Measuring startup funding across five layers, from cash need to outcomes

The board of an England startup receives a term sheet for £1,000,000 in gross equity. It needs to know whether cash will clear, what remains after costs, how long the money may last and what funded work should deliver. The scenario follows those checks from the offer through to later outcomes.

Examples use named UK datasets checked on 4 September 2026. Company records, market statistics and causal evaluation answer different questions: a funding announcement or busy pipeline cannot establish that a business has succeeded. The board keeps those evidence types apart while deciding what the offer means.

The board receives a conditional offer

The company treats the offer as a decision problem, not a success measure. The board must decide when to activate a fallback, which route deserves more work, whether the offer leaves enough net cash, where diligence is blocked and whether delivery remains within budget.

Each measure gets a definition, population, denominator, unit, period, source, owner, reporting frequency, quality check and limitation. The GOV.UK Service Manual performance-data guidance advises service teams to define success and collect information that helps them understand and improve performance; the startup adapts that discipline to its own governance rather than copying public-service targets.

Related demand signals can help frame questions, but the board must state what evidence would change its own decision. A measure is useful when its meaning and role in a decision are clear, not simply because it is available or easy to count.

The board sorts evidence into five layers before interpreting the offer. Cash describes available money and when it reaches an unsafe level; process records progress through research, application, diligence and completion; transaction measures gross funds, fees, net proceeds, repayment or ownership; delivery follows spending and milestones; outcomes examine later company, investor or programme results.

The layers occur at different times. A signed agreement can improve process status while cash remains unchanged. Cleared funds complete a transaction measure, but do not demonstrate a product milestone or profit. A metric dictionary keeps each definition and denominator beside its result, including what that result cannot establish.

Cash checks separate promise from liquidity

Before treating the term sheet as funding, the finance lead records unrestricted cash, committed receipts, scheduled payments, monthly net use and the date cash reaches the board minimum.

Actual cash is compared with the approved forecast, with material variances explained. Those checks establish the fallback point before the offer changes any decision.

An application, meeting or term sheet cannot pay salaries before money arrives. Uncertain funding stays outside reconciled cash and appears in a separate scenario, with its supporting evidence and conditions recorded. That separation prevents a hoped-for receipt from concealing the date when cash may reach an unsafe level.

Minimum, target and maximum funding amounts are connected to milestones, and net proceeds are updated as current fee and deduction information becomes available. If only part of the expected amount arrives, the company revises its activities and the evidence expected from the smaller budget. The planning template can support route comparison and process tracking, while the cash record distinguishes cleared money from conditional receipts.

The bank confirms what arrived

In the worked calculation, the term sheet states £1,000,000 gross equity and cash fees and expenses total £60,000, leaving £940,000 net proceeds. If £940,000 arrives on 1 November and monthly net cash burn is £80,000, runway is about 11.8 months: £940,000 divided by £80,000.

After completion, the finance lead reconciles the £940,000 receipt to the bank and the £60,000 costs to invoices. Gross amount, deductions, cash fees, net proceeds and receipt dates are reported separately. An announced amount is not equivalent to cash, and runway alone cannot show whether funded work will achieve its milestone.

For debt, the record includes the interest basis, fees, total scheduled repayment, security, guarantee and covenant headroom. For equity, it shows fully diluted ownership before and after new shares, options and conversion. Governance and transfer rights stay in the decision record because a percentage alone does not capture them.

For grants, the record shows eligible cost, award contribution, company contribution, spend and claims. An announced award has a separate field from cash received. The route comparison explains why grants, loans and equity should not be treated as interchangeable funding measures.

Route records reveal where progress stalls

The programme lead follows evidenced stages for each route, from research and approval to contact through submission, screening, diligence, proposal, documents and conditions, ending only when cash clears. This record makes it possible to see what is blocked instead of treating all activity as completed funding.

For every case, the log captures the entry date, owner, outstanding request, next decision and closure reason where supplied. Loans, equity and grants do not share a single conversion rate: their populations and processes differ, so stage movement and elapsed time are reported by route.

Emails sent and meetings booked may help manage workload, but they are weak performance measures rather than proof of qualified progress. A smaller set of verified, suitable targets can support better decisions than broad, unqualified outreach. The launch review offers a related way to assess a service against named criteria, not a substitute for company-level evidence.

The budget is tied to delivery evidence

The board maps the £940,000 net receipt to activities, owners, evidence and dates. It follows actual spend, forecast variance, milestone completion and unresolved dependencies, while scheduling investor reporting, debt payments, covenants, grant claims and filings. The receipt is connected to planned work without being mistaken for a delivery result.

Outputs are defined separately from outcomes, with evidence and a period for each. A new employee, completed prototype or customer pilot may be an output; repeat revenue, cost reduction or technical performance can be a later outcome. Change requests are checked against contract, grant, loan and board restrictions rather than left until a later funding round reveals drift.

External numbers remain context, not verdicts

The board uses market statistics as context, not proof of company success. ONS Business demography, UK: 2024, table 1 reports active enterprises, births and deaths under the Inter-Departmental Business Register method. A business birth is not a funding round or application; the business births distinction matters when defining demand.

HMRC 2026 EIS and SEIS statistics, table 1, reports companies and investment within the defined tax schemes. Later periods can be revised as further compliance information arrives, and those totals are not all startup equity.

The British Business Bank Equity Tracker 2026 announcement, table 1, reports sector and regional activity and notes that a small number of large deals can drive substantial regional changes. Where the underlying material supports it, the board pairs investment value with deal count, stage and distribution, and states whether geography means registered office, operating site, founder residence or investor location.

The board does not estimate England from a UK total by applying population share without a sound source method. Market totals, company records and evidence about causal impact remain separate, even when they appear together in a report.

A fair comparison requires a defined population

Before selecting a source, the board writes the benchmark question and fixes company age, stage, sector, instrument, amount, geography and period. ONS Business demography, UK: 2024, table 1 reports that around 40% of new businesses survive five years, but that is a population benchmark, not a seed-stage conversion rate.

The denominator and stage definition still determine any funding benchmark. The team uses medians and distributions alongside means or totals when available, then tests sensitivity to exceptional deals, missing values and stage boundaries. Comparisons retain the definitions used by each source rather than implying that unlike populations are interchangeable.

An internal service cohort is selected and must not be labelled as the England market. The board archives the exact table, extraction date and calculation, then rebuilds the series when official data are revised. Figures based on different definitions are not joined into an apparently smooth trend.

Causation is tested before it is claimed

A change observed after funding does not prove that funding caused it. Earlier sales work, staff, market conditions and chance may also contribute. Where possible, the board records its comparison case before the outcome is known, so the planned test is not rewritten in hindsight.

HM Treasury quality guidance for policy impact evaluation examines counterfactual quality and distinguishes outcome change from causal impact. Its public-policy context does not make every method suitable for a startup, but the distinction remains useful when results are described.

A before-and-after review can help operations while leaving many explanations unresolved. Forecast against actual uses a company plan recorded beforehand but depends on forecast quality. Contribution analysis tests the chain from money to activity and result; matched comparison, difference-in-differences and randomised designs require stronger data, assumptions and specialist review.

The board uses wording that the design supports. Associated with, contributed to and caused describe different levels of evidence and are not interchangeable for rhetorical effect. An announced round, a cleared receipt and a later outcome are separate observations, not automatic evidence of a causal chain.

The board pack keeps the trail intact

The reporting pack has four views: cash, forecast variance, funding need and fallback date; process stages, evidence gaps and next decisions; transaction economics; and delivery, obligations and outcomes. A copyable workbook can use five tabs named Cash, Process, Transaction, Delivery and Outcome.

Each tab retains metric name, definition, population, denominator, unit, period, source, owner, frequency, quality check and limitation. The template can be opened in Excel, Google Sheets, Power BI or Looker Studio. A data-quality panel records stale sources, missing values, reconciliation differences, revisions and access limits, with every chart linked to the metric dictionary.

One reporting date is used across the pack. The board does not mix today's bank balance, last month's pipeline and an old market report without clear labels. After review, it records actions, owners and deadlines. Internal dashboards may contain personal or commercially sensitive information, so access is role-based and access rights have a change log.

A separate public or investor view is made from approved fields instead of circulating a screenshot. Bank, accounting, CRM, application, cap-table and contract records are reconciled before automation. Tests include duplicate cases, missing fields, delayed receipts and a corrected forecast; a correction register records old value, new value, cause, affected decisions and recipients notified.

Old definitions are preserved when a metric changes. A low error count is not a useful target if staff become reluctant to report mistakes. Headline numbers are traced through the formula to the record, and an independent reviewer is asked to explain what the result proves and what it does not.

For publication, the report states who produced it and why, along with sources, extraction date, population, denominator, period, geography, formula, method, missing data, revisions, conflicts and limitations. Small groups are suppressed or aggregated where disclosure risk exists, and official releases are reopened on publication day.

The government's Evaluation Registry guidance sets duties for relevant central-government evaluations and discusses proportionality. Private startups are not automatically covered, but pre-specifying material analyses can reduce hindsight bias. Statistical, financial, evaluation and privacy review is arranged as appropriate, with confidential source data kept apart from public evidence.

Rules taken from the decision

The scenario leaves a short set of general rules: distinguish the measurement layers, make cash status explicit, reconcile transaction values and match claims about impact to the strength of the evidence. Treat named UK datasets as context, not evidence of a company's own success.

Before the next board review, the team checks that each measure has a definition, owner, source and limitation; that the dashboard uses one reporting date; and that the comparison case was recorded before the outcome. Those checks keep the offer, delivery evidence and later results distinct in the board's decisions.