QuickBooks Online works fine for one business. The trouble starts when bookkeepers try to run portfolios of clients on it. Past 15 clients, the math stops working. Software costs scale linearly. Hours per week climb. Rule maintenance becomes a part-time job. This article sits inside the bookkeeping automation guide at /topics/bookkeeping-automation and answers one question: when does it make sense to replace QBO at the portfolio level?
QBO is single-tenant by design. Each client gets a separate file, a separate login, and a separate set of bank rules. Nothing transfers. AI training memory across clients? It doesn't exist. Batch operations across the portfolio? Per file only. A single dashboard for all your clients? You bounce between tabs.
This is a scale-mismatch, not a product-quality problem. QBO is built for one company. Bookkeepers run 15, 25, 40 clients. Different job, different tool. Below we cover five specific breakpoints, an honest "when QBO is still right" section, the 15-client time and money math, a comparison of the four serious alternatives in 2026, and a quick migration overview.
When does QuickBooks stop scaling for bookkeepers?
QBO scales fine up to about 15 clients. Past that, the math breaks. Each new client adds a $115/mo Plus subscription, roughly an hour per week of categorization work, and a fresh set of bank rules that don't transfer from your other clients. At 15 clients you're paying ~$1,725/mo in QBO subscriptions, spending 15+ hours per week on categorization, and rebuilding the same rules over and over. The five breakpoints: no portfolio-level AI memory, no portfolio dashboard, no batch operations, bank rules that stay per-client, and per-client billing that compresses margin. Replace QBO at the portfolio level when you cross 15 clients and integrations aren't blocking you. Layer Growthy over QBO when integrations are sticky.
QBO was designed for the small business owner managing their own books. One company file, one chart of accounts, one set of bank rules. The product is excellent at that job. Intuit supports thousands of US financial institutions, pours billions into the platform each year, and just announced a 15-25% price hike on May 1, 2026 because demand is strong [Intuit, 2026 pricing announcement].
The problem isn't the product. It's the architecture. Single-tenant means each new client opens a separate file. There's no shared layer that learns across your portfolio. Every bank rule you build for client A stays trapped in client A's file. Every login is separate. Every dashboard is per-client. For one business owner doing their own books, this is fine. For a bookkeeper running 15 engagements, it's a tax on every hour of your week.
A portfolio is not 15 separate single-business engagements stacked on top of each other. It's a single book of business with shared patterns. Your 15 clients use overlapping vendors. Stripe, Gusto, Adobe, AWS, and Google Workspace probably appear in 10 of them. The same rule for "Adobe = Software Subscription" should apply across all 15. In QBO, you build it 15 times.
A portfolio also needs a single dashboard. You want to answer: which 3 clients haven't closed April yet? Which 2 have unreviewed transactions over $50? In QBO, you log into 15 separate files to find out.
The portfolio model needs portfolio-aware tooling. QBO doesn't ship that, because that's not what QBO was built to do.
Bookkeepers consistently hit a wall around 12-15 clients on a QBO plus manual workflow. The numbers tell why. 15 clients times roughly an hour per week of categorization equals 15 hours of coding work before reconciliations or close. Add 15 QBO Plus subscriptions at $115/mo and you're paying $1,725/mo in software cost before bookkeeper time. Margin compresses. Burnout follows. Most solo bookkeepers stop growing right here, at the exact point where they should be scaling into a real practice.
QBO's bank rules learn nothing from your other 14 clients. If you spend an hour teaching client A's file that "Stripe Transfer" maps to "Sales Income," that rule does not transfer to client B. You start over.
Growthy AI categorization works on patterns across your full portfolio. It learns vendor patterns, amount ranges, memo strings, and counterparty behavior across every client at once. A new client added to the portfolio inherits 85% first-import accuracy because the model has already seen Stripe and Gusto and Adobe in 14 other files.
That difference compounds. Every new client on QBO is a fresh setup project. Every new client on a portfolio model lands closer to "ready to review" on day one.
QBO has no concept of "your book of business." You have 15 separate files, 15 separate logins (or one login that switches contexts), and 15 separate dashboards. Want to see which clients haven't closed for April? Open each file. Want to see total unreconciled transactions across the portfolio? Open each file. Want to see who has uncategorized transactions over $500? Same.
A portfolio dashboard answers those questions in one view. Close status by client, exception count by client, transactions awaiting review across all clients. That's a different product category. QBO doesn't sit in it.
When QBO releases a new feature or you change a process, you propagate the change file by file. Lock April for all 15 clients? 15 separate operations. Update a vendor name? 15 times. Add a custom GL account? 15 times.
Batch ops at the portfolio level are how solo bookkeepers handle 15 clients without burning out. Lock the period for the whole portfolio in one click. Push a new vendor mapping across every client. Roll forward year-end balances at the portfolio level. QBO has none of that, because QBO doesn't see your portfolio. It sees 15 unrelated companies.
QBO bank rules are powerful inside one file and useless across files. The rule "Vendor contains 'GUSTO' then category 'Payroll Expenses'" works perfectly for the client you built it for. It does nothing for the other 14 clients on Gusto.
You can copy rules client by client. Bookkeepers do this. It's manual, slow, and breaks every time Gusto changes the bank descriptor. Then you rebuild 15 times. The structural problem isn't the rules engine. It's that QBO has no portfolio layer where shared patterns live.
QBO Plus is $115/mo per company file as of May 2026 [Intuit pricing page]. At 15 clients, that's $1,725/mo in subscriptions before you add Hubdoc, payroll, or your practice management tool. At 25 clients, $2,875/mo. The marginal cost per client doesn't drop. Linear scaling.
Growthy alpha is $99/mo per seat with 5 companies per seat. 15 clients = 3 seats = $297/mo. 25 clients = 5 seats = $495/mo. The marginal cost per client drops as you scale. That's the whole point of a portfolio pricing model.
If you're at 5-10 clients with no plans to grow past 15, QBO economics still work. $575-$1,150/mo in software is fine when you're billing each client $300-500/mo. Don't migrate just because there's a newer tool. Migrate when QBO becomes the bottleneck on your growth, not before.
If your clients are stable, your processes are dialed in, and you have no roadmap to add 5+ clients in the next year, leave QBO alone. The migration cost (20-40 hours of setup time) won't pay back inside a year.
QBO has the broadest 3rd-party ecosystem in small business accounting. Thousands of supported banks, native Gusto integration, Square POS sync, Shopify, BigCommerce, Stripe, and dozens more direct integrations [Intuit integration directory].
If your clients depend on a specific QBO integration that doesn't exist elsewhere yet, switching costs more than it saves. A retail client running QBO POS sync. An e-commerce client with a custom QBO Shopify integration. A construction client using a job-costing add-on that only writes to QBO. In those cases, layering Growthy over QBO (Mode B) makes more sense than ripping out QBO entirely.
If your client's CPA pulls tax data directly from QBO via UltraTax, Drake, ProConnect, or Lacerte, switching to a new GL adds friction at tax season. The CPA has to learn a new export format. Your client pays in either money or risk.
Some CPAs accept CSV exports from any GL. Others insist on direct QBO pull. Ask before you migrate. If the CPA insists on QBO, layer Growthy over QBO with bank rules vs AI categorization running side by side. You get portfolio AI without breaking the CPA's workflow.
At 5 clients on QBO + manual workflow, expect ~5 hours per week on bookkeeping. Categorization takes the bulk; reconciliation and close are spread across the month. Margin healthy at standard $75-150/hr bookkeeper rates.
At 15 clients on QBO, the same workflow scales to roughly 15 hours per week on categorization alone. Add reconciliations, close work, and admin and you're at 25-30 billable hours per week. That fits in a part-time week if you don't sleep, exercise, or take new business calls.
At 25 clients on QBO without tooling, the math doesn't work. 25 hours per week on categorization plus another 15 on close means a 40-hour bookkeeping week before you do anything else. Burnout territory.
5 clients on QBO Plus = $575/mo software. 15 clients = $1,725/mo. 25 clients = $2,875/mo. QBO ships native receipt capture (QBO Receipts) bundled with the subscription; Hubdoc bundles with Xero, not QBO. Dext or AutoEntry adds $20-30/mo per OCR seat for portfolios needing more capable OCR. Practice management on top: Karbon at $59/mo or TaxDome at $60/mo per user.
Total stack at 15 clients on QBO: roughly $1,800-1,900/mo in software. At 25 clients: roughly $2,950-3,050/mo. That's before bookkeeper time.
At 15 clients billing $400/mo each, gross revenue is $6,000/mo. Software at $1,800/mo eats 30% off the top. Bookkeeper time at 25 hours/wk × 4 weeks × $75/hr = $7,500/mo opportunity cost. The math goes negative on a value-of-time basis. Bookkeepers stop here because the next client adds $400/mo in revenue and $1,000/mo in stress.
Growthy is built for bookkeepers running portfolios. Per-seat pricing at $99/mo alpha (5 companies per seat), AI categorization that learns across your portfolio, and a portfolio dashboard that shows close status by client. 85% first-import accuracy, 90%+ on returning clients. Strong on the categorize and match layers (layers 2 and 3 of the 4-layer automation framework). Honest gap: native receipt OCR is on the roadmap, not shipped today. Hubdoc, Dext, or AutoEntry handle OCR upstream.
Best for: bookkeepers at 15+ clients who want portfolio-aware tooling and per-seat pricing.
Xero is the dominant non-QBO GL globally, especially outside the US. Strong product, clean API, broad bank coverage. Architecturally similar to QBO: single-tenant per file. Same portfolio limits. Different country footprint and a stronger position with international bookkeepers.
Pricing in the US starts around $20/mo and runs to $80/mo per company depending on the tier. Marginal cost per client still scales linearly. If you switch from QBO to Xero, you get a different product but the same single-tenant constraint.
Best for: bookkeepers serving international clients or those moving away from QBO for product preference rather than portfolio scale.
Puzzle is engineering-led, designed for venture-backed startups and founders who want a clean modern GL. Strong dashboards, good UX, accrual-aware. Less mature multi-client tooling. The product targets the founder buying directly more than the bookkeeper running portfolios.
If you serve venture-backed startups and your founders push you toward Puzzle, it works. For solo bookkeepers running varied small business portfolios (restaurants, retail, professional services), the fit is weaker.
FreshBooks started as invoicing for freelancers and grew into a light GL. Strong on invoicing, time tracking, and project profitability for service businesses. Weaker on accrual accounting, inventory, and complex GL work. Per-client pricing.
Best for: bookkeepers serving freelancers, agencies, and service businesses where invoicing is the dominant workflow. Not a fit for portfolios with significant accrual or inventory work.
QBO
Growthy
Xero
Puzzle
FreshBooks
Every new client adds a full subscription. 15 clients on QBO Plus = 15 subscriptions = $1,725/mo. The marginal cost per client never drops. You pay full sticker for client 1 and full sticker for client 15.
This pricing model assumes the buyer is the business owner, not the bookkeeper. It works fine for the small business owner buying QBO for their one company. It punishes the bookkeeper running 15 engagements.
Growthy alpha is $99/mo per seat with 5 companies per seat. 15 clients = 3 seats = $297/mo. The marginal cost per client drops as you fill each seat. Adding the 6th client to a 5-client seat triggers a new seat. Adding the 7th-10th clients fills it.
This pricing model assumes the buyer is the bookkeeper running a portfolio. It scales sub-linearly with client count.
Below 5 clients, QBO Plus is cheaper sticker (one subscription vs one seat). At 5 clients, the math is roughly even. Above 5 clients, the per-seat model wins.
At 15 clients: QBO ~$1,725/mo, Growthy ~$297/mo. At 25 clients: QBO ~$2,875/mo, Growthy ~$495/mo. The gap widens as you scale.
Migrate (Mode A) when QBO itself is the bottleneck. Per-client billing is killing margin, your portfolio dashboard pain is acute, and your clients don't depend on a specific QBO-only integration.
Layer (Mode B) when integrations are sticky. Your clients have QBO Shopify, QBO POS, or QBO-only payroll add-ons that don't exist elsewhere. Layer Growthy on top of QBO for the AI categorization and portfolio dashboard. Keep QBO as the GL of record.
Most bookkeepers start in Mode B and migrate clients to Mode A one at a time as integrations get unblocked.
Run a 30-day migration plan from QBO with parallel run and decision gates. Week 1: COA mapping and AI training on historical data. Week 2: parallel run on one client. Week 3: parallel run on 3-5 clients. Week 4: cutover for clients where parallel run matched, defer clients where it didn't.
The parallel run is the safety check. If Growthy's categorizations match what you'd code in QBO, you're ready. If they diverge, you fix the AI training or defer that client to Mode B.
Keep QBO active for 60 days post-cutover. If a client's tax preparer needs a year-end QBO export, you can pull it. If you discover a categorization issue, you can compare against the QBO baseline. After 60 days with no issues, downgrade or cancel the QBO subscription.
The full bookkeeper automation stack with practice management, OCR, and reporting tools sits in the bookkeeper automation stack.
Yes, in most cases. Growthy exports a clean trial balance, GL detail, and journal entries in CSV format. Most CPAs accept CSV. If your CPA insists on direct QBO pull (some firms require it for UltraTax, Drake, ProConnect, or Lacerte direct integration), layer Mode B instead. Keep QBO as the GL of record and Growthy as your portfolio workflow layer.
Ask the CPA before you migrate. The answer changes by firm.
Yes. The recommended pattern is to migrate one client first, run parallel for 30 days, then expand. Pick a client where you have full chart of accounts control, low transaction volume (under 200/month), and a flexible tax preparer.
After the first migration, you'll know the exact time cost (typically 4-8 hours per client for COA mapping and AI training). Plan accordingly for the rest.
Yes. Growthy imports your QBO chart of accounts, vendor list, customer list, and historical transactions via QBO API. Historical data is what trains the AI categorization model. The more history Growthy ingests, the higher first-import accuracy on subsequent clients.
Plan for 2-4 hours of import + review per client. Not zero-touch but not painful.
Bank rules don't migrate directly because Growthy uses AI categorization rather than rule-based matching. The good news: AI categorization replaces most bank rules. The 5-10 high-confidence repeat vendors (Stripe, Gusto, recurring SaaS) where rules genuinely win can stay in QBO if you're running Mode B, or be replicated as confidence-bypass overrides in Growthy if you're in Mode A. See the breakdown of bank rules vs AI categorization for when each approach wins.
Plan for 6-10 weeks for a clean 15-client migration. Week 1-2: setup and AI training. Week 3-4: parallel run on first 3 clients. Week 5-6: parallel run on next 6 clients. Week 7-8: parallel run on final 6 clients. Week 9-10: cutover and cleanup.
You can do it faster if you're full-time on it. Most bookkeepers spread it over 8-10 weeks while serving existing clients.
Growthy is in alpha. SOC 2 Type II is on the roadmap. Today, security posture includes encryption at rest, encryption in transit, role-based access controls, and audit logging. If your clients require SOC 2 Type II for data handling (common in regulated industries), layer Mode B over QBO until Growthy completes SOC 2.
Ask about current security posture and roadmap timing during your alpha onboarding.
Growthy is bookkeeping software, not a CPA firm. This content is educational, not professional advice. Full disclaimer.
Related: Bookkeeping Automation for Multi-Client Bookkeepers, Multi-Client AI Bookkeeping, Migrate Manual to AI: 30-Day Plan, Bookkeeper Automation Stack
Free during alpha. Read-only access. You review every sync.
CPA firm partner who got tired of watching bookkeepers click categorize 500 times a day. Built Growthy to fix it.
View all articles →Growthy is dedicated to helping businesses of all sizes make informed decisions. We adhere to strict editorial guidelines to ensure that our content meets and maintains our high standards.
Bank rules vs AI categorization across 8 dimensions. Hybrid stack wins: rules for high-confidence repeat vendors (Stripe, payroll, recurring SaaS), AI for everything else. Real numbers.
The 4-layer framework for evaluating bookkeeping automation (import, categorize, match, report) plus a 7-tool comparison: Growthy, Digits, Puzzle, Hubdoc, Pilot, Docyt, QBO Live.
