Scratch Development vs. Package Software vs. SaaS: How to Choose
A neutral comparison of scratch development, package software, and SaaS across cost, timeline, and customizability, with criteria for choosing based on how unique your operations are.
What Are Scratch Development, Package Software, and SaaS?
System development can broadly be procured in three ways: scratch (custom) development, package software, and SaaS (cloud services). Each differs significantly in cost structure, degree of customization, and where maintenance responsibility sits. Choosing without weighing these against your own operational requirements often leads to overinvestment or, conversely, functional gaps that disrupt operations. This article compares the three approaches neutrally and outlines criteria for choosing among them. For the full ordering process, see the System Development Ordering Guide.
Background: Why the Choice of Approach Has Become a Management Issue
As small and midsize businesses push forward with digitalization, development budgets remain limited even as the need to digitize operations keeps expanding. The spread of low-cost SaaS and no-code tools has widened the range of options, but that abundance of choice has itself become a new source of difficulty. It is now common to start with SaaS and fill gaps with scratch development where needed, rather than committing to a single approach outright — more companies are planning for a combination from the start.
Defining the Three Approaches
- Scratch development: Building a system entirely custom to your business, from requirements definition through design and implementation, commissioned to a development company
- Package software: Deploying pre-built business software and adding customization (add-on development) as needed — common in accounting, sales management, and production management
- SaaS (cloud service): Using a ready-made cloud service delivered over the internet on a monthly or annual subscription — kintone, Salesforce, and freee are representative examples
The Structure of the Problem: Why the Wrong Choice Leads to Failure
Selecting an approach without understanding these differences tends to produce predictable failures: starting with SaaS only to find the operations too unique for the tool to handle; over-engineering a scratch build until costs blow past budget; or adopting a package whose accumulated customization fees end up matching a scratch build's cost anyway. The root cause is usually deciding based on price or brand recognition alone, without first assessing how standard the operation actually is and how much it is likely to change going forward.
A Five-Axis Comparison
| Axis | Scratch Development | Package | SaaS |
|---|---|---|---|
| Initial cost | High (millions of yen, scales with scope) | Moderate (hundreds of thousands to a few million yen) | Low (from zero to a few tens of thousands of yen) |
| Time to deploy | Long (several months to over a year) | Moderate (1–3 months) | Short (same day to a few weeks) |
| Customizability | High (can match operations exactly) | Moderate (within add-on scope) | Low to moderate (within settings/integration scope) |
| Maintenance | Depends on in-house team or vendor (maintenance contract needed) | Vendor maintenance plus in-house handling | Bundled by the vendor (low in-house burden) |
| Asset value | High (remains a company asset) | Moderate (depends on license terms) | Low (usage right only; disappears on cancellation) |
*Costs vary widely by industry, scale, and requirements — always confirm with quotes from multiple vendors.*
The table shows a simple trade-off: SaaS wins on cost and speed, while scratch development wins on customizability and asset value when operations are highly unique. Packages sit in between, suited to companies whose processes align reasonably well with a standard workflow.
Which Approach Fits Your Operations: Judging by Uniqueness × Scale
In practice, the decision comes down to two axes: how unique your operations are, and your business scale and investment capacity.
- Standard operations (accounting, attendance, business card management) at small scale: SaaS is the first candidate — aligning with an industry-standard workflow keeps both cost and operational load low
- Standard operations but larger scale, multiple locations, or complex approval flows: Package plus add-on customization is often the practical choice
- Highly unique operations (proprietary production management or order logic): Scratch development becomes a candidate, but the return on investment needs careful evaluation
- Highly unique but limited investment capacity: Consider the hybrid approach described below (SaaS plus connective development)
The Practical Middle Ground: Hybrid (SaaS Plus Connective Development)
A growing pattern is the hybrid approach: use an existing SaaS or package for the core, and cover only the company's unique business logic with small-scale scratch development such as API integration or add-ons. This keeps initial costs lower than building everything from scratch, while still benefiting from automatic updates to standard features, and concentrates the development budget on the parts that genuinely differentiate the business. See The Limits of No-Code Development for where no-code tools reach their ceiling, and the System Development Cost Guide for cost benchmarks.
A Practical Selection Process
- Inventory your operations: Separate current workflows into non-negotiable requirements and requirements you can compromise on
- Check existing products first: Confirm what percentage of your requirements a SaaS or package can already meet
- Estimate the ceiling on customization costs: If package add-on fees keep accumulating, calculate the break-even point against scratch development
- Get quotes from multiple vendors: Present the same requirements to several development companies or vendors and compare cost and proposal content
- Check the risks around maintenance and cancellation: SaaS raises data-migration concerns at cancellation; scratch development raises concerns about the ongoing relationship with the development company
Frequently Asked Questions
Can we switch from SaaS to scratch development later?
Technically yes, but data migration and redesigning workflows takes both cost and time. Planning for a hybrid approach from the start, and confirming upfront whether the SaaS can export data via API, makes a later migration much easier.
What's a typical cost range for package customization?
It varies too widely by scope and industry to state a single figure. Minor adjustments close to standard features cost far less than a large-scale overhaul of the underlying business logic. Ask vendors to break down customization items in detail when quoting.
Should small companies ever consider scratch development?
If your business process itself is a source of competitive advantage — a proprietary order-handling logic, for instance — scratch development can be worth considering even at small scale. But this should rest on a projected return on investment, validated against quotes from multiple vendors.
Summary
There is no inherently superior choice among scratch development, package software, and SaaS — the right answer depends on how unique your operations are and how much you can invest. Start by inventorying your operations, check how far existing products can take you, and reserve custom development for the gaps. That approach tends to produce the best cost-effectiveness. Always confirm final cost and timeline with quotes from multiple vendors before deciding.
Feel free to contact us
Contact Us