You've got a problem. Maybe it's inventory tracking that's leaking money. Maybe it's a support team drowning in tickets. Maybe it's a workflow that runs on WhatsApp and Google Sheets.
The question hits: do we build custom software, or do we buy something off the shelf?
Every founder, every operator, every CTO in Nigeria asks this question. And most get it wrong — not because they're lazy, but because they're guessing. They're comparing vibes, not numbers.
Here's the framework we've used across 16+ client engagements to make this decision actually defensible.
The Real Cost Problem
Before we get to the questions, you need to understand what you're actually comparing.
Buying looks cheap on day one. A SaaS subscription costs ₦50k–₦500k per month. That's nothing next to a custom build.
But buying compounds. Per-seat pricing. Feature gateups. Integration fees. Migration costs when the vendor gets acquired or jacks up renewal prices. The average SaaS stack grows 20–37% in cost at renewal when AI features are added.
Building looks expensive upfront. A custom internal tool in Nigeria runs $30k–$150k depending on complexity. That's real money.
But building settles. Year 2 and beyond, you're paying 15–25% of initial cost for maintenance. No per-seat inflation. No vendor lock-in. No surprise renewal hikes.
The real comparison isn't "₦50k/month vs. ₦10m upfront." It's 5-year total cost of ownership — including maintenance, integration, migration risk, and the opportunity cost of your engineers building commodity features instead of your competitive advantage.
The 7 Questions
Score each question. At the end, tally your answers. The framework does the rest.
Question 1: Is This How You Win Customers?
Ask: Would a customer notice if we did this differently than our competitor?
If the answer is yes — if this workflow is part of why customers choose you over alternatives — you should seriously consider building.
If the answer is no — if this is a backend function that no customer ever sees or cares about — buying almost always wins.
Example: Your booking system that offers instant confirmations with custom pricing logic? That's how you win. Your payroll system? Nobody's choosing you because of how you pay staff.
Question 2: How Many People Use This Daily?
Ask: How many team members interact with this system every single day?
This is where the math gets real.
| Daily Users | SaaS Annual Cost (₦2k/user/mo) | 5-Year SaaS Cost | Custom Build (₦35m) | 5-Year Build Cost |
|---|---|---|---|---|
| 5 | ₦120k | ₦600k | ₦35m | ₦48m |
| 20 | ₦480k | ₦2.4m | ₦35m | ₦48m |
| 50 | ₦1.2m | ₦6m | ₦35m | ₦48m |
| 100 | ₦2.4m | ₦12m | ₦35m | ₦48m |
| 200 | ₦4.8m | ₦24m | ₦35m | ₦48m |
| 500 | ₦12m | ₦60m | ₦35m | ₦48m |
Rule of thumb: Below 20 users, SaaS almost always wins. Above 100 users, custom starts winning — especially if the SaaS charges per seat.
Question 3: What's the 3-Year Vendor Risk?
Ask: If this vendor gets acquired, pivots, or triples their price in 18 months — what happens to our business?
Score vendor risk on four vectors:
- Pricing stability — Do they have flat pricing, or consumption-based that can spike?
- Acquisition exposure — Is this a company that could be bought and gutted?
- Data lock-in — Can you export your data cleanly if you leave?
- Feature gating — Are the features you need locked behind higher tiers?
High vendor risk = stronger case for building.
We've seen Nigerian businesses burned by this. A logistics company switched to a SaaS fleet tracker, then the vendor changed pricing 6 months later. No export path. They rebuilt from scratch anyway — except now they had 6 months of lost data.
Question 4: Can You Own This for 5 Years?
Ask: Do we have (or can we hire) a technical partner who will maintain this for the long haul?
Custom software without long-term ownership is worse than buying. You'll end up with expensive abandonware — a system nobody understands, nobody maintains, and nobody wants to touch.
Before you commit to building, make sure you have:
- An in-house dev team that understands the stack, OR
- A trusted agency relationship (like Meshgryd) that will be there in year 5, OR
- A clear handover plan with documentation and knowledge transfer
If you don't have any of these three, buy. Come back to building when you do.
Question 5: Does This Process Change Frequently?
Ask: How often does this workflow change?
| Change Frequency | Recommendation |
|---|---|
| Rarely (payroll, invoicing) | Buy. Stable processes don't need bespoke tools. |
| Quarterly (your internal ops) | Build or heavily customize. Off-the-shelf can't keep up. |
| Constantly (still figuring out the business) | Buy. Don't freeze a process you haven't settled on yet. |
This one trips up startups. You're still validating your business model, but you want to build a custom operations platform. Don't. Use duct-tape and spreadsheets until the process stabilizes, then build the thing that matters.
Question 6: What Happens If the Tool Disappears Tomorrow?
Ask: If this vendor shuts down in 18 months, what's our recovery plan?
- "We switch to an alternative in a week" — Buying is fine.
- "We lose 3 months of operations" — Buying is risky.
- "We lose the business entirely" — Build. Ownership matters more than convenience.
This is the question most people forget to ask. They evaluate the tool's features, not the tool's absence.
Question 7: Does AI-Coding Collapse the Build Cost?
Ask: Can AI-assisted development cut the build time and cost enough to change the math?
This is the 2026 variable that didn't exist three years ago.
AI coding tools have compressed build estimates by 60–80% for internal tools, dashboards, and workflow automation. What used to take 12 weeks now takes 2–4 weeks with an experienced team.
The new break-even rule: If a custom build can ship a working version in under 90 days that solves 70% of the use case, the build side wins.
But not everything benefits equally. Payment rails, compliance systems, and infrastructure with deep certification requirements still need traditional development timelines.
The Scorecard
Add up your answers:
| Question | Build Score | Buy Score |
|---|---|---|
| Q1: Competitive differentiator? | Yes = +3 | No = +3 |
| Q2: Under 20 daily users? | — | Yes = +3 |
| Q3: High vendor risk? | Yes = +2 | Low = +3 |
| Q4: 5-year technical ownership? | Yes = +3 | No = +3 |
| Q5: Process changes quarterly? | Yes = +2 | Rarely = +2 |
| Q6: Tool disappears = business lost? | Yes = +3 | Manageable = +3 |
| Q7: AI cuts build to under 90 days? | Yes = +2 | No = +2 |
Interpretation:
- Build score ≥ 12: Build custom. You have a clear case.
- Buy score ≥ 12: Buy SaaS. Stop overthinking it.
- Both scores 8–11: You're in blend territory — buy the commodity core, build your differentiating layer on top.
The Hybrid Approach Most Miss
The build-vs-buy framing is a false binary. The best answer is usually a mix:
Buy the commodity, build the differentiator.
- Buy Stripe for payments. Build your custom billing logic on top.
- Buy a CRM. Build your unique workflow engine that feeds it.
- Buy infrastructure. Build the AI agent that acts on your private data.
Gartner's research shows 76% of enterprise software spend now flows into blended stacks — not pure build, not pure buy.
The Nigerian Context
This framework matters more in Nigeria because the stakes are different:
- Vendor lock-in is harder to escape. Payment gateways, data residency, and connectivity issues make migration painful.
- Per-seat pricing hits harder. When your team is 50 people paying $50/seat/month, that's $36k/year — real money in naira.
- Technical talent is available but scarce. Finding engineers who'll maintain your system for 5 years is harder than finding ones who'll build it once.
- Infrastructure costs are unpredictable. Hosting, bandwidth, and uptime considerations are different from Silicon Valley assumptions.
Meshgryd was built for this context. We don't just build — we push back on bad specs, design for how you'll operate in 5 years, and ship with observability, logging, and rollback built in.
Next Steps
- Run the 7 questions on your most painful workflow this week
- Score it honestly — don't let your preference for "building" or "buying" bias the result
- If you're in blend territory, you need a partner who can build the custom layer while you buy the commodity core
Meshgryd offers a free systems audit — we'll map your current stack, score each component, and show you exactly where to build, where to buy, and where to blend.
Meshgryd Systems builds intelligent digital infrastructure for Nigerian businesses. Software engineering, AI & automation, IoT systems. 16+ companies served. Predictable delivery is a design decision.