Square + Xero deep dive

Executive summary

Essential Care's revenue-in, revenue-accounted-for, and revenue-collected pipelines each break in distinct ways. On the invoicing side, a single staff member (Mohamed) is manually creating every invoice through Square, producing roughly 1,334 invoices per month, and has driven median invoicing latency from 60 days in September 2025 to 14 days in April 2026 — a genuine operational turnaround. Over the same window, the records show $121,787 in client co-contributions fell past 91 days overdue, concentrated in the December 2025 – February 2026 cohort that was invoiced late and never followed up systematically.

On the accounting side, Xero shows a $2.34M YTD net loss (10 months to 30 April 2026) versus a positive prior-year comparable, with the balance sheet carrying $252,496 in unremitted PAYG, an outstanding Square clearing account balance, and a director-vehicle asset that warrants documentation review.

The invoicing and accounting findings are the same story seen from two ends of the same pipe. The Amaka integration was built for retail point-of-sale: it only pushes paid Square transactions into Xero as daily summaries. Any unpaid Square invoice is invisible to Xero. Any bank transfer paying a Square invoice that staff have marked “paid” in Square arrives in a Xero clearing account and stays there. The architecture is wrong for Essential Care's use case. The fix is a Xero-first invoicing flow with BPAY, covered in Section 7.

Five numbers that anchor the rest of this document

Square invoicing — operational deep dive

3.1 — Production curve

Mohamed became the sole invoice creator during FY25-26. His monthly output shows a clear ramp as he absorbed the historical backlog.

The January dip reflects the summer holiday period. February is the peak at ~104/day — a large backlog consolidation push. The trajectory proves operational capacity; it also exposes single-point-of-failure risk. If Mohamed takes leave, output falls to zero and the overdue pile grows proportionally.

3.2 — Latency: the improvement and the hangover

Latency = (Invoice Date − Service Date). The whole-period average is 69.0 days, median 70, P90 119. The raw headline is misleading because it includes the pre-catch-up period. The monthly trend tells the real story.

The team is now near-current on new work — a 14-day median sits well inside sector norms. However, the December 2025 cohort, issued at a 94-day median lag, was delivered to clients who had long since forgotten the underlying service, had no payment prompt in context, and in many cases had already paid by some other route that was never matched. That cohort is where the $121K 91+ day bucket sits today.

Distribution of latency, full period

• 1–7 days: 29% of invoices — weekly batching pattern
• 8–14 days: 39% — largest cohort, fortnightly batching
• 15–21 days: 14%
• 22–30 days: 8%
• 31+ days: 10% — the long tail driving the overdue pile

3.3 — When did the backlog build up?

For each Invoice-Date month, what $ value is still sitting open today?

The $53,917 December 2025 peak is the single worst month — the same month invoicing latency was at its worst (94-day median). October and November 2025 are now well-aged (91+ bucket territory). Every month past January 2026 is the “new” overdue pile forming in real time.

3.4 — Aging buckets (Square Aging Report)

The 91+ bucket is 45% of all outstanding $, across 4,680 invoices and 1,000 customers. Each week it grows as invoices in the 61–90 bucket age into it, unless collection accelerates. Industry realisation on 91+ aged receivables in Australian small-value consumer billing is typically 40–55% recoverable; the remainder requires formal write-off or collection-agency escalation. A collection-agency engagement is already in place for a subset — scope, amount, and fee basis are open items.

3.5 — Top debtors: the long-tail shape

The top debtors are many small invoices per customer, not a few large ones. Contarino is 40 separate $26–$45 invoices. This is structural to CHSP client co-contributions (small, recurring, per-service) and it means the chase strategy must work at the statement-of-account level — one monthly summary per customer — not the individual-invoice level. Chasing 40 separate $30 reminders per client is not viable and is almost certainly what has driven the existing complaints about incorrect “overdue” SMS reminders.

3.6 — Service-stream and delivery mix

Service-stream breakdown by invoice count: Domestic Assistance (DA) is the largest single cohort; Home Maintenance (HM) is second; Transport has the smallest per-invoice count but the highest $ per invoice.

Delivery mix over the period:

• SMS: 94.7% (consistent across the period)
• Email: 5.3%
• Share Manually / None: 0.05%

No paper invoicing exists today. Clients who cannot engage with SMS (digital exclusion, unsubscribed, wrong mobile number) receive nothing actionable. PostGrid Australia (approximately $1.80–$2.50 per letter, print and mail) is the recommended channel for this cohort — estimated 150 letters per month at roughly 10% of monthly volume, triggered when an invoice ages past 14 days without digital engagement. Annual cost: $3,600–$4,500.

3.7 — Payment method split (Square dashboard, Jul 2025 – Apr 22 2026)

Square fees total $7,960.94 on the dashboard versus $3,178 in Xero — approximately $4,800 understated in P&L, because Amaka is not passing through full fee data. This is a P&L accuracy item for the accountant to correct at close.

The 16.6% bank-transfer group is where the Xero clearing problem originates: when a client pays by bank transfer without quoting the invoice reference — which is very common, since family members transferring from their own accounts have no easy way to know what reference to use — the payment arrives, staff mark the Square invoice as “Invoice payment” in Square's UI, but the bank deposit in Xero has no matching counterpart, so it lands in SQ-600010 and sits.

Square ↔ Xero integration mechanics

This section explains, in technical detail, why the clearing balance exists at all.

5.1 — What the Amaka integration actually does

From Amaka's own documentation:

Amaka only syncs paid transactions. It creates a single daily summary invoice in Xero for the previous day's Square settlements. Unpaid Square invoices never appear in Xero under any configuration. This is by design — Amaka was built for retail and hospitality, where payment is taken at point of sale.

Any configuration tweak leaves this architecture in place. The only fix is to change where the invoice originates.

5.2 — Why bank transfers land in clearing

1. Mohamed creates a Square invoice.
2. The client's spouse or relative pays via bank transfer with reference “Dad's gardening” or no reference at all.
3. Money arrives in EC's bank account. Xero's bank feed creates an uncategorised deposit.
4. A staff member spots the deposit, figures out (or guesses) who it's from, and manually marks the Square invoice as paid under “Invoice Payment” status.
5. Square now thinks the invoice is paid. Amaka never learns about it (Amaka only watches card settlements).
6. The Xero bank feed still shows an uncategorised deposit. The user codes it to SQ-600010 Square Other Payment Clearing as a best-guess.
7. Nothing ever ties the clearing credit back to the Square invoice. It accumulates.

Result: $73,138 in the clearing account as at 22 April 2026 (up from $990 twelve months earlier). The account had been reduced to approximately $16.5K by 4 May 2026 following targeted clean-up work.

5.3 — The “zero Xero receivables” problem

Xero's aged-receivables report against both Square contacts (Square Card Payment + Square Other Payment) shows $0 outstanding. This is false: Square has $273K of open invoices. But because Amaka never pushes them to Xero, Xero's financial statements materially understate receivables. The CHSP Financial Acquittal will contain unexplained reconciling items unless this is fixed before close.

Three-stage remediation plan

Stage 1 — Stop the leakage (Week 1–2)

1. Export all open Square invoices and all bank transactions July 2025 – April 2026; run a fuzzy-match reconciliation (amount, date window, family-name keyword) to identify which clearing entries pair with which Square invoices.
2. Disable Square overdue SMS reminders until matching is complete — stops generating incorrect “you owe $X” messages to paid clients.
3. Freeze ad-hoc Square invoice creation in favour of the Stage-2 Xero flow (no new bank-transfer unmatched items).
4. Migrate the operational Google Sheet from a personal Gmail to an EC-owned Google Workspace domain (separate privacy-law issue).
5. Triage the 21 manually-created overdue Xero invoices ($66,918) — confirm which to chase, waive, or write off.

Stage 2 — Switch to Xero-first invoicing with BPAY (Weeks 3–6)

1. Configure Xero invoice templates per service stream (DA visit, Transport one-way / return, HM job) with co-contribution rates pre-filled and Xero contact records pre-linked.
2. Add BPAY via Parakeet (~$30–$50/month). Each invoice gets a unique CRN. Bank transfers against that CRN auto-reconcile in Xero regardless of who the payer name is — this solves the family-member-paying problem structurally.
3. Twilio SMS + Make.com workflow: when Xero issues an invoice, Make.com fires a Twilio SMS (~$0.08/message): “Hi [Name], your invoice for $X is ready — pay here: [Xero-hosted link].” The Xero payment page shows BPAY, bank transfer, and card options.
4. Square becomes a card-payment terminal only — integrated as a payment method on Xero invoices, not as an invoicing system.
5. Xero bank-feed matching rules — common client-name patterns, recurring amounts, known family-payer names → Xero auto-suggests matches for bank transfers.
6. PostGrid Australia for paper — triggered via Make.com when Xero flags a delivery method of “Manual” or when SMS fails. Estimated 150 letters per month.

Stage 3 — Automate invoice creation (Weeks 6–12)

1. Make.com automation: when the operational sheet marks a job “Complete,” auto-create a draft Xero invoice with client, service type, date, amount. Draft lands in a team inbox for a 30-second review and one-click send.
2. Replace cell-colour status tracking with a proper dropdown: Not Invoiced → Draft Created → Invoice Sent → Paid → Reconciled. Tracked by data, not colour.
3. Monday morning reconciliation digest — automated email: invoices issued this week, payments received, unmatched bank items requiring attention.
4. The current manual-invoicing role would be largely eliminated under Stage 3. A management decision on redeployment (compliance, client service, scheduling) would be needed at that point.

Costs and ROI

Annual savings target (conservative): $58K–$76K/year:

• Manual-invoicing role redeployment (salary + super): ~$42K
• Accounts team time recovered (dispute investigation): ~$6.5K–$10.5K
• Co-contribution slippage recovered: ~$6K–$18K
• Complaint-handling time: ~$3K–$5K

Plus one-time: a portion of the clearing balance recovered + a portion of the $121K 91+ bucket collected (both dependent on matching success and collection policy).

Risk register — systems + financial