Tax Preparation Software: Build vs Buy

What Drake, UltraTax CS, Canopy and TaxDome actually do well

Nearly every firm reading this should buy, and the part you might be tempted to build is the part you definitely should not.

Drake Tax, Lacerte, ProSeries, UltraTax CS and CCH Axcess Tax carry the calculation and the filing. Forms change every year, states diverge, electronic filing schemas move, and the vendors absorb all of it. Rebuilding a tax engine is not a project, it is a permanent obligation, and no firm has ever regretted keeping one. On the practice side, Canopy, TaxDome, Karbon, Jetpack Workflow and XCM give you tasks, due dates and client records that work out of the box. SmartVault, TaxCaddy, Liscio and ShareFile solved document shipping. SafeSend Returns handles delivery and signature collection properly, including the identity verification a remote Form 8879 requires.

Those tools also carry security groundwork you would otherwise write: encryption at rest, access logging, multi factor authentication and the audit trail your written information security plan has to describe.

So buy if you are a single office under roughly 1,200 returns with one dominant service line. TaxDome or Canopy plus Drake will cover you, and a custom platform will cost more than the coordination it saves. Buy also if your partners are three years from selling, because an acquirer will migrate you to their stack regardless. We say this to firms who called us to build, and we mean it.

Where they stop: nobody can answer where is this return

It is 22 March. Your firm administrator has a client on hold asking whether her 1040 is done. The answer needs three logins: the tax engine to see whether the return is prepared, the portal to see whether the brokerage 1099 ever arrived, and the practice tool to see whether the reviewer signed off. None of them agree. The whiteboard behind the front desk says waiting on K-1, which was true on 4 March, and the preparer who wrote it is in another office and has moved on to eleven other files.

The reason no product fixes this is that none of them is allowed to be the authority. Your tax engine has a preparer entered status that means whatever the last person to touch it decided. The practice tool derives status from task completion, which lags because preparers close tasks in batches on Friday. Neither knows the client emailed a corrected 1099-B on Tuesday. Integrations between them sync client names and due dates, not the state machine, so the state machine ends up in a human head.

Two other gaps follow from the same root. Document intake solved shipping but not sorting: a client uploads 43 pages as one scan called taxes, someone splits and files it, and three days later the preparer discovers the K-1 from the client's LLC is missing. Extraction tools hand you data and stop, because they do not know what the client had last year. And delivery is where a well prepared return dies: the 8879 goes out on 2 April, nobody signs, and nobody notices until 12 April because signature status lives in one system and the deadline lives in someone's head. Meanwhile a return transmitted on 28 March came back rejected from Modernized e-File over a dependent identification mismatch, and the acknowledgment landed in a shared inbox three people half watch.

The arithmetic: per return and per seat fees against a build

Your stack has three meters. Practice management and portals are priced per user per year. Delivery platforms are priced per return. Tax engines are priced per return, per module or as an unlimited bundle. Write all three down before comparing anything, because only the per return meters grow with the season you are trying to expand.

Run your own season. Suppose you file 2,400 returns with 30 staff. Practice management at $900 per user is $27,000, delivery at $6 per return is $14,400, and a portal adds several thousand more, so call the coordination stack $47,000 a year before the tax engine.

Now price the coordination itself. If each return gets its status looked up eight times across intake, preparation, review, signature chasing and delivery, and each lookup burns three minutes hunting across systems, that is 24 minutes per return. On 2,400 returns it is roughly 960 hours, and at a $95 blended loaded cost it is about $91,000 of your busiest people doing nothing that touches a tax position. That figure excludes returns that sit for nine days because everyone assumed someone else owned them.

The crossover sits near 1,500 returns a season across two or more offices, or roughly 25 staff seats. Below 1,200 returns in one office, buy. Between 1,200 and 1,500, buy and fix your intake process, which is cheaper and recovers more hours. Above 1,500 across offices, a $130,000 first release plus $26,000 migration plus 18 percent annually clears its cost in the second season.

What a custom coordination layer actually costs

Across more than 2,000 delivered projects, Digital Heroes sees this category land in two bands. A focused first release covering the return ledger, document intake with classification and a missing items engine, the signature and acknowledgment loop, and one live dashboard runs $60,000 to $130,000 and ships in 12 to 16 weeks, deliberately timed to land before a season rather than during one. A full platform adding capacity forecasting, billing and realisation, a client portal, multi year rollover and multi office reporting runs $150,000 to $400,000 phased over 6 to 12 months.

Two lines belong in the budget. Data migration runs 10 to 25 percent of build cost, and it sits at the top of that band here because five to ten years of prior year documents have to load for the year over year comparison to work on day one. That item is routinely underestimated and deserves its own line. Year two onward runs 15 to 20 percent of build cost annually, which in a seasonal business buys you the pre season hardening as well as support.

What pushes price up in tax specifically: the number of tax engines you read from, since a firm running both Lacerte and CCH Axcess after an acquisition roughly doubles the integration surface. State electronic filing coverage. Hosted environments such as Rightworks or Citrix that constrain how you reach the data. And the security work the FTC Safeguards Rule and IRS Publication 4557 require, including your written information security plan, encryption, access logging and multi factor authentication.

The four situations where building wins

Regulatory fit. The FTC Safeguards Rule and IRS Publication 4557 set expectations your written information security plan has to describe and your systems have to honour: access logging on every view of taxpayer data, retention policy, and identity verification on remote Form 8879 signatures. Section 7216 consent tracking becomes a hard requirement the moment you want to use return data for anything beyond preparing the return, including analytics your partners will eventually ask for. A build lets you record consent per client per purpose rather than filing a signed page and hoping.

Scale economics. Past roughly 1,500 returns across two offices, coordination cost exceeds the entire licence stack and keeps climbing while return count stays flat.

A workflow that is your competitive advantage. A missing items engine that compares this year's document set against last year's filed return, knows by entity name that two of three K-1s have arrived, and chases at 7pm by text with a one tap upload link, is where firms recover the most hours. The chase is the work, and no shared vendor tool will give you a version your competitors do not also have.

Integration sprawl across three or more systems. Tax engine, portal, practice tool, delivery platform, extraction service, billing, and an inbox holding every decision anyone actually made.

How to decide in a week, in the middle of March

Do this during season, not in September, because September will flatter you.

Monday: pick 20 returns at random and time how long it takes to establish the true status of each, counting logins. Tuesday: take every return sitting in the same stage for more than seven days and find out who believed they owned it. Wednesday: count outstanding 8879s and sort them by deadline exposure, then check how many nobody had chased in the previous 72 hours. Thursday: pull every electronic filing rejection from the last month and identify who owned each and how long it sat. Friday: ask your administrator to hand over the spreadsheet, and count how many fields in it exist in no system you pay for.

If Friday's count runs past six fields and Monday averaged more than four minutes a return, the conditions are met. If the practice tool answered everything, keep it and spend the money on staff, which is almost always the better purchase.

When you do cross the line, fund a paid discovery phase first, and insist on a delivery plan built around your calendar rather than a vendor's. Anyone proposing a February go live has never watched a firm in season. Digital Heroes writes a signed product requirements document covering the return ledger, the intake model, consent tracking and acceptance criteria before code exists, and you keep that document whichever firm builds. We contract through India LLP, US LLC and UK LTD entities so intellectual property assigns under law your own advisers already read.

We are the wrong firm if you want a tax engine written, or if your administrator will not be in the room during discovery. Software built without that person becomes shelfware by February.

Book a 30-minute call with Digital Heroes and get a written plan and a fixed quote within 48 hours.