The false opposition

There's no universal answer between SaaS and custom. The two approaches are complementary: most companies use a dozen SaaS daily (email, CRM, payroll, accounting, e-signature) and that's excellent. That said, some business bricks deserve custom development because they carry the differentiation of the company.

This article helps identify which bricks belong to SaaS and which deserve custom.

The basic rule — differentiator or commodity?

The central question: is this business process differentiating for your activity, or is it a commodity shared by all players in your sector?

Commodity → SaaS

Accounting, payroll, generic HR, email marketing, standard CRM, e-signature, helpdesk: these are identical processes from one company to another. A proven SaaS does better and cheaper than any custom build.

Buying Sage or QuickBooks for payroll doesn't make you less competitive — all companies in your sector have the same kind of solution.

Differentiation → custom (often)

Your own business logic, your specific way of managing clients or projects, your production process, your specific client relationship: that's what makes the difference vs your competitors. It's rarely well-served by generic SaaS.

Wanting to handle a specific workflow with a generic SaaS leads to two frequent outcomes:

  1. You bend to SaaS constraints and lose specificity (and value)
  2. You hack with custom fields, jury-rigged integrations, hourly Zapier scripts — and inherit a fragile gas plant

The 5-criteria grid

Criterion 1 — Volume of users and transactions

SaaS advantaged: low volume (few users, few hundred transactions/month). Custom cost is disproportionate.

Custom advantaged: high volume (hundreds of users, thousands/millions of transactions). SaaS fees scale quickly (often per user or transaction) and custom becomes economically viable.

Criterion 2 — Process specificity

SaaS advantaged: standard process, identical to what all your competitors do.

Custom advantaged: unique process, which constitutes your know-how or competitive edge.

Criterion 3 — Integration with the rest of the IT system

SaaS advantaged: limited integration needs, or standard integrations via documented APIs.

Custom advantaged: strong interactions with other bricks of your system, data flowing in real time between modules, complex dependencies.

Criterion 4 — Data ownership and sensitivity

SaaS advantaged: standardized data, low sensitivity, compliance covered by SaaS.

Custom advantaged: strategic data (your client base, catalog, algorithm), sovereignty matters, specific regulatory compliance.

Criterion 5 — Predictable evolution

SaaS advantaged: stable need or evolution within the editor's roadmap.

Custom advantaged: frequent, sometimes unpredictable evolutions, that don't depend on a third-party editor's calendar.

The 4 typical cases

Case 1 — SaaS is the obvious choice

An SME needs a CRM to track prospects and clients. Modest volume, standard process, important but well-served data. → HubSpot, Pipedrive, Salesforce depending on size. Custom is absurd.

Case 2 — Custom is the obvious choice

An SME has a production process with very specific business rules, multiple stakeholders, complex temporal dependencies, and 50+ projects managed in parallel. Market ERPs cost €50-200k/year in licenses, require 6-12 months of configuration, and never cover 100% of the specific need. → Custom: €60-120k initial, €12-18k/year maintenance, covers 100% of needs and evolves with the activity.

Case 3 — Relevant hybrid

A marketplace uses custom for its core business (directory, matchmaking, moderation) and plugs SaaS for commodity functions: Stripe for payment, SendGrid for transactional email, Mixpanel for analytics. It's the healthiest configuration: custom where there's differentiation, SaaS where there's commodity.

Case 4 — Migration from a stifling SaaS

A company has used for 3 years a SaaS that no longer covers 30% of its needs. Annual cost: €12k. Cost of workarounds (Zapier, in-house scripts, manual re-entry): €25k/year of lost time. Migration to custom: €40k initial. ROI: 2 years. → Often profitable, but to be validated precisely.

SaaS traps

1. Price dependency

A SaaS that costs €50/user/month for 10 users (€6k/year) costs €60k/year for 100 users. If your activity scales, your SaaS cost follows — often linearly, no amortization.

2. Vendor lock-in

Your data lives at the editor's. Exiting is never trivial: limited exports, proprietary formats, features not reproducible elsewhere. The more you're committed, the more expensive it is to leave.

3. The editor's pace

A feature critical for you may never arrive — or arrive in 3 years. You don't control the roadmap.

4. Imposed evolutions

The editor can change its interface, pricing, model, sometimes even close the product. You take what comes.

Custom traps

1. Initial over-investment

Building custom what already exists in SaaS costs more for a worse result. Never redo what exists and is well-done elsewhere.

2. Underestimated maintenance

Custom demands ongoing maintenance. Budget 15-25% of initial cost per year. Otherwise, the product drifts and loses value.

3. Vendor dependency

If you have only one vendor who knows your code, you have a breaking point. Mitigation: clean and documented code, access to your Git repo, training an internal reference person or choosing a vendor whose work you can audit.

4. Over-specification

Wanting to cover EVERYTHING in V1 explodes budgets and timelines. Better an MVP at €30k that ships in 3 months and evolves, than a dreamed project at €150k that ships in 18 months and is already obsolete.

The practical rule

| Type of need | Reflex | |---|---| | Accounting, payroll, tax | SaaS always | | Generic CRM | SaaS in 90% of cases | | Email marketing, newsletter | SaaS (Brevo, MailerLite, ConvertKit) | | Helpdesk, customer support | SaaS (Crisp, Intercom, HelpScout) | | Analytics, A/B testing | SaaS (Plausible, Posthog, Mixpanel) | | Document management | SaaS most often (Notion, Confluence) | | Specific business platform | Custom | | Marketplace, directory | Custom | | Critical internal tool | Custom if volume + specificity | | Client area / portal | Custom if differentiating | | Configurator, simulator, quote | Custom if proprietary logic |

  1. Everything that's commodity → SaaS (10 to 15 SaaS total in an SME, that's normal)
  2. The differentiating core business → custom (1 to 3 bricks depending on company size)
  3. The two communicate via clean integrations (webhooks, API, sometimes a light middleware)

It's the configuration of all companies that have a high-performing product platform: custom where it matters, SaaS everywhere else.

Going further

Evaluating custom relevance for a specific need? Let's discuss — I'll orient you honestly, including toward a SaaS if it's the best option.