Startup Finance Kit · Guide

Startup Finance Kit: a guide to the ten workbooks

Kit release CR02-2026-10-13-v1 · data frozen on 13 October 2026 · ten Excel workbooks in English · amounts in EUR

For every file, this guide says why you need it, what it shows, what you can change, how it connects to the others and where its limits are. It takes about 25 minutes to read; the glossary in section 7 is for reference.

1. What this kit is#

The kit is a working set of finance and operations templates for an early-stage company: one integrated financial model, one quick model for top-down what-ifs, three reference books, three planning tools and two monthly reports. Each workbook is standalone, keeps its data inside the file and tells you which cells you may change.

It is written for founders who need a fundraising model, a budget and a runway view; for finance leads and analysts who want a reference structure for assumptions, KPIs, accounts and monthly reporting; and for students who want one complete, checked example instead of isolated formulas.

The case. All files describe one fictional company, Mirovest GmbH, a Berlin start-up founded in January 2027. It runs a social-investing and copy-trading marketplace for retail investors in Germany and Austria: users connect their own exchange accounts, follow verified lead traders and, once the company holds a crypto-asset service licence, can copy their trades automatically. Mirovest never holds client money or crypto-assets; it earns subscription fees, copy fees and business licence fees.

Revenue is calibrated to realistic levels, and the company runs lean: 16 people, software and AI tools budgeted at €450 per FTE a month, and fractional (part-time, bought-in) CFO and HR services from January 2028. A €12m Series A in July 2028 is the last planned round.

The plan runs for 60 months, January 2027 to December 2031, with invented actual figures for January to June 2027.

This is an illustration, not advice. Mirovest does not exist, most of its numbers are illustrative assumptions, and its regulatory, tax, payroll and accounting treatments have not been verified by a lawyer, tax adviser or auditor. Use the kit to learn the mechanics and as a starting structure for your own work. Section 8 has the full disclaimers.

If you open only one file, open D10: its Dashboard shows the plan on one page, and its How_To_Use sheet explains the rest. For a quick what-if, open D20.

2. How the files fit together#

2.1 Data flow#

All ten files rest on one set of common inputs, frozen on 13 October 2026: assumptions, people, cost lines, accounts, KPI definitions, policies, sources and an invented ledger for January–June 2027. That input pack is not part of the download; each workbook embeds the part it needs. There are no links between the files. An arrow below means “built from” or “frozen from”, never a live connection.

COMMON INPUTS   one frozen data pack (13 Oct 2026): assumptions, people, cost lines,
                accounts, KPI definitions, policies, sources, ledger Jan-Jun 2027
      |
      v
REFERENCE       D01 Assumptions        D02 KPI dictionary       D03 Chart of accounts
BOOKS           register               (what each KPI means)    and cost centres
                (what each input is)                            (where each posting goes)
      |
      v
MODELS          D10 Integrated financial model  ---------->  D20 Quick model
                3 scenarios x 2 versions,                     top-down what-ifs, seeded
                60 months, live formulas                      from the D10 Original Budget
      |
      v
PLANNING        D11 Unit economics     D12 Headcount and        D13 Budget FY2027
TOOLS           (seeded from           payroll (checked         (frozen copy of D10,
                D10 Base)              against D10 People)      Base, months 1-12)
      |
      v
MONTHLY         D18 Burn and runway, 30 Jun 2027      D19 Cost report, 30 Jun 2027
REPORTING       (actuals + frozen D10 latest          (actuals vs frozen
ON ACTUALS      estimate + D13 budget burn)           D13 budget)

What is frozen. Anything one file takes from another was fixed on 13 October 2026. If you edit D10, the companion files stay as they are; if you edit D12 or D20, D10 stays as it is. To carry a change across, make it in each file. D10 also keeps four frozen elements inside itself: the original budget (Budget_Original, a protected sheet), the released input values and key results (Scenario_Snapshot), the first-licence-year expense projection that sets the capital requirement (REG.FROZEN rows on Liquidity_RegCap) and the actual journal (Inputs_Actuals).

2.2 Conventions in every file#

• Colours. Blue text on pale yellow is an input. Black is a formula; green is a formula that reads another sheet. Type only into blue/yellow cells.
• Layout. On the time-series sheets of D10 and the companion files, months run from column H (M1, January 2027) to column BO (M60, December 2031). Column A holds the row ID; column E holds the 60-month total or closing value, built by the row’s reducer (SUM, END, AVG, RATIO, MIN/MAX, FIRST or NA). D20 has a compact layout of its own.
• Signs. Calculation sheets hold costs as positive numbers; the P&L shows them as negative numbers.
• Undefined values. A ratio without a valid denominator is stored as 0 with a VALID flag of 0, and display rows show “n/m”, “not reached” or “none within horizon”.
• Checks. D10, D20, the planning tools and the reports each have a Checks sheet; D03 has Mapping_Checks. Zero means the figures tie, within €0.01 for money; D20 shows OK or CHECK. In D11, D12 and D13 each tie has a frozen EXPECTED row (the D10 value) and a live DIFFERENCE row. After your own edit, a difference on the lines you changed is expected, not an error.
• Your copy. No file has macros, add-ins or links to other workbooks. Your edits stay in your file. Save an edited copy under a new name: a new release replaces the file, and your edits do not carry over.
• Programs. Tested in Microsoft Excel (Windows, Microsoft 365): opens, recalculates, edits and saves. LibreOffice 26.8: opens and recalculates with identical results; saving from LibreOffice is not claimed. Google Sheets: not tested. Section 6 has the details.

3. The files, one by one#

D10: Integrated financial model#

Why you need it. D10 is the core of the kit. Revenue drivers, team, costs, VAT, taxes and funding produce a monthly P&L, balance sheet and cash flow for 60 months. Use it to size a funding round, set the annual budget, see how long cash lasts and check that own funds stay above the regulatory requirement. After each month-end it re-forecasts from actuals (the latest estimate). Its Dashboard puts the plan on one page; its Valuation sheet shows dilution and an indicative exit.

What it shows.

Start with the Dashboard and the status lights on the Cover, then read PnL, CashFlow, BalanceSheet and Liquidity_RegCap.

Dashboard and Valuation. Both follow the selected case and version and only display values calculated elsewhere.

• Dashboard: one printed page by year, 2027–2031: revenue by stream, margins, EBITDA, net income, cash, burn, runway, team, customers, ARR (subscriptions and B2B), the recurring revenue run-rate and LTV/CAC, plus the milestones (when EBITDA and operating cash flow turn positive for good, and the lowest cash), the status lights and the Downside note.
• Valuation: the round table (amount, pre- and post-money valuation, price per share, new shares, fully diluted ownership), the note conversion, the cap table at exit and an indicative exit: 4.0 times the recurring revenue run-rate of December 2031 (M60), plus cash, less any note outstanding, with MOIC and IRR per investor.

In Base, the exit gives an enterprise value of €27.5m and an equity value of €34.0m; the Seed’s €4m returns 1.56 times (10.7% a year), the Series A’s €12m 0.93 times (−2.0%), and the founders keep 30.4%, fully diluted. These values are illustrative, not a valuation opinion.

What you can change (How_To_Use, sections 1–3):

• Case and version: Inputs!E3 (1 Base, 2 Upside, 3 Downside) and Inputs!H1026 (1 Original Budget, 2 Latest Estimate). The unit-economics view uses Inputs!H1023:H1025 (cohort, channel, country).
• Assumptions: Inputs!H11:J203, with Base, Upside and Downside in columns H, I and J where the cell is yellow. Black or green scenario cells follow Base by a rule.
• Monthly profiles, such as marketing spend: Inputs_TS!H11:BO50. Revenue-stream switches (present, absent, stop_new) are at Inputs_TS!H3585, H3587, H3589 and H3591, with stop months in H1500:H1503.
• People and costs: Inputs_People!H11:U70 holds 16 planned people and 44 spare slots; Inputs_Opex!H57:T70 holds 14 spare cost lines. The 46 governed cost lines are changed on Inputs.
• Funding and actuals: the FUND.* rows on Inputs, and the journal on Inputs_Actuals. The optional Series B is off: to add it, check the FUND.B terms (by default €15m at €60m pre-money in October 2030) and set SeriesBIncluded (Inputs!H1039) to 1.
• Exit: month at Inputs!H171, multiple at H172.

Worked edit (How_To_Use, section 6): a €2,000-a-month software line lowers 2027–2031 EBITDA by €126,000, the subscription plus the 5% contingency it attracts, and December 2031 cash by €128,499.94.

Status lights. Cover cells E18:E21 count the non-zero check rows in the worst month. Input and Integrity must be 0; if Integrity turns non-zero after an edit, undo the edit. Plan rows are feasibility warnings, such as cash below the buffer or capital below the requirement. Evidence rows flag policies still to verify, untested platforms and, after any edit, a stale snapshot. The shipped Base budget shows Input 0, Integrity 0, Plan 4 and Evidence 2. The Plan warnings are cash below the buffer in June and July 2027 and negative equity until the Seed arrives in August 2027 (the convertible loan is debt); the path without future rounds repeats them and adds a capital shortfall from January 2028. Evidence 2 means that the policies are still to verify and that Google Sheets is not tested.

How it connects. D10 is built from the same inputs as D01–D03. Its Base values seed D11, its People schedule is the control for D12, its 2027 budget is frozen in D13, its latest estimate is frozen in D18, and its Original Budget seeds D20. Nothing flows back.

Limits.

• D10 holds 1.41 million live formulas and recalculates all of them after every edit. For a series of edits, set Formulas > Calculation Options to Manual and press F9 at the end.
• Capacity is 60 people, 60 cost lines and 60 months. The funding structure (founders’ capital, the convertible note, Seed, Series A and the optional Series B), the revenue streams and the countries (Germany, Austria) are fixed. Not modelled: deferred tax, capitalised development, leases, dividends and option expense.
• The exit applies one multiple to one month’s run-rate: it shows how dilution and returns work, not what the company is worth.

D20: Quick model#

Why you need it. D10 builds the company bottom-up: precise, but slow to change. D20 works top-down: you type revenue, growth and costs per year, and it extrapolates 60 months of P&L, cash, funding and an indicative exit for all three cases at once. Use it to sketch a plan, test a what-if in a meeting or check a D10 result for plausibility. It has 11,500 formulas (D10: 1.41 million) and recalculates at once.

What it shows. The Dashboard gives the headline results, the milestones, the runway and the check status, with the three cases side by side and four charts; the Cover repeats the key results. Monthly and Annual hold the P&L and cash flow. Rounds_Exit shows each round’s price, new shares and ownership (simple dilution), then the exit with MOIC and IRR per round. Reconciliation compares the Base with D10.

What you can change. Inputs!D5 picks the case shown; all three are always calculated. Each case has its own column (D Base, E Upside, F Downside), with the allowed range and a status that reads CHECK when a value is outside it:

• Revenue per stream (subscriptions, copy fees, B2B licences): the amount for FY2027 and FY2028, spread over the months by ramp profiles (each month’s share of its year, on the Ramp sheet); growth for FY2029–FY2031, December on December; and a launch-year row for a stream with no revenue in December 2028, since growth cannot start from zero (in the starting values, the B2B licences).
• Gross margin by stream and year. Early years can be strongly negative, because fixed platform costs come before revenue.
• Costs: marketing; FTE times a fully loaded cost per FTE, with an annual raise; other costs plus a share of revenue; capex with a useful life. FTE, marketing and other costs follow monthly profiles in FY2027–FY2028 and are spread evenly afterwards.
• Interest and tax: interest on positive cash; one tax rate on profit after loss carry-forward.
• Funding and exit: opening cash and the founders’ shares; up to four priced rounds (month, amount, pre-money); exit month and multiple.

Reconciliation with D10 (Base; for reference, not a tie):

Revenue matches by year because the growth rates were set to reproduce it; EBITDA for 2028–2030 differs by up to about €65,000 a year because one cost per FTE cannot follow the changing team mix. Cash differs because D10 also has VAT, working capital, payment lags and later tax payments; MOIC is higher because D20 has no option pool and treats the note as equity from month 1.

What it simplifies. No VAT, no working capital, no users or individual hires, no MiCA own-funds test and no cash buffer. Tax is one combined rate (25.955% in the starting values) on the year’s profit after loss carry-forward, paid in December. The convertible note is a priced round in month 1, valued at €4.86m pre-money so that it buys the same 5,145 shares as on conversion in D10, without interest. The exit run-rate is twelve times all revenue of the exit month, not only recurring revenue as in D10.

When to use D20 and when D10. D20 is for direction: a first plan, a range of outcomes, a plausibility check. Use D10 when the answer depends on what D20 leaves out: the capital test and the cash buffer, VAT and payment timing, payroll and tax detail, actuals, or a budget you will report against.

How it connects. Every starting value comes from the D10 Original Budget, planned funding path, of the matching case. Nothing syncs in either direction.

Limits. Three revenue streams, five years, four rounds and one exit, and no capital test: for that, use D10.

D01: Assumptions register#

Why you need it. A model is only as credible as its inputs. D01 lists every input D10 uses, with its value per scenario, unit and origin. Use it to review a model, to answer “where does this number come from?”, and as a pattern for your own assumptions log.

What it shows. Assumptions: 203 inputs with scenario values, unit, provenance, source or rationale, policy, range, Verify flag and scenario rule. Of these, 146 are illustrative assumptions, 27 reported facts, 3 forecasts and 27 engineering choices; 90 carry a Verify flag. The other sheets hold the monthly Profiles, Input_Detail (every supporting field, about 41,600 rows), Sources (534 dated records, 170 with a URL) and Policies P01–P10.

What you can change (How_To_Use). ScenarioNo on the Cover (C28) switches the Active column. You can edit blue/yellow Base values, absolute case values and case parameters; inherited cases stay formulas.

How it connects. At release, D01 and D10 hold the same IDs, values, units and provenance.

Limits. D01 documents the inputs; it drives nothing. Edits stay in D01, so change D10 to move results. No source was fetched online at release. Sources without a URL cite the case’s own planning documents and research notes, which are not published. A blank cell means “not applicable”, never zero.

D02: KPI dictionary#

Why you need it. CAC, LTV, MRR and runway each come in several variants. D02 fixes one definition per KPI, so the board, investors and the team calculate it the same way. Use it as a board-pack appendix or as a dashboard specification.

What it shows. 36 KPIs, each with its formula meaning, numerator and denominator, eligible population, timing, reducer, unit, the rule for when it is undefined and its D10 row. They cover users and subscribers, ARPU, MRR, ARR (subscriptions and B2B) and the recurring revenue run-rate, which adds copy fees; B2B retention, three CAC variants, margins, LTV and payback, burn, runway, buffer gap, capital headroom, funding need, burn multiple, cost variance and cost per FTE-month. D10_Ties lists every D10 row each KPI uses.

What you can change. Nothing that changes a result. You may fill in the Owner column. The content is the same in every scenario, so D02 has no scenario selector.

How it connects. Each KPI points to its D10 row, and D11, D18 and D19 use the same definitions.

Limits. D02 defines KPIs; it calculates none of them. Its row references point to the released D10, so check them if you restructure your own copy of D10.

D03: Chart of accounts and cost centres#

Why you need it. You can compare plan and actuals only if both use the same accounts and cost centres. D03 is that bridge. Use it to set up bookkeeping, to map a ledger export before a budget-vs-actual review, or to design management reports.

What it shows.

• Accounts: 93 logical accounts.
• Posting_Map and Posting_Legs: 201 posting sources, each with exactly one owner, and their 524 debit and credit legs.
• Allocations and Mapping_Checks: 4,192 weights that report every leg in eight views (statement, nature, function, cost centre, vendor, recurrence, commitment, behaviour), with a check that each set sums to 1.
• Reference lists: cost centres, management categories, fictional vendors and eight functions, including research and development, which carries the payroll and depreciation of the engineering and product cost centres.
• How_to_map_your_GL: an 11-step method for mapping your own ledger.

What you can change. Your own account numbers (Accounts, column B) and the allocation weights (Allocations, column E). The content is the same in every scenario, so D03 has no scenario selector.

How it connects. Every actual and budget posting in D10, D13, D18 and D19 uses these account IDs and cost centres; D18 and D19 carry copies.

Limits. Logical IDs such as GL.OPEX.OPX.MISC govern the model. The numbering follows a German standard layout for illustration only (section 8), and only 2 of the 93 accounts carry a provisional number, so map the rest to your own verified chart. D03 shows how to map a journal; it does not create or import one.

D11: Unit economics calculator#

Why you need it. Before you spend on acquisition, you want to know whether a customer earns back what it cost to win them. D11 answers this for one cohort and channel in a small, fast file. Use it before a campaign or to test the marketing assumptions in the plan.

What it shows. Calculator: 36-month discounted LTV, cohort CAC, LTV/CAC, discounted payback and the cumulative path by month of age. Cohort: revenue, variable cost and contribution per connected user, month by month. B2B: separate client economics. Retention, Calendar_Seeds (frozen D10 CAC, arrivals and rates), Tail_Check and Checks complete the file. At release, the October 2027 launch cohort on the paid channel shows an LTV of €32.33 against a CAC of €104.46, which includes two months of pre-launch spend: LTV/CAC is 0.31, and payback is not reached within 36 months.

What you can change. Quick_Start lists every input:

• start months, the Austria share, subscription prices and copy fees;
• conversion, churn, variable costs, VAT, B2B terms and the discount rate;
• cohort month (Inputs!E52), channel (E54), CAC per channel (E55:E58) and ARPU per stream;
• the retention curve (Retention!H11:AQ11), which should not rise with age.

Worked edit: a paid CAC of 25 instead of 104.46 lifts LTV/CAC to 1.29 and brings payback to 27 months. Checks then shows a difference of −79.46 against D10, which is expected.

How it connects. D11 is seeded from D10 Base. Six ties check it against D10: five on the launch cohort and paid channel, and one on a fixed cohort that starts in February 2031 (Tail_Check), whose later months fall after December

2031. Other cohorts and channels differ by design.

Limits. The horizon is 36 months with no terminal value, and B2B stays separate from retail users. Cohorts other than the launch cohort take their CAC from the frozen D10 calendar, and cohort months after December 2031 use frozen rates from D10’s extended columns, which run to November 2034. Exchange rebates (R3) are off.

D12: Headcount and payroll plan#

Why you need it. People are most of a start-up’s cost. D12 turns a hiring plan into the full German employer cost per person and month: salary with raises, each social-insurance contribution up to its ceiling, levies and accident insurance, plus recruiting, software and AI tools, and IT equipment. Use it at budget time and before each hire.

What it shows. Summary gives headcount, FTE and payroll by component per month (€7.76m over 60 months at release). Cost_Centres splits costs by cost centre, Hiring_Bridges shows recruiting and software as costs and IT as capex, and Payroll covers every person and month. The input sheets hold the 16 people and 44 spare rows, the founders’ salary and status steps, and rates and ceilings for 2027–2031.

What you can change (Quick_Start). Round months and switches, contribution rates, recruiting costs and the IT replacement cycle. Per person: cost centre, hire and exit month, FTE, salary and raise, insurance status, tools and recruiting route; a spare row becomes active once you enter its hire and exit months. Yearly pension rates and ceilings. Worked edit: hiring the finance manager (P17) in April instead of January 2028 saves €23,412.43 of payroll over the plan.

How it connects. D12 computes payroll itself; 17 ties compare each component with D10. Make the same edit in D10 to carry it into the model.

Limits. Like D10, D12 has 60 person rows: 16 used and 44 spare. In the case, the founding managing directors are treated as exempt from social insurance until the Series A (policy P07, to verify). Rates and ceilings for 2027–2031 are forecasts, so verify them before real use. There are no bonuses.

D13: Budget FY2027#

Why you need it. A budget is the yardstick for the year, so once approved it must not move. D13 is the approved FY2027 Base budget, frozen. It is the reference for budget-vs-actual in D19 and burn variance in D18, and it shows the detail a working budget needs.

What it shows. Summary: the monthly P&L from revenue to net income. For FY2027, revenue is about €5,900, because the app launches in October, and EBITDA is −€1.54m. P_and_L shows every account × cost centre × month. Cash ends 2027 at €3.59m, Headcount reaches 13 people by December, and Capex lists the asset additions. Frozen_Data and Posting_Inputs hold the values.

What you can change. Nothing, by design. For a what-if, overwrite a value on Frozen_Data in your copy. In the Quick_Start example, a €1,000 higher January cost lowers EBITDA by €1,000, while cash, capex and headcount stay unchanged because they read their own rows.

How it connects. D13 is a frozen copy of D10 (Base, Original Budget, months 1–12); D10 keeps the same original budget on Budget_Original. 59 ties check D13 against both.

Limits. D13 covers Base only, calendar 2027, with a fixed set of postings. P&L edits do not flow into cash, tax or headcount; make those changes in D10, and payroll detail in D12.

D18: Burn and runway report, June 2027#

Why you need it. This is the monthly cash conversation: cash on hand, how fast it is used, burn against budget, and when cash or regulatory capital would run short, both with the planned rounds and with no new money. Founders use it every month and in board meetings.

What it shows.

• Runway: first negative cash, first buffer breach and first capital shortfall on both funding paths, plus June cash, buffer, headroom, trailing burn and burn multiple.
• Burn: actuals for January–June and the latest estimate to December 2031, from opening to closing cash, for both paths.
• Burn_Variance: actual burn against budget, with a note for each month on what drove the difference.
• Levers: three cash-saving levers, at zero until you enter a saving.
• Supporting sheets: the journal, bank, VAT invoices, FTE, frozen D10 results and checks.

At release, in Base, cash on 30 June 2027 is €576,550. With planned rounds cash stays positive, but it dips below the six-month buffer target in July 2027, the month before the Seed arrives. With no future rounds it turns negative in November 2027. Net burn for January–June was €424,650, 5.3% below budget.

What you can change (How_To_Use):

• Case: Cover C28 selects Base, Upside or Downside; actuals stay the same. With 3, Runway shows the first statutory shortfall in October 2031, and Burn shows headroom of about −€103k in that month.
• Levers: Levers!B11:D13 hold the monthly saving, one-off cost and start month.
• Postings: amounts on Journal, VAT_Invoices, FTE_Actuals and Bank, within the existing postings. Keep the IDs and keep debits equal to credits.

The other values on the Cover, such as the cut-off month and the buffer multiplier, are fixed report constants.

How it connects. D18 shares its ledger with D10 and D19. Its forecast is the frozen D10 latest estimate for each scenario, and its budget burn is the original budget, tied to D13.

Limits.

• The report date and the postings are fixed. New postings or a new cut-off need a regenerated file.
• D10_LE holds frozen results, not the D10 engine.
• Levers show direct cash effects only, with no knock-on to tax, interest, staffing or capital.
• Before the licence (January 2028 in Base), negative headroom is a readiness gap, not a breach.

D19: Cost report, June 2027#

Why you need it. This is cost control: what was spent, against which budget, where and why. D19 compares accrued costs for January–June 2027 with the frozen FY2027 budget and adds a commentary template. Budget owners and the finance lead use it at each month-end.

What it shows.

• Cost_Report: actual against budget, monthly and year to date, in euros and per cent; a positive variance means over budget. It also shows non-recoverable VAT, a bridge from accrued cost to cash paid, and cost per FTE-month.
• YTD_Detail and Monthly_Detail: costs by account and cost centre.
• Cost_Views: the same costs by cost centre, nature, vendor, recurring or one-off, committed or discretionary, and fixed or variable.
• Commentary: six topics, each with context drawn from the postings (for example, a May legal overrun of €13,800), and columns for cause, action, owner and due date.

At release, actual costs are €497,112 against a budget of €520,374 (4.5% under), which is €14,203 per FTE-month.

What you can change. The Commentary text (C11:F16), and amounts on Journal, VAT_Invoices, FTE_Actuals and Bank within the existing postings.

How it connects. The budget is D13, the actuals are the ledger shared with D10 and D18, and the accounts come from D03.

Limits. D19 always uses the original Base budget, which is the same in every case, so it has no case selector. The postings and the report date are fixed, as in D18. Budget is assigned to vendors through the vendor templates of the actual postings; budget with no single named vendor stays under the placeholder UNKNOWN_FORECAST, and a vendor without a budget line, such as the bank, shows n/m.

4. Typical workflows#

Nothing syncs between files: when a step changes a figure that another file also uses, repeat the change there.

Build a fundraising model.

1. Sketch the plan in D20: revenue, growth, team and rounds.
2. In D01, note which inputs are illustrative or flagged Verify: those are the numbers investors will question.
3. In D10 (Base, Original Budget), replace the case inputs with yours: prices, conversion and churn, marketing spend, team, costs and rounds. Switch off any stream you do not have.
4. Check that the Input and Integrity lights show 0. Then read the Dashboard, CashFlow and Liquidity_RegCap: when cash falls below the buffer, when it would turn negative without future rounds, and the funding need.
5. Switch to Upside and Downside to show the range. The Valuation sheet shows dilution and an indicative exit; for a later round, switch on the optional Series B.
6. Use D11 to test whether paid acquisition pays back, and D12 to detail hiring cost, then repeat those changes in D10. Use D02 definitions for every KPI in the deck.

Test a quick what-if.

1. In D20, pick the case to display on Inputs!D5; all three are calculated anyway.
2. Change that case’s column: for example, lower the FY2029–FY2031 growth, add FTE or move the Series A later.
3. Read the Dashboard, with the three cases side by side. Checks must read OK.
4. If the result matters, rebuild it in D10 with drivers (prices, conversion, churn, hires, cost lines, rounds) and check what D20 cannot: the capital requirement and the cash buffer.

Budget season.

1. Plan next year’s hires in D12, then enter them and next year’s cost lines in D10.
2. Make sure every line has an account and a cost centre in D03.
3. Use D13 as the pattern for the approved budget: account × cost centre × month, plus cash, capex and headcount, frozen once approved. Later changes go into the latest estimate.

Monthly close and budget vs actual.

1. Map your ledger export to accounts, cost centres and vendors with How_to_map_your_GL in D03.
2. In D19, read Cost_Report, then the detail sheets and Cost_Views, and record explanations in Commentary.
3. In D18, read Runway, Burn and Burn_Variance, and test savings with Levers.
4. Set D10 to version 2 to see the latest estimate.

D18 and D19 hold a fixed set of invented postings; for your own close, copy their layout and logic.

Prepare a board pack. Start with the D10 Dashboard, which prints on one page for the selected case and version. Add the statements and KPIs from D10 with its status lights, Runway and Burn from D18, and Cost_Report, Cost_Views and Commentary from D19, with D02 as the definitions appendix. The monthly sheets of D10 print all 60 months on one A3 page width, so the type is small: use Annual or copy the months you need. If you show the Downside, label it as a stress case; if you show the Valuation sheet, label it as illustrative.

Check unit economics before a marketing push.

1. In D11, choose channel and cohort, enter your CAC, ARPU, retention and costs, and read LTV/CAC and payback.
2. In D10, raise the INPUT.ACQ.SPEND profile for the months of the push. Watch Unit_Economics and the buffer and funding need on Liquidity_RegCap.
3. Keep the CAC variants in D02 apart: the marginal paid CAC of a period is not the cohort CAC that LTV/CAC uses.

5. Scenarios and versions explained#

5.1 Base, Upside and Downside#

The three scenarios share the model, the actual months and the funding plan. They differ in demand, prices and some costs. The main differences, from D10 Inputs:

The funding plan is the same in all three: founders’ capital (€25,000) and a €1.0m convertible loan in January 2027, a €4m Seed at €8m pre-money in August 2027 and a €12m Series A at €20m pre-money in July 2028. No later round is planned; a Series B can be switched on (section 3, D10). In Base, EBITDA stays negative in every month to December 2031, although the monthly loss narrows to about €26,000. Operating cash flow is still negative in every month through November 2031 (M59). The Dashboard shows it positive from M60, but that single month is a year-end accrual effect, not a turn: the €9,000 annual-accounts cost for 2031 is booked in December and paid in March 2032, and without it M60 would be about −€2,900.

Key results at release (D10, Original Budget, planned rounds):

The Downside is a deliberate stress case. It keeps the common funding plan, with no round after July 2028. Weaker demand and lower prices keep it loss-making, so own funds fall below the MiCA own-funds requirement from October 2031 (M58) to December 2031, by at most about €0.72m in M60, while cash stays positive. The model is not adjusted to hide this. The D10 Cover, How_To_Use and Dashboard say:

Downside (3) is a deliberate funding stress case: its capital shortfall is intentional, and its Plan warning on Checks stays visible.

In the original budget with the release inputs, own funds fall below the MiCA own-funds requirement from M58 (October 2031) and stay below it to the end of the horizon (M60); the largest shortfall is about EUR 721k in M60 (CHECK.PRUDENTIAL). Month-end cash stays positive throughout.

Latest estimate (VersionNo 2): short from M58, by at most about EUR 697k in M60. The model never adds balancing funding.

How_To_Use labels the selector itself: “ScenarioNo: Inputs!E3 = 1 Base, 2 Upside, 3 Downside (a deliberate funding stress case: see rows 6–9).” D18 and D19 add: “The warning is intentional: an illustrative stress result, not a forecast or a legal assessment.”

A compliant Downside plan would need new equity before October 2031 or lower costs; this version does not model that. You can test a remedy yourself, for example by switching on the optional Series B.

5.2 Original Budget and Latest Estimate#

The version selector in D10 is Inputs!H1026.

• 1, Original Budget: the plan made before the year began, forecasting January 2027 to December 2031. Its Base values are also frozen in Budget_Original and in D13, and they are the yardstick for cost and burn variances.
• 2, Latest Estimate: actuals for January–June 2027 from the journal, then a forecast from July 2027 that restarts from the June closing balances, not from the plan.

Actuals are the same in every scenario, but the tests applied to them (capital requirement, headroom, funding need, buffer) follow the selected scenario. D11, D12 and D13 use Base and the Original Budget; D20 starts from the Original Budget of each scenario; D18 uses the Latest Estimate for all three scenarios plus the original budget burn; D19 compares actuals with the original Base budget.

5.3 Actual months: January–June 2027#

The first six months have invented actuals: 1,137 journal postings covering opening funding, payroll, suppliers, VAT and capex, plus a bank statement, VAT invoices and FTE by person and month. They are pre-revenue: the licence application is filed in June 2027, and the app launches in October 2027. The cut-off is 30 June 2027.

5.4 Two funding paths#

D10 and D18 show every cash and capital test twice: with the planned rounds, and with no future rounds. The D10 Cover defines the second path: “Future-rounds path removes all post-June rounds, including signed rounds.” Here that removes the Seed and the Series A. The model then recomputes what depends on those rounds, such as round fees, founder pay, note maturity, interest and tax; in D10 these rows start with ALT. In Base without future rounds, cash turns negative in November 2027 and falls to about −€9.3m by November 2031. The model never adds funding to close a gap. It reports the gap.

6. How the kit was verified#

Verification answers one question: does each workbook calculate what its written rules say, in the program you use? It does not show that the assumptions are realistic or that the treatments are legally correct (section 8). All checks were run on the final release files, identified by their checksums.

1. An independent re-implementation, compared cell by cell. A second calculation engine was written from the written rules and the input data alone, without access to the code that builds the workbooks or to their results. Its output was compared with copies of D10 recalculated by Excel for three scenarios, two versions and two funding paths, plus the actual months and the frozen original budget: 14 sets and 458,562 values, including the run-rate, milestones, round table and exit valuation, and the Series B rows, which must stay at zero while the round is off. No value differed beyond tolerance (€0.01 for money, 0.0000001 for ratios, 0.000001 for counts; flags, dates and share counts exact). The eight companion files were checked separately against the same source data, D10 and D13; all eight pass. D20 follows its own simplified rules, so it is outside that comparison: a separate implementation of those rules reproduced 9,766 of its values with no difference.

2. Microsoft Excel on Windows. All ten files were opened, fully recalculated, saved and reopened with no error cells. In D10, all six combinations of scenario and version were switched and gave the expected values at strict tolerances. D10 was also tested in eight cases of 3,003 values each: the baseline, a stream switched off, new sales stopped, a replaced person, a replaced cost line, the worked edit, full capacity (all 60 person and 60 cost slots filled) and one row beyond capacity, which correctly raised the capacity check; the edited copies were saved and reopened. The worked edits in D11, D12 and D13 and test edits in D18 (the Downside selected) and D19 (one FTE halved) were saved as copies and reopened, and D20 was run through 18 cases. Every value matched its expected value.

3. LibreOffice 26.8. All ten files open and recalculate in LibreOffice 26.8.1 with the same values as Excel. In D10, all six scenario and version settings and the edit tests give the same results, as do the worked edits in D11, D12 and D13. Saving from LibreOffice is not claimed: it writes numbers with 15 significant digits and turns TRUE/FALSE values into formulas. If you work in LibreOffice, keep the downloaded file as your reference copy.

4. Google Sheets. Not tested yet, and neither are other spreadsheet programs or other versions of Excel.

5. Release scan. An automated scan read every part of every release file, checking for macros, links to other workbooks, hidden sheets, local file paths, author details and a list of forbidden terms. All ten files passed.

6. Independent audit. A reviewer who built none of the files recomputed, from the common inputs and for all three scenarios, revenue, the P&L, cash, the run-rate, the milestones, the round table, the exit values and the capital headroom. Every figure agreed with the workbooks to within €0.0000001. The verdict was “correct with issues”: the open points concern presentation and wording only, and none of them changes a statement, a KPI, a status light or a tie.

The checks cannot tell you whether the Mirovest assumptions fit your business, whether the treatments in policies P01–P10 are right (they remain “to verify”), or whether a copy you have edited is still consistent. For the last, use its Checks sheet.

7. Glossary#

8. Disclaimers#

These statements appear in the workbooks and apply to this guide.

Mirovest GmbH is fictional. Illustrative calculation — not investment, tax or legal advice.

Crypto-assets are highly risky. Copy trading can lose money, including the entire amount invested.

Actuals are fictional. Every person, vendor and transaction is invented.

Verify flags and source access dates are inherited from the frozen data. No source was fetched online.

On the Downside scenario:

The warning is intentional: an illustrative stress result, not a forecast or a legal assessment.

On the valuation:

Illustrative calculation for the fictional Mirovest GmbH — not a valuation opinion and not investment, tax or legal advice.

Approved limitation statement:

Regulatory, tax, payroll and accounting treatments (P01–P10) are illustrative assumptions for a fictional company and have not been verified by a lawyer, tax adviser or auditor. Account codes follow a DATEV SKR04-style layout for illustration; they are not an official DATEV mapping.

The ten policies are: P01 service classification and licence passport; P02 prudential expense base and own funds; P03 copy fee charged as principal, and trader payouts; P04 app-store sales and commission; P05 supplier VAT and input-VAT recovery; P06 exclusion of exchange rebates; P07 founder payroll and social-insurance status; P08 tax and book treatments; P09 funding and share terms; P10 audit from the licence year.