Custom build vs. SaaS: when does each one make sense for African SMEs?
Off-the-shelf SaaS is faster and cheaper to start. Custom builds fit your business better and compound in value over time. Here's a clear decision framework — and the situations where each one is the obvious wrong choice.
The “build or buy” decision is one of the most consequential calls a growing business makes. Get it right and your operations compound for years. Get it wrong and you’re either paying SaaS fees for software you’ve outgrown, or sunk a year and tens of millions of Naira into a custom build that doesn’t actually outperform the off-the-shelf option you skipped.
We’ve built both. We’ve migrated clients between them in both directions. Here’s the honest framework we use to advise on the decision.
The two extremes, stated plainly
SaaS means you rent software from a vendor who built it for thousands of customers like you. Examples: QuickBooks, Salesforce, HubSpot, Asana, Odoo Online, Shopify.
Custom build means you (or your partner) design and develop software that fits your exact business — built once, owned by you. Examples in our portfolio: E5 Solstice (solar ERP), SwiftRoute (logistics platform), DevOS (multi-tenant SaaS for real estate developers).
A third option that often gets overlooked: configurable platforms — Odoo Enterprise, NetSuite, Salesforce — sit in the middle. They’re SaaS products you heavily customize. Most “custom build” decisions in the SME market are actually configurable-platform decisions in disguise.
When SaaS is obviously the right choice
Pick SaaS without hesitation if:
1. Your process is generic, not differentiating
If you’re running a standard B2B sales pipeline, standard double-entry accounting, standard project management — these are solved problems. Thousands of companies do them exactly the same way. The SaaS vendor has spent more on UX than your custom build ever will. Buy it.
Test: can you describe your need in five words a vendor would understand? “We need a CRM.” “We need accounting software.” “We need a help desk.” If yes — buy.
2. You need it live in weeks, not quarters
SaaS deploys in days. Custom takes months at best. If the cost of waiting is bigger than the cost of imperfect fit, that’s SaaS territory.
3. The total contract value is under ~₦5M per year
The math: a basic custom build starts around ₦15-25M to deliver and another ₦3-8M/year to maintain. If your needs would otherwise cost less than ₦5M/year on SaaS, custom never pays back.
4. Your business is changing too fast to specify
If you genuinely don’t know what your operations will look like in 12 months, freezing requirements into a custom build is dangerous. SaaS lets you change your mind cheaply.
5. You don’t have technical leadership in-house
Custom requires somebody who can hold a vendor accountable, evaluate trade-offs, and make decisions about architecture and scope. If you can’t put a competent person on the buyer side of the table, custom builds spiral. SaaS doesn’t require that role.
When custom is obviously the right choice
Build custom (or heavily configure a platform) if:
1. The process IS the business
If your competitive advantage is how you handle a specific workflow — and that workflow doesn’t look like anyone else’s — SaaS will average you into mediocrity. A logistics company whose differentiation is real-time visibility can’t run on the same generic dispatch tool every competitor uses.
Test: would a customer notice if your operations behaved like every other player in your space? If yes, you need software that mirrors your specific advantage.
2. You’re paying enough SaaS to justify the investment
A useful threshold: when SaaS subscriptions cross ₦15M/year and the system still doesn’t fit, the math starts favoring custom. At ₦30M/year and growing, the ROI on a custom build is usually overwhelming.
3. You’re hitting integration walls
Mid-sized businesses often end up with 6-10 SaaS tools that don’t talk to each other. Each one was good in isolation. Together they create a data nightmare. A custom build that consolidates those into one platform — owned by you, integrated by design — recovers operational sanity.
This is exactly what DevOS does for real estate developers: consolidates finance, HR, governance, and operations onto a single platform with strict tenant isolation, eliminating the spreadsheet-and-email tax of running on disconnected SaaS.
4. You need data sovereignty or regulatory control
Some workloads can’t reasonably run on foreign SaaS:
- NDPC-regulated personal data with cross-border transfer constraints
- CBN-regulated financial workflows with audit trail requirements
- Government parastatal data that’s required to remain on-soil
- Health data, regulated education records
The cost of custom is rounding error compared to the regulatory risk of getting this wrong.
5. You’re building a SaaS yourself
If your business is software — selling a platform to customers — then custom isn’t optional. The question is whether you partner with someone who’s shipped multi-tenant SaaS before or learn the hard way.
The expensive middle ground: when neither is clearly right
Most SME decisions don’t fall cleanly into either bucket. They sit in the middle: SaaS sort of fits but not really, custom would fit perfectly but feels like overkill. Here’s how to think about it.
Try in this order
-
Off-the-shelf SaaS with default configuration. If it works, ship it. Done.
-
Configurable platform — Odoo, NetSuite, Salesforce — heavily configured to your needs. This catches 60-70% of SME use cases that “feel like custom” but aren’t.
-
Configurable platform + small custom layer for the 10-20% that’s truly differentiating. This is the sweet spot for most growing businesses: leverage someone else’s product where it’s generic, build what’s genuinely yours.
-
Fully custom build only when steps 1-3 demonstrably can’t carry the workload.
The mistake we see most often: skipping straight to step 4 because it sounds more impressive, or because a developer sold the idea of “having your own platform.” That decision routinely costs 5-10× more than necessary.
The opposite mistake: staying on step 1 long after the business has clearly outgrown it, accumulating SaaS sprawl, then trying to fix everything at once with a moonshot custom rewrite.
The 3-year cost comparison
For a 100-300 person business with moderately complex needs:
| Path | Year 1 | Year 2-3 | 3-Year Total |
|---|---|---|---|
| Pure SaaS sprawl (6 tools) | ₦18M | ₦40M | ₦58M + integration pain |
| Configurable platform (e.g. Odoo Enterprise) | ₦25M | ₦18M | ₦43M |
| Configurable + small custom | ₦35M | ₦20M | ₦55M (huge ROI gain) |
| Fully custom | ₦60M | ₦25M | ₦85M (only worth it if differentiating) |
The “configurable + small custom” path usually wins because it captures the SaaS economics for generic work and the custom advantage for differentiating work, with neither’s downsides.
Three questions that resolve most build-vs-buy decisions
When advising clients on this call, three questions clarify 80% of cases:
1. “If we picked the wrong tool, how much does it cost to switch?” SaaS: cheap to switch (low lock-in is the point). Custom: expensive to switch (you own it). Higher switching cost → more upfront diligence → favor SaaS for ambiguity, custom only for high-confidence cases.
2. “Will customers ever interact with this system directly?” If yes → you need full control of UX. Strongly favors custom or heavily-configured platform. If no (purely internal) → SaaS is usually fine.
3. “Does our SaaS bill exceed our salary for a mid-level engineer?” At that point, an engineer can usually save you money plus build differentiation. Below it, the SaaS economics dominate.
What we’d advise if you asked us today
For most growing African SMEs we work with: start with a configurable platform (Odoo, NetSuite, or similar) and only build custom for the genuinely-differentiating layer. This captures 80% of the value of a custom build at 30-40% of the cost. (Before you take a vendor pitch on Odoo or NetSuite, read the real cost of ERP in Nigeria — the per-user license is the smallest line item.)
The exceptions, in our portfolio:
- Industry-specific workflows that no platform models well — solar asset lifecycles spanning 25 years (we built E5 Solstice for exactly this), real-time logistics dispatch with event-driven architecture (SwiftRoute), multi-tenant SaaS where isolation is critical (DevOS — read the architecture write-up)
- Regulated workloads where data sovereignty trumps cost (Nigerian fintechs in particular — NDPC compliance shapes the build vs. buy math)
- Software products where the product is the business
If your situation fits one of those, custom is worth the investment. If it doesn’t, configurable platforms have probably caught up to your needs since the last time you evaluated them. If you’d like a no-pitch reality check on which side your situation falls, our Design & Build practice is exactly this conversation.
If you’d like an honest second opinion on a specific build-vs-buy decision — including a no-pitch read on whether your case is genuinely in the “build” bucket — start a conversation. We’ve made the call (correctly and incorrectly) enough times to have strong opinions and no skin in selling you either path.