The short answer
Accounting firms typically need custom software when client workflows do not fit QuickBooks Online, Drake, or Lacerte — or when they add a niche service like R&D credits, franchise bookkeeping, or multi-entity consolidation. Until then, keep buying: a small custom build shouldn't redo what a ledger or a professional tax package already does.
When a build does make sense, it usually leaves the ledger and the tax package alone and fills the gaps between them: intake, engagement status, and reporting across systems. A focused tool like that typically lands between $5,000 and $15,000 and takes 2–4 weeks. And because Pythn Development is a DBA of Finlock Accounting, a build that needs its own accounting layer plugs into the Finlock API instead of rebuilding one. Our accounting firms page lists what gets built; this post is about whether to build at all.
Three pain points, and when each is worth fixing
Most of the problems worth solving in a firm aren't inside any single tool. They're in the gaps between tools: the double entry, the status questions, the “where did that document go?” moments. Each can be a software problem or a process problem, and it's worth knowing which before anyone writes code.
1. Data entry
The same client details get typed into the portal, the practice-management tool, and the ledger. It's also easy to measure: count how often a value gets copied between screens in a normal week. If it's a handful of copies, it isn't worth a build. If it happens for every client, every period, it's a sync job, and keeping systems in sync is what business automation and integration work is for.
2. Document collection
Chasing clients for paperwork is a process problem first, and a portal with a consistent request list solves a lot of it. Custom work earns its place when the request list depends on the engagement, so someone rebuilds it by hand for each client, or when documents arrive in one system while their status is tracked in another.
3. Reporting
Partners want answers across the whole client list: which engagements are stuck, what's waiting on the client, where the hours went. When those answers come from three exports pasted into a spreadsheet every week, there's a reporting layer waiting to be built — one that reads from the systems you already use instead of adding another place to type.
One test covers all three. A problem that shows up once a year is rarely worth a build. A problem that repeats for every client, every month, and grows each time the client list does is worth pricing. That growth is the real tipping point, and it arrives at a different client count for every firm.
Build vs buy for accounting firms
The rule from our buy-vs-build framework holds here: custom software makes sense when the monthly cost of workarounds exceeds the one-time cost of building. What's specific to accounting firms is which parts of the stack should stay bought.
- Keep the specialist software. QuickBooks Online for the books and professional tax software such as Drake or Lacerte are not things a small firm should pay to rebuild.
- Build the connective layer. Intake that feeds the tools you keep, status tracking across them, and reports that combine them. This is where the double entry lives.
- Build for the niche service. If you deliver a niche service in a way no product models, the workaround behind that service is the first thing to price.
Before pricing any of it, check what each product exposes: a documented API, an import and export file, or nothing at all. That answer moves the price, because each integration adds roughly $2,000–$5,000 depending on how well the other system is documented. Our guide to connecting QuickBooks, Stripe, and your CRM covers what that work involves.
Put maintenance in the math too: it typically runs 15–20% of the build cost per year, and our guide to maintenance costs after launch has the detail. The build side, at least, is a known number: Pythn fixes scope and price in a written design document before any code is written.
Where the Finlock API fits
Pythn Development is a DBA of Finlock Accounting, our sister company. For a firm weighing a build, that matters in one specific place: the accounting layer.
A build that only moves data between the tools you already have doesn't need books of its own; it works with the ledger you use. Some builds do need an accounting layer: ledgers, invoicing, payroll, reporting, or tax. Rather than rebuilding that wheel, a Pythn build plugs those pieces directly into the Finlock API, so the software has production-grade accounting from day one and the custom work stays on the workflow that's actually yours.
The practical effect is one source of truth for the financial data, instead of a custom ledger that has to be reconciled against everything else. The integration itself — the API layer and the background jobs that keep data in sync — is API and backend work, shipped with API docs, integration tests, and a deployment runbook. There's more on the relationship on our about page.
Three patterns: solo, five-person, and multi-location
These are illustrative patterns, not client stories. Real firms mix them; start from whichever is closest to yours.
The solo practitioner
Take a solo practitioner with a steady list of individual and small-business clients, running QuickBooks Online, a tax package, a portal, and e-signature, where only the owner needs to know where everything stands. The honest answer here is usually not to build: the workarounds cost one person's time, and better use of the tools already paid for is the cheaper fix. If a build happens at all, it's small — an intake form with some logic and email routing, in the $1,000–$5,000 band.
The five-person firm
Take a five-person firm: a partner, two preparers, a bookkeeper, and an admin. Now the pain is coordination — who has which client, what's missing, what's waiting on the client, what's ready for review. That status lives in a spreadsheet and in people's heads, and busy season strains both. This is the shape a focused internal tool fits: intake that builds each client's request list, assignment, and a status board across the tools the firm keeps. It typically lands in the $5,000–$15,000 band and takes 2–4 weeks.
The multi-location franchise bookkeeper
Take a bookkeeping practice that serves many locations of one franchise brand, each set up as its own company. The work is repetitive by design: the same setup, the same checklists, and the same reports for every location. The pain is scale, because each new location adds another copy of the manual steps, and the cross-location view is assembled by hand. This is where a full platform makes sense: every location's status in one place, exceptions flagged instead of hunted for, and reports generated rather than pasted together. A full platform typically sits in the $15,000–$50,000 band and takes 2–6 months, and if it needs its own accounting layer, that's where the Finlock API comes in.
Frequently asked questions
Should a small accounting firm build custom software?
Only when a specific, recurring problem costs more to work around than to fix. The signals that a build is worth pricing: the same data typed into two systems for every client, engagement status that lives in one person's spreadsheet, or a niche service no product fits. If none apply, get more out of the tools you already pay for. Our custom vs SaaS quiz is a quick first check.
What is the Finlock API?
Pythn Development is a DBA of Finlock Accounting, our sister company. The Finlock API is what Pythn builds plug into when a project needs an accounting layer — ledgers, invoicing, payroll, reporting, or tax. You get production-grade accounting from day one without Pythn rebuilding the wheel.
Will custom software replace QuickBooks Online, Drake, or Lacerte?
Usually not. When Pythn builds for firms, the ledger and the tax package typically stay, and the new software handles intake, status, and reporting around them. How tightly it can connect to each product depends on what that product exposes, which is worth checking before anything is scoped.
How much does custom software cost for an accounting firm?
Focused tools, such as a branded client portal or an intake workflow, typically land between $5,000 and $15,000 and take 2–4 weeks. Full practice platforms usually fall between $15,000 and $50,000, depending on integration count and existing systems, and take 2–6 months. The cost calculator gives a band in five questions; the cost guide explains what drives it.
Does Pythn work with firms outside the Finger Lakes?
Yes. Pythn is based in Geneva, NY, and works on-site with accounting firms in the Finger Lakes, the Rochester metro, and Western New York, and remotely with firms across the United States.