River
Y CombinatorBacked by Y Combinator
FREE TEMPLATE

13-Week Cash Flow Forecast Template

A weekly cash grid seeded from your own transaction history, plus the variance ledger that tells a late payment apart from a slipping collection cycle.

Free download  ·  No account needed

Variance Ledger

[Entity], week ending [__ ____ 20__]

One row per line that missed. Opened when the week locks, against the forecast that was in force at the time.

What every cash spreadsheet already has

Line itemForecastActualVarianceVariance %Trigger
blankblankblankcomputedcomputedcomputed

What it does not

CauseClassReverses inAssumption changedReversed?
one sentenceTiming or Structurala week ending datea named rulechecked, not assumed

Three triggers, and the third is the one that pays

TriggerSet inCatches
Over the dollar floorForecast AssumptionsThe obvious misses
Over both the dollar and the percentage floorForecast AssumptionsSmall lines moving hard
Same line, same direction, twice runningAny size at allA recurring cost error, which is small every week and permanent

A timing call with no reversal week is a deferral, not a classification. This sheet will not take one.

Two customers miss by forty thousand dollars in the same week. One moved a payment run by seven days; the other has been paying eight days slower every month since October. Actual minus forecast gives both the same row. A court-appointed monitor filing a weekly cash flow does not stop there. The Fifteenth Report in the AbitibiBowater proceedings puts a paragraph reference beside every variance, pointing at prose that names the cause and whether it reverses. That column is the whole difference.

This pack builds that column in. Every Variance Ledger row carries a cause, a class of Timing or Structural, and, when it is timing, the week it will reverse in. Take the illustrative quarter shipped with it: a fastener distributor whose rolling four-week forecast error is 22.6%, and 1.9% once reversed timing variances are taken out. Both numbers are true. The gap between them is one national account that moved its payment run from the 14th to the 21st.

Placement comes from behaviour, not paper. Receipts Detail carries each customer's observed days to pay beside their stated terms, and Disbursements Detail carries the notice period every discretionary line needs, which is what turns a failing week into a list of options. Payroll is the line that never moves: the semiweekly deposit schedule puts a Tuesday payday and its tax deposit in one week and a Thursday payday's in two. Seed it from bank statement PDFs, or download the Word and CSV files.

One quarter, worked all the way through

The Variance Ledger, the 13-Week Model, Forecast Accuracy, and the filed report behind the design.

Variance Ledger

Halcourt Supply Co., an illustrative fastener distributor. Four weeks locked in January.

WeekLineVarianceCauseClassReversesReversed?
16 JanNational accounts(240,000)Centrelane moved its payable run from the 14th to the 21stTiming23 JanYes
16 JanProfessional fees(12,000)The advisory firm did not invoice in JanuaryTiming30 JanNo
30 JanProfessional feesreclassThe retainer ended on 31 December, so no invoice was ever comingStructuraln/anew row
30 JanCore collections(42,100)The 31 to 60 day bucket went 38 to 44 to 47 days, and 61 to 90 grew 118,400Structuraln/an/a
23 JanPayroll2,900Warehouse shift premium effective 5 January, never added to the assumptionStructuraln/an/a
23 JanFreight3,200Two expedited shipments covering the 3/8 hex stockoutTiming13 FebPending
16 JanInventory(4,800)13,500 of week 1 reversing in, against an 8,700 mill surcharge on PO-2304Structuraln/an/a
23 JanOther operating2,900Cannot explain: nine expense codes, nothing above 600openn/an/a

Payroll is 1.2% and clears no floor. Two consecutive adverse runs pulled it, and six runs remain in the horizon, so the line nobody would flag is worth 17,400.

13-Week Model

Weekly summary. Forward weeks start from the actual close, not from the plan.

WeekStateOpeningReceiptsPaymentsDebtEndingHeadroom
W1 09 JanLocked520,000480,540696,6000303,94053,940
W2 16 JanLocked303,940496,200407,10031,400361,640111,640
W3 23 JanLocked361,640753,750708,1400407,250157,250
W4 30 JanLocked407,250593,300449,7000550,850300,850
W5 06 FebForecast550,850503,500694,9000359,450109,450
W6 13 FebForecast359,450791,500437,20031,400682,350432,350
W7 20 FebForecast682,350516,500711,3300487,520237,520
W8 27 FebForecast487,520647,500554,0000581,020331,020
W9 06 MarForecast581,020510,500918,9000172,620(77,380)
W10 13 MarForecast172,620804,500423,20031,400522,520272,520
W11 20 MarForecast522,520523,500718,8400327,18077,180
W12 27 MarForecast327,180656,500474,0000509,680259,680
W13 03 AprForecast509,680517,500640,9000386,280136,280

Week 9 carries payroll, the second pre-build tranche, the annual insurance renewal and the rent, in a week with no national account receipt. Payroll clears. The 250,000 covenant does not.

Forecast Accuracy

Every closed week scored twice, because one of those scores is about the model and the other one is not.

WeekForecast endingActual endingErrorError %Ex reversed timingEx timing %
09 Jan334,500303,940(30,560)6.2%(2,360)0.5%
16 Jan607,840361,640(246,200)32.4%(19,700)2.6%
23 Jan166,640407,250240,61047.5%6100.1%
30 Jan578,750550,850(27,900)4.4%(27,900)4.4%
Rolling 422.6%1.9%

Error is measured against forecast receipts, and the forecast is rebased on the prior week's actual close, so nothing here is compounded drift. Week 3 forecast an ending balance 83,360 under the covenant and closed 157,250 over it, and all of that swing is one payment seven days late.

Where the design comes from

Ernst & Young Inc. as Monitor, Fifteenth Report in the AbitibiBowater proceedings under the Companies' Creditors Arrangement Act, filed with the SEC on 8 September 2009. Receipts variance table, US$000.

ReceiptsPara.ActualForecastVarianceVariance %
A/R Collections26(i)131,992176,096(44,104)(25%)
Intercompany A/R Settlement26(i)27,861027,861N/A
Total A/R Collections159,853176,096(16,243)(9%)
Collections on Behalf of Joint Ventures26(ii)39713,276(12,879)(97%)
Other Inflows26(v)19,41912,3097,11058%
Total Receipts209,366201,6807,6864%

The Para. column resolves, and this is what it resolves to

RefLineThe Monitor's own words
26(i)A/R Collections"This variance is primarily due to an increase in accounts receivable aging."
39(i)Restructuring Costs"This variance is primarily due to timing and is expected to reverse as various professional service firms submit their invoices for payment."
28(iii)Utility Payments"Due to the agreement the ACI Group has with its various power suppliers, it is required to pay a certain fixed amount on a weekly basis."

Two of those are timing and one is a collection cycle deteriorating, and the grid alone cannot tell you which is which. The Para. column is the Variance Ledger's Cause and Class, thirty years early.

What's in the pack

01

Variance Ledger sheet

Forecast snapshotted at lock time, actual, variance, the cause in one sentence, the class, the reversal week, the assumption it rewrote, and whether the reversal actually happened.

02

13-Week Model sheet

Locked weeks carry actuals and forward weeks start from the actual closing balance. Headroom against the minimum balance is a row, so a lender's question has an answer rather than a colour. For this morning rather than this quarter, the daily cash position pack opens on usable balance and dates the float.

03

Timing or Structural

The judgement guide, with the worked pairs. A payment run that moved seven days changes nothing. A cohort stretching from 38 days to 46 changes every week after it.

04

Receipts Detail sheet

One row per open invoice, placed on the cohort's observed days to pay rather than its terms, with an aging bucket, a confidence, and a reason on every override. Chasing those receipts on a dated commitment is the AR collections pack.

05

Disbursements Detail sheet

Every bill and committed order with stated terms beside observed terms, a committed or discretionary flag, and the notice period each line needs before it can move.

06

Forecast Assumptions

Every rule behind a forward week: the derivation, the period and sample it came from, the date it last changed, and the ledger row that changed it.

07

Forecast Accuracy sheet

Error per closed week measured twice, with and without reversed timing variances, plus the counts by class and how many reversals landed on the week they were promised.

08

Weekly Cash Memo

Fixed sections in a fixed order: position, the week's variances with causes, assumptions changed, the payroll test, what cannot be explained yet, and the decisions due.

09

Cash Coverage Note

Written only when a forward week fails a test. Every line that could move, its notice period, and the date the decision has to be taken, ordered by that date.

How to use it

  1. 1

    Take it blank, or seeded

    Download the empty Word and CSV files with no signup, or open the pack in River and hand it your exports instead.

  2. 2

    Send the raw history

    Bank transactions as far back as they export, both agings, and the payroll calendar. Statement PDFs and a four-header-row export are both fine.

  3. 3

    Lock the week, then class it

    Each Friday, write the actuals against the snapshotted forecast and settle every row as timing or structural, with a dated reversal week on each timing call.

  4. 4

    Roll, and test the calls

    Advance the horizon, push each structural finding into the assumption it rewrote, and check every reversal week that has now closed.

Frequently asked questions

Is this template free?

Yes. The whole pack downloads as Word documents and CSV sheets with no signup and no card. Edit with AI is the optional path for anyone who would rather hand River their exports than fill nine files by hand. Other packs are in the template library.

What is the difference between a timing and a structural variance?

A timing variance means the cash is still coming and only the week moved, so it carries the week it reverses in. A structural variance means the forward weeks are wrong until an assumption changes, so it names the assumption. A timing call that never reverses was structural all along.

Do I need bank transaction history to use it?

No. Every document and sheet works blank, and the download is the blank version. History is what lets River derive each customer's real days to pay and each vendor's observed terms. With little history behind you, the startup cash flow forecasting guide covers the first version.

What format are the downloaded files?

Four Word documents and five CSV sheets in a single zip. The documents open in Word, Pages or Google Docs; the spreadsheets open in Excel, Numbers or Sheets. No conversion step and no format that needs River to read it. When a week comes up short, the payment prioritization memo turns the gap into that Friday's run.

Is this the right shape for a lender or a restructuring adviser?

It is the shape they already read. A weekly cash flow projection attached to a forbearance agreement runs actual against budget line by line, which is what the model and the ledger produce, with the memo as the covering explanation.

How does this fit with the monthly close?

It runs beside it, not inside it. The weekly model is cash on the bank statement; the month end close checklist is the accrual ledger. Locked weeks hand the close a clean cash figure, and the close hands back the footed balances a normalized trial balance produces.

What does Edit with AI actually do?

A free River account is created and this pack installs as a private workspace. The agent starts ready for your bank export and your two agings, derives what each customer and vendor actually does, and places every open item. It writes nothing until you send it something.

Know which misses matter, and which ones are Tuesday

Download the blank pack as Word and CSV files, or open it in River and let the agent seed thirteen weeks from your own transaction history.

Edit with AI