitor can buy the same API - zero differentiation.
Build: 2 engineers × 4 months × $15,000/month each = $120,000 Implementation Cost. Maintenance ~$30K/year. 2-year total: $180,000. High Execution Risk - ML projects routinely miss timelines by 2-3x.
Hire: $220K/year × 2 years = $440,000. Most expensive on paper, but brings Knowledge Capital that adapts continuously.
Best answer: stage them. This is investment sequencing. Buy the API immediately ($8K/month) to close the competitive gap in 2 weeks. Simultaneously start hiring an ML engineer. Over months 3-12, the ML hire builds a custom engine using your proprietary user data, creating a Data Moat the API cannot replicate. Discontinue the API once the custom engine is live. 2-year cost: ~$72K (API for 9 months) + $440K (hire) = $512K.
This is expensive, but recommendations are the competitive moat in a marketplace. If better recommendations lift conversion by just 5% on 200,000 users, that is thousands of additional transactions per month - likely millions in incremental Revenue over 2 years. You build what differentiates you. You buy what bridges the gap while you build.
Build, Buy, or Hire is where opportunity cost becomes an operational decision. Every dollar and every hour allocated to one path cannot work somewhere else - this is the Capital Allocation tradeoff applied to capabilities. The decision depends on Time to Value: when Revenue is at stake, speed often dominates cost. Downstream, it feeds into Capital Budgeting (how you evaluate and approve these investments), NPV analysis (comparing the present value of each option's costs and benefits), and Payback Period (how quickly each option recoups its investment). It shapes your Cost Structure: Build can produce Capital Assets (engineering costs may be capitalized on your Balance Sheet under applicable accounting rules, though many companies expense them), Buy creates recurring expenses on your Operating Statement, and Hire adds to your Labor line as a Fixed Obligation.
Disclaimer: This content is for educational and informational purposes only and does not constitute financial, investment, tax, or legal advice. It is not a recommendation to buy, sell, or hold any security or financial product. You should consult a qualified financial advisor, tax professional, or attorney before making financial decisions. Past performance is not indicative of future results. The author is not a registered investment advisor, broker-dealer, or financial planner.