The decision between purchasing a SaaS subscription and building (or commissioning) a custom system to support the same workflow.
A build-vs-buy decision depends on five variables: workflow stability, integration depth, vendor moat, cost trajectory and compliance requirement. NodeSparks publishes two free decision tools: the five questions on the services page for intake triage and the six-row matrix inside the SaaS Replacement Playbook for the full scoping call.
See also/glossary#workflow-specificity A measure of how closely a team's actual workflow matches the path a SaaS tool was designed for. High specificity = strong fit. Low specificity = the team is fighting the tool.
High workflow specificity (the team uses 60%+ of the SaaS features as designed) is a buy signal. Low specificity (under 20% of features used, with workarounds patching the rest) is a build signal. It is the first row of the six-row matrix in the SaaS Replacement Playbook.
See alsoThe defensible value a SaaS vendor provides beyond UI and API access — typically unique data, compliance posture, marketplace effects, or network effects.
Vendor moat is the bar against which "could we rebuild this" is measured. Apollo has a moat via its B2B contact database. BILL has a moat via its payments network. A scheduler that wraps Google Calendar with a nicer UI has no moat. Build-vs-buy decisions hinge on whether a credible moat exists; if no moat, the SaaS is a UI layer the team is renting at a markup.
See also/glossary#workflow-stability A measure of whether a workflow has changed materially in the last 6+ months. Stable workflows are good rebuild candidates; volatile ones are not.
Workflow stability is the cheapest sanity check before any build-vs-buy decision. If the workflow has been settled for 6+ months without major requirement changes, the engineering hours invested in rebuilding pay back. If the workflow is still being defined, building is rebuilding the rebuild within months — burn engineering capacity that should be on product work.
See also/glossary#cost-trajectory The direction and slope of a SaaS subscription's recurring cost over the next 24 months — flat vs growing linearly with seats or usage.
Cost trajectory is a build-vs-buy signal. A flat ~$80/month subscription does not move the math; a per-seat tool growing 30% YoY with headcount eventually does. A stack costing more than €500 a month in total, growing linearly, is the canonical replacement candidate.
See also/glossary#saas-replacement The practice of identifying single-purpose SaaS tools in a stack and rebuilding the specific workflows they support as custom systems the team owns.
SaaS replacement is not cloning a vendor product. It is identifying which 20-30% of a SaaS tool the team actually uses, rebuilding that specific workflow on owned infrastructure, and discontinuing the subscription. Typical candidates: outreach platforms, schedulers, content engagement, basic CRMs, scraping subscriptions. Typical non-candidates: payroll, accounting, e-signature with legal weight — anywhere the vendor compliance posture is the product.
See alsoThe accumulation of overlapping or under-used SaaS subscriptions in a team's stack over time, typically growing faster than headcount or actual workflow needs.
SaaS sprawl is the structural pattern where teams add subscriptions faster than they retire them. Tracked across mid-market companies, SaaS spend has grown 4-6x faster than headcount over the last five years. Sprawl shows up as feature overlap (CRM, sales engagement tool, and lifecycle marketing all storing the same contact data) and as paid-for-but-unused seats.
See alsoThe recurring monthly cost a team pays to keep a workflow running across SaaS tools that markup the underlying compute and API calls.
Automation tax is the gap between what a workflow actually costs to execute (cents in API calls + cents in compute) and what the SaaS bill is for the same workflow. At SME scale this gap is typically 10-100x. The tax is paid in exchange for managed UI, integrations, and not having to run infrastructure — legitimate for some teams, severe enough to justify rebuild for others.
See alsoA software agent that automates outbound sales prospecting — list-building, personalization, send sequencing — at a per-contact or per-agent SaaS price.
AI SDR vendors (AiSDR, Artisan, 11x.ai) bundle list-building, enrichment, personalization, and email send into a single SaaS product. Pricing typically runs $250-$2,500/month depending on tier and contact volume. The same workflow can be built as a custom outreach agent on Postmark + Hunter + a small LLM for $74-$172/month, with break-even on engineering hours typically inside year one.
See alsoSoftware that takes over manual daily work that no SaaS solves cleanly: outreach personalization, content engagement, reporting, internal handoffs, candidate routing.
Ops automation is Lane 01, the first of the two lanes NodeSparks works in. It targets recurring manual work that costs ~10-15 hours/week per affected person and which existing SaaS tools either do not address or address with too much friction. Typical pattern: Slack-native agent + LLM + 2-3 API integrations, deployed on a small VPS.
See also/glossary#custom-build-cost The total cost of replacing a SaaS subscription with a custom system. Two components: a one-time build billed against an agreed maximum and an ongoing infrastructure plus API cost.
A custom build has two costs. The build itself is billed against a maximum agreed in writing before any code exists, at €3,000 for one job and €6,500 for one system. Running it is $5 to $30 a month of infrastructure plus the API calls. Nothing is paid before work starts and there is no per-seat scaling to renew. Break-even on a €6,500 system lands around month 16 at €500 a month once the hosting and the API bill are netted off. Above that it comes sooner.
See also/glossary#eu-saas-sovereignty The practice of running an SME tech stack on EU-hosted infrastructure rather than routing data through US-hosted SaaS — driven by GDPR risk, latency, and pricing.
EU SaaS sovereignty is more achievable in 2026 than it was in 2024. Hetzner-hosted n8n, Postmark EU region, OpenAI EU Residency, Anthropic EU endpoints, and EU-hosted Postgres providers (Neon, Supabase) cover most SME workflows. Trade-off: slightly fewer integrations available out of the box, slightly higher engineering cost to wire things together. Benefit: simpler GDPR posture, lower latency for EU users, often lower price.
See alsoReviewed September 2026. Free to cite and adapt under CC BY 4.0. Missing a term, or think one is wrong? Tell us.
