Finance & FP&A

How to build a 13-week cash flow forecast as Restaurant and Hospitality Founders

Finance & FP&AFor Restaurant and Hospitality Founders3 apps12 steps~24 min to set up

Your 13-week cash flow forecast lives in a Google Sheet that's three days stale the moment you close it. You manually pull last week's sales from Toast or Square, guess at the food and beverage invoices that haven't hit QuickBooks yet, and estimate payroll from a 7shifts export you copy-pasted. Every Friday, you spend 90 minutes updating a spreadsheet your bookkeeper won't even see until month-end. Meanwhile, you're making real decisions — do you open a second location, hire a sous chef, extend your bar program — based on numbers that are already wrong. You need a rolling 13-week view of actual cash in and cash out, updated daily, not a snapshot you stitch together once a week from four different logins.

Build this on Starch → See the apps used

Finance & FP&AFor Restaurant and Hospitality Founders3 apps12 steps~24 min to set up

Outcome

What you'll set up

A live 13-week cash flow forecast that updates daily from your bank feeds, showing exactly how many weeks of operating cash you have before you need to pull from a line of credit or make a hard call on staffing.

Scenario models that let you test 'what if covers drop 20% in January?' or 'what if I add a private dining room and take on a $40K build-out?' before you commit — with runway and break-even dates recalculated automatically.

A spend anomaly feed that flags when your linen vendor charges double, a new subscription appears on your account, or your food cost run-rate blows past your target margin — caught in days, not discovered at month-end.

The Starch recipe

Apps, data, and prompts

The combination of Starch apps, the data sources they pull from, and the prompts you use to drive them.

Apps used

Runway Analysis Scenario Analysis Transaction Insights

Data sources & config

Starch syncs your Plaid bank account data on a schedule — transactions categorized, balances updated daily — which powers both the Runway Analysis and Transaction Insights apps. Stripe syncs on the same schedule if you process card payments through Stripe directly. For operators on Square or Toast, Starch connects to Square from its integration catalog (agent queries it live when your forecast runs), and automates Toast through your browser — no Toast API required. QuickBooks syncs on a schedule for your accrued invoices and vendor bills. Paylocity or ADP sync on a schedule for payroll runs, so your labor cash outflows hit the forecast the day payroll posts, not the day your bookkeeper reconciles.

Prompts to copy

Build me a 13-week cash flow forecast using my Plaid bank feed as the cash-in source. Pull in weekly net deposits grouped by day-of-week so I can see the typical Tuesday trough and Saturday spike. Project forward 13 weeks using the trailing 8-week average weekly net cash flow. Flag any week where ending cash drops below $30,000 — that's my floor before I need to tap my credit line.

Create a scenario that models what happens to my 13-week runway if covers drop 15% in weeks 4 through 8 — the post-Valentine's shoulder period — while my fixed costs stay flat. Show me break-even week and lowest cash balance in that window.

Show me every transaction in the last 60 days grouped by vendor category — food and bev suppliers, labor-related charges, utilities, rent, linen and smallwares, credit card processing fees, and marketing. Flag any vendor where this month's total is more than 20% higher than the prior 30-day average.

Run these in Starch → or paste them into your favorite agent

Walkthrough

Step-by-step

1 Connect your business checking and any savings or reserve accounts via Plaid — Starch syncs transactions and balances on a schedule, so your starting cash balance is always current to the prior business day.

2 Connect QuickBooks on a schedule so Starch can pull open bills and scheduled vendor payments — invoices you've approved but haven't paid yet show up as future cash outflows in the correct week.

3 Connect your payroll provider (Paylocity or ADP) on a schedule so each payroll run appears as a cash outflow in the week it processes, not as an estimate.

4 If you're on Square, connect it from Starch's integration catalog so the agent can query daily net settlement amounts live when your forecast refreshes. If you're on Toast, Starch automates the settlement export through your browser — no API needed.

5 Open the Runway Analysis app and tell Starch: 'Build me a 13-week cash flow forecast. Use Plaid net deposits as weekly cash inflow. Use QuickBooks scheduled bills and payroll run dates as outflows. Project forward 13 weeks and flag any week where ending cash falls below $30,000.'

6 Review the first draft — Starch will surface the 13-week waterfall. Verify that your rent payment, which may be a large single-day outflow, lands in the correct week and isn't being averaged across the month.

7 Open the Scenario Analysis app and build your downside case: tell Starch 'Reduce weekly inflow by 15% in weeks 4 through 8, hold all outflows fixed, and show me ending cash balance by week and the lowest point in that window.'

8 Build a second scenario for a planned investment: 'Add a one-time $40,000 cash outflow in week 3 for a private dining room build-out. Show me how runway changes and what week cash hits its floor.'

9 Open Transaction Insights and prompt: 'Group all Plaid transactions from the last 60 days into these categories: food and bev suppliers, labor, utilities, occupancy, linen and smallwares, processing fees, marketing. Flag any vendor where 30-day spend is more than 20% above the prior 30-day period.'

10 Set a weekly automation: tell Starch 'Every Monday at 7am, refresh the 13-week cash flow forecast, check whether any week's ending cash is below $30,000, and send me a Slack message with the lowest projected cash week and the balance.' Starch builds this as a scheduled automation — no code, just describe it.

11 At the end of each accounting period, reconcile your actual cash position against week-1 of the prior forecast to measure forecast accuracy. Tell Starch: 'Compare my actual Plaid ending balance for the week of [date] against what the forecast projected 13 weeks ago. Show me the variance in dollars and percentage.'

12 Share a read-only view of the forecast with your accountant or business partner — they see the same live numbers you do, without needing a separate login to QuickBooks or Plaid.

See this running on Starch

Connect your tools, describe what you want, and the agent builds it. Closed beta is free.

Try it on Starch →

Worked example

Week of March 3, 2026 — Pre-St. Patrick's Day cash plan for a 120-seat neighborhood bar and grill

Sample numbers from a real run

Opening cash (March 3)	48,200
Week 1 net card settlements (Square, est.)	32,400
Week 1 food and bev vendor payments (Sysco, local farms)	-14,800
Week 1 payroll run (Paylocity, 22 FOH + BOH staff)	-19,600
Week 1 rent (pro-rated March 1 ACH)	-8,500
Week 1 ending cash	37,700
Week 2 net card settlements (pre-St. Pat's uptick)	41,200
Week 2 food and bev (extra keg and wing inventory order)	-18,400
Week 2 payroll run	-19,600
Week 2 ending cash	40,900
Week 7 projected ending cash (post-holiday drop)	28,300
Week 7 flag — below $30,000 floor	-1,700

This operator runs a 120-seat bar and grill in a neighborhood with a strong St. Patrick's Day lift. When she builds the 13-week forecast in Starch the first week of March, the Plaid feed shows $48,200 in operating cash. The first two weeks look strong — settlements jump to $41,200 in week 2 as the St. Patrick's Day weekend lands, but she's also pre-ordered extra inventory from her beer distributor and a local chicken wing supplier, so food and bev outflows spike to $18,400 that week against a normal $14,800. Payroll stays flat at $19,600 because she's using existing staff plus two temporary bartenders she's already scheduled in 7shifts. The forecast shows week 7 — early April, two weeks after the holiday hangover — as the cash floor: $28,300, which triggers Starch's $30,000 flag. That's the signal she needed. She uses the Scenario Analysis app to model a 10% price increase on bar items starting week 5: the scenario shows week 7 cash recovering to $33,100, above the floor. She runs the promotion, adjusts her pour cost targets in the Transaction Insights dashboard, and checks back in week 6 to see whether the actual settlement trend matches the model. It does, within 4%. Her bookkeeper, who previously saw this picture in mid-April, is now looking at the same numbers on Monday morning.

Measurement

How you'll know it's working

Weeks of operating cash remaining at current net burn — updated daily from Plaid, not monthly from the bookkeeper

Weekly labor cost as a percentage of weekly net sales — payroll run from Paylocity divided by Square or Plaid settlements in the same 7-day window

Food and beverage cost run-rate versus target margin — 30-day vendor spend from Transaction Insights divided by 30-day net sales

Cash floor week in the 13-week window — the lowest projected ending balance and which week it occurs

Forecast accuracy — actual Plaid ending balance versus what the prior 13-week model projected for that week, tracked as a percentage variance to improve the model over time

Comparison

What this replaces

The other ways teams handle this today, and how the Starch version compares.

Google Sheets built manually from Toast, Square, and QuickBooks exports

Free and fully customizable, but you rebuild it from scratch every Friday and the inputs are already 48-72 hours stale by the time you open the file.

QuickBooks Cash Flow Planner

Built into QuickBooks and decent for a single-entity business, but it can't pull in POS settlement data from Square or Toast directly, so your inflow projections are based on invoices rather than actual bank deposits — meaningless for a cash-heavy restaurant.

MarginEdge

Excellent for food cost and invoice management specific to restaurants, but it's not a cash flow forecasting tool — it won't show you a 13-week waterfall or model scenarios like a slow January.

Float (cash flow forecasting SaaS)

Purpose-built for 13-week cash flow and integrates with QuickBooks, but it doesn't pull in your POS settlements, payroll runs, or reservation trends, so you still manually layer in the hospitality-specific line items that drive most of your week-to-week variance.

Your bookkeeper's monthly close

Accurate, but you see the picture 3-4 weeks after the period closes — by the time you know January was bad, February is already half over.

On Starch RECOMMENDED

One platform — runway analysis, scenario planning, transaction insights all running on connected data. Setup in plain English; numbers stay current via scheduled syncs and live agent queries.

Try it on Starch →

FAQ

Frequently asked questions

My POS is Toast, not Stripe. Can Starch still build this forecast?

Yes. Starch automates Toast through your browser — no Toast API needed. It reads your daily net settlement reports the same way you would by logging in yourself, then feeds those numbers into the cash flow model alongside your Plaid bank feed. Square users can connect Square directly from Starch's integration catalog; the agent queries it live when your forecast refreshes.

I run three locations on different bank accounts. Does the forecast consolidate them?

Yes. Connect each business checking account via Plaid — Starch syncs all of them on a schedule. When you build the forecast, tell Starch which accounts to roll up into a consolidated 13-week view and which to show separately. You can also ask it to flag the location where cash is tightest at any given week.

Will this replace my bookkeeper?

No, and it's not trying to. Your bookkeeper handles the close, tax prep, and accrual accounting. Starch gives you a live operational view so you're not flying blind between closes. The two are complementary — your bookkeeper will actually appreciate that you're not calling them every week asking for a cash balance.

How accurate is a 13-week forecast for a restaurant, given how much volume swings week to week?

The model is only as good as the assumptions you feed it. Starch uses your trailing 8-week average net cash flow as the default baseline, but you can — and should — override individual weeks for known events: a private dining buyout you've already booked in Resy, a holiday weekend, a planned closure for a kitchen renovation. The more known outflows you add from QuickBooks bills, the tighter the outflow side gets. Track forecast-versus-actual each week and the model improves over time.

Is my bank data stored by Starch? What about PCI compliance?

Starch is not SOC 2 Type II certified today — that's worth knowing if your franchisor or investors ask. Plaid handles the bank connection under their own security framework; Starch stores the synced transaction data in its own database. For PCI scope, card processing happens entirely in your POS (Toast, Square, Stripe) — Starch only reads the settlement totals, never raw card data.

My QuickBooks has open bills that haven't been approved yet. Will those show up as outflows?

Starch syncs QuickBooks bills, payments, invoices, and vendors on a schedule. Unapproved or draft bills will sync as data, but you can tell Starch in your prompt to include only bills with a 'due within 13 weeks' status, or to flag unapproved bills separately so you know the outflow number might be incomplete. One current note: QuickBooks report-level views like the P&L summary are temporarily unavailable, but entity-level data — individual bills, payments, vendors — syncs normally, which is what the cash flow model uses.

Can I set an alert when my projected cash drops below my floor?

Yes. Tell Starch: 'Every Monday morning, run the 13-week forecast and send me a Slack message if any week's projected ending cash falls below $30,000 — include the week date and the projected balance.' Starch builds that as a recurring scheduled automation. You can also add a daily version if you want same-day visibility when a large unexpected charge clears.