An impact investment memorandum should let another person understand why funding was approved, declined or made conditional, which uncertainties were accepted and who will oversee implementation. Its purpose is to connect an identified need, a plausible intervention, an appropriate legal relationship and a decision about resources. For the investor, it preserves the reasoning available when the decision was made, so that later results can be assessed against the original justification. It should support scrutiny rather than simply make a project presentation more persuasive.
Turn an intention into a decision others can examine
The starting point is a project with identifiable objectives, beneficiaries and indicators. However, identifying the intended participants and measurements does not establish that the intervention will work, that the applicant can deliver it or that the financial instrument is suitable. These questions need separate treatment. An organization may deliver services well but struggle with accountability. A business may generate revenue while causing social harm. A promising initiative may need a donation because it has no realistic source of repayment.
A useful opening states the requested decision, amount, recipient, purpose, instrument, duration, principal risks and proposed conditions in a few paragraphs. Evidence and analysis then support that summary. Readers should find both the strongest reason to provide funding and the strongest reason to decline it. A trustworthy memorandum identifies missing information, admits uncertainty and explains why the recommendation remains defensible. If every argument points toward approval, the document may be functioning as a sales presentation rather than an examinable decision record.
Define the legal relationship before discussing returns
Donations, sponsorships, grants or subsidies, and investments seeking financial returns create different relationships. A donation supports an agreed purpose under its conditions without ordinarily creating equity ownership or a right to financial earnings. Sponsorship involves specified benefits in exchange, such as institutional visibility, which must be described and assessed. A grant or subsidy follows its applicable regime and instrument. Where financial returns are expected, the memorandum must explain the economic right, who must fulfill it and the potential source of payment.
A Brazilian nonprofit association does not offer equity interests in the same way as a limited liability company, or LTDA. Describing a funder as entitled to a percentage of the nonprofit does not resolve that incompatibility. The parties must examine a relationship consistent with its legal nature and project. Business status also needs careful checking: an individual microentrepreneur, or MEI, cannot admit equity partners while remaining a MEI. These questions come before valuation, conversion or exit discussions and require legal support appropriate to the case.
Loans, receivables transactions, equity and convertible instruments allocate rights, losses and enforcement options differently. Expected revenue from hoped-for future sales is not an existing, eligible receivable. The memorandum should name the actual relationship instead of relying on an ambiguous label such as partnership funding. For a donation, describe expected results through social outcomes and project delivery. Collective benefits do not become promised financial returns to the donor, and attaching the word investment to the contribution does not create a repayment obligation.
Define the intervention and the people it could benefit
An impact thesis starts with a concrete difficulty and a defined population. Improving employability among vulnerable people is a broad intention. Providing training and job-search support to adults in a particular area who face transport and family-care barriers creates a more assessable proposal. The memorandum should explain eligibility, selection, the barriers the intervention will address and the needs outside its scope. This makes it possible to examine whether the activities and budget fit the people the project intends to reach.
Clear boundaries also expose a common distortion: selecting people already close to achieving the desired outcome and then attributing their success to the intervention. If hiring is the only valued indicator, staff may favor applicants with recent experience and full availability, excluding people facing greater obstacles. Investors should examine how admission criteria, service adaptations and targets interact. A high placement rate could reflect effective support, easier starting conditions or both. Recognizing this incentive conflict before implementation helps prevent misleading interpretations of later reports.
Direct beneficiaries should also be distinguished from other stakeholders. Employers may benefit from access to candidates, while participants seek income, safety and work compatible with their responsibilities. These interests overlap in some respects and diverge in others. Record who was consulted, the needs they described and the choices that followed. A community leader's account is relevant evidence, but it cannot represent every participant's circumstances. Listening across different conditions of participation helps reveal barriers that a single spokesperson or aggregate target may conceal.
Explain how activities could produce the intended outcome
The causal chain should connect resources, activities, outputs and expected changes. Funding pays for staff, accessibility and follow-up; staff deliver training and support; participants complete stages; some obtain work and remain employed for an agreed period. Every connection contains an assumption. Completing a course does not guarantee employment, and an initial hire does not establish lasting improvements in income or well-being. The memorandum should distinguish delivery the organization controls from outcomes influenced by employers, household circumstances and other external factors.
For each important link, record supporting evidence, weaknesses and alternative explanations. Attendance might improve because transport assistance removes a barrier, but also because selection favored people with fewer constraints. Job placements might reflect the project, an expansion in local hiring or both. This does not require turning every initiative into an academic experiment. It requires avoiding the assumption that sequence proves causation: a change observed after participation was not necessarily caused by participation. The strength of the claim should match the strength of the evidence.
The chain also guides intermediate indicators. If the main hypothesis concerns removing access barriers, counting certificates alone will be insufficient. Transport-related absences, suitable schedules, referrals and withdrawals may show where delivery needs adjustment before the cycle ends. Explain which signals would lead staff to change an activity and which would require reconsidering the underlying thesis. Changing a class schedule can preserve the original rationale; discovering that the proposed support does not address the main barrier may require a more fundamental decision.
Assess additionality without presenting estimates as certainty
Additionality asks what the funding and intervention add to what would probably happen without them. Financing additionality concerns whether the project would proceed at the same scale, speed and quality without this contribution. Impact additionality concerns whether beneficiaries would achieve similar outcomes without the intervention. These are separate questions. A project may depend entirely on a donation yet create little additional change if it provides a service already available and accessible to the same population.
Look for verifiable evidence: committed resources, unused capacity, waiting lists, alternative services, access barriers and comparable previous outcomes. A claim that no other funder will support the initiative needs examination. New funding might replace a contribution another funder would otherwise provide. That does not automatically make the decision inappropriate, but it changes the justification. The memorandum should explain whether the contribution expands services, brings them forward, improves quality or mainly changes who pays for activity that would have happened anyway.
Comparisons can help assess attribution, provided their limitations remain visible. People who did not receive support may differ from participants, and historical data may reflect different hiring conditions. Where evidence is weak, use an explicitly qualified reference assumption or range rather than an unsupported precise causal figure. State what was observed and what relies on assumptions. This lets the committee test the conclusion without confusing an attractive explanation with established causality, and helps prevent a persuasive narrative from carrying more weight than its evidence warrants.
Choose indicators that guide decisions and respect people
Each important indicator needs a definition, denominator, source, reporting frequency, responsible person and rule for missing data. Training completion might mean minimum attendance, a practical assessment or a specified combination of activities. These definitions yield different numbers. Sustained employment also requires an agreed follow-up period and confirmation criteria. Without consistent rules, apparently comparable reports may measure different things, and performance may improve on paper simply because an interpretation changed. Missing responses should remain distinguishable from verified positive or negative outcomes.
Combine delivery, outcome and protection measures. Useful examples include participants admitted, verified completions, employment status and complaints about services. Read them together: a high completion rate achieved by excluding people with disabilities differs materially from a similar rate achieved through inclusive access. Breakdowns should answer a defined question and receive appropriate protection. In small groups, removing names may not prevent identification. Reporting more detailed categories can therefore carry a privacy cost that must be justified by their usefulness.
Measurement costs belong in the budget. Excessive proof requirements can consume staff time, distress participants and divert resources from support. Minimum operational records, sample-based checks and targeted investigation of inconsistencies may offer a proportionate approach. Explain why each item is needed and which decision it can change. Collecting information merely because it might someday be useful creates exposure without a defined benefit. Information quality includes proportionality, reliability and consistent interpretation, rather than volume alone.
Organize evidence by source, limits and unresolved questions
Readers should be able to distinguish examined documents, applicant statements and analysts' inferences. A signed contract, management spreadsheet and interview have different evidentiary limits. For each decisive claim, retain a document reference, version, reviewer and identified limitation. This prevents an estimate mentioned in a meeting from becoming a confirmed fact through repeated revisions. A clear reference trail also allows a later reviewer to understand what was available at the time without assuming that subsequent information informed the original decision.
Diligence should cover the institutional capacity relevant to delivery: identity, authority to sign, staff experience, financial position, existing commitments, important suppliers and payment controls. A history of serving participants may support an assessment of operational capacity, but does not establish outcome quality by itself. Photographs and testimonials can illuminate experiences without demonstrating scale, causality or absence of harm. Consistent evidence from different sources is more useful than a large collection of promotional materials repeating the same claim.
Unresolved matters need explicit treatment. Some block a decision, such as uncertainty over who can bind the organization. Others can become conditions before disbursement, such as completing a data-access protocol. Others remain inherent uncertainties, such as employers' future willingness to hire. State how each gap affects the recommendation, who will address it and who will verify the response. Receiving a document does not necessarily resolve the underlying problem: a written protocol, for example, still needs workable responsibilities and access arrangements.
A worked example: a donation for economic inclusion
Consider an entirely hypothetical nonprofit association requesting a R$ 240,000 donation for a twelve-month cycle of training and employment support for 120 adults. Planning assumptions are 96 completions and 60 sustained employment outcomes under an agreed definition. These are arithmetic targets, not observed results or promises of hiring. The contribution is restricted to the project and creates neither equity ownership nor financial remuneration for the donor. For this exercise, all 60 employment outcomes concern people among the 96 completers. That overlap is an assumption requiring individual confirmation in actual measurement.
The budget allocates R$ 144,000 to training and follow-up, R$ 36,000 to accessibility and participant protection, R$ 24,000 to monitoring and verification, R$ 24,000 to coordination and administration attributable to the project, and R$ 12,000 to a reserve. These amounts total R$ 240,000. Explain what each category covers and how shared costs were allocated. Treating all administration as waste would hide necessary work, including financial reconciliation, supervision and incident response. The allocation should make those functions visible enough to assess their necessity and cost.
With the full budget spent, planned costs are R$ 2,000 per person admitted, R$ 2,500 per expected completion and R$ 4,000 per expected sustained employment outcome. The respective denominators are 120, 96 and 60. Each ratio answers a different question; choosing whichever looks most favorable would distort the comparison. Preserve all three and distinguish planned from actual costs. Actual ratios depend on verified expenditure and results. None of these calculations independently measures causality, job quality or whether the support reached those facing the greatest barriers.
Connect disbursements, delivery and cash availability
In the same example, proposed disbursements are R$ 120,000 at the start, R$ 80,000 after review of the first quarter and R$ 40,000 after review of the second. Forecast outflows across the four quarters are R$ 60,000, R$ 80,000, R$ 70,000 and R$ 30,000. Both inflows and outflows total R$ 240,000. The forecast assumes full use of the budget, including the reserve. If the contingency is unnecessary, the agreement must determine how the remaining balance is treated.
Assuming no other resources and receipt of each installment at the beginning of its scheduled quarter, cash closes the first quarter at R$ 60,000. In the second, receiving R$ 80,000 and paying R$ 80,000 leaves R$ 60,000. In the third, receiving R$ 40,000 and paying R$ 70,000 leaves R$ 30,000. Paying R$ 30,000 in the fourth reduces cash to zero. Budget sufficiency and cash sufficiency are therefore related but distinct: obligations may fall due before an approved installment is released.
The first-quarter review must occur early enough to decide the second installment. If its R$ 80,000 is delayed, the existing R$ 60,000 cannot cover all R$ 80,000 of forecast second-quarter outflows, creating a R$ 20,000 gap. Specify notice procedures, priorities and scheduling adjustments without assuming rescue funding. Conditional installments may limit donor exposure, but poorly designed conditions can destabilize staff and participants. The review timetable should protect essential services and make clear how commitments will be handled while a release decision remains pending.
Interpret results and costs without inventing attributable impact
To explore additionality, suppose a preliminary analysis estimates that 36 of the 120 people would obtain sustained employment without the project. If 60 employment outcomes were observed, the difference from that reference would be 24. Dividing R$ 240,000 by 24 gives R$ 10,000 per estimated additional outcome. The calculation is correct under the assumption, but its meaning depends entirely on whether 36 is a defensible reference. It is not an assured causal measurement merely because the arithmetic is precise.
Sensitivity analysis makes this dependence visible. If 48 people would have achieved the outcome without support, the estimated difference becomes 12 and the cost per estimated additional outcome rises to R$ 20,000. If the reference is 24, the difference becomes 36 and the cost is approximately R$ 6,670. Expenditure and the 60 observed outcomes remain unchanged across all three scenarios. Only the counterfactual assumption changes. The committee should therefore examine its basis before using the ratio to compare proposals or declare one intervention more effective.
The remaining participants also require attention. The 24 who did not complete may have encountered barriers the project could address. The 36 who completed without achieving sustained employment may have acquired useful skills, but that does not make them successes under the original employment indicator. Additional benefits need their own evidence. Preserve the original target, report secondary outcomes separately and explain shortfalls. Changing the language after delivery should not conceal a result below expectations or erase the experience of people whose needs remained unmet.
Make beneficiary protection a condition of delivery
Identify plausible harms arising from the intervention itself. In this example, risks include disclosure of economic circumstances, humiliating selection procedures, harassment, misuse of images and referrals to unsuitable work. A statement of respect is insufficient. Protection requires assigned responsibilities, accessible reporting channels, staff guidance, referral criteria and response procedures that consider the safety of the person reporting a concern. The memorandum should show how protection works during everyday delivery and who can act when normal reporting lines are implicated.
Access to services should not depend on permission to use a participant's image or personal story in promotion. Assess which information the donor actually needs. Aggregate reports may support oversight while individual records remain in restricted environments accessible only to authorized people. Data processing requires examination of its purpose and applicable requirements; a generic authorization does not replace that analysis. Guidance from Brazil's National Data Protection Authority, ANPD, is a reference entry point, while the specific arrangements still require appropriate assessment.
A serious allegation may justify suspending a supplier or activity without automatically abandoning participants. Combine immediate protection, investigation and safe continuity where possible. If services pause, specify who communicates, what support remains and how retaliation will be prevented. Investors should monitor response quality rather than reward silence through a zero-complaints target. A channel receiving reports can indicate trust and detection capacity. An absence of complaints can also reflect fear or inaccessible procedures, so the count alone cannot establish that delivery is safe.
Test adverse scenarios before authorizing funding
A weaker-outcome scenario might spend the same R$ 240,000 but produce only 72 completions and 36 sustained employment outcomes. Cost per completion would be approximately R$ 3,330, and cost per observed employment outcome approximately R$ 6,670. Cost per person admitted would remain R$ 2,000 if admissions stayed at 120. Reading these indicators together prevents a stable entry measure from hiding lower effectiveness. Investigate causes before concluding that execution was poor: delivery failures and external changes can require different responses.
Cost pressure creates another decision. If training and follow-up exceed their R$ 144,000 budget by R$ 9,000, the R$ 12,000 reserve could absorb the difference if the agreement permits it and the required authorization is obtained. The main category becomes R$ 153,000, leaving R$ 3,000 in reserve, while the overall ceiling remains R$ 240,000. The reserve is part of that ceiling. Moving money from it does not create new funding, and the same amount cannot also be counted as remaining contingency capacity.
Some scenarios may require reduced scope or closure of an activity. Losing critical staff, an unworkable referral pathway or persistent protection failures can make continuation harmful. Define warning signals, response periods, decision authority and treatment of existing commitments. Interruption criteria should neither punish every fluctuation automatically nor permit indefinite continuation despite mounting evidence. A proportionate response considers severity, persistence and available remedies. It also accounts for the practical consequences of stopping, including how participants and committed resources will be protected.
Record disagreements, conflicts and negotiated conditions
Committee members may value different dimensions. Financial analysis may regard reliance on one donor as excessive, while impact analysis may consider the underserved population a reason to accept that exposure. Record the disagreement, its factual basis, the response and its effect on the recommendation. Artificial consensus is unnecessary. An approval accompanied by a reasoned dissent can be more transparent than unanimous minutes that remove material objections. The record should show which concern remained and why those approving considered the accepted risk defensible.
Conflicts of interest need concrete identification and treatment. An evaluator may have ties to the applicant, a supplier may belong to a director, or an intermediary may be paid only if funding proceeds. These relationships do not automatically establish misconduct, but they affect incentives and independence. Record the relationship, any voting restriction, independent review and checks on agreed terms. A generic declaration of no conflict has little value if nobody examined the relationships capable of influencing selection, pricing or approval.
Every negotiated condition should address an identified problem at a reasonable compliance cost. Detailed monthly cash reporting may be useful, while monthly outcome reporting may be excessive when change takes months. Linking an installment solely to hiring can encourage opportunistic participant selection. Reviews in this example could combine reconciled financial execution, capacity to continue and functioning protection controls. Social targets guide assessment without necessarily becoming automatic payment triggers for outcomes outside the organization's full control. Conditions should also identify who verifies compliance and what evidence is sufficient.
Assign responsibilities and preserve the decision trail
Distinguish who recommends, approves, signs, releases funds and verifies their use. Within the implementing organization, one person may coordinate activities while another checks payments and reconciles cash. On the funder's side, monitoring may be handled separately from opportunity origination. Complete separation may be impractical in a small organization. Proportionate independent reviews and recorded approvals can help compensate for concentrated duties, although they do not eliminate the underlying risk. The memorandum should describe the actual arrangements rather than an organizational chart nobody follows.
The final record includes the version considered, supporting documents, decision, conditions, responsible people and relative follow-up deadlines. When a condition is met, record the evidence examined and who accepted it. When a condition changes, preserve its earlier wording and explain the revision. This trail helps distinguish new information, analytical error, inadequate execution and a conscious choice to accept risk. It also prevents later reviewers from treating a revised requirement as though it had governed the original decision.
Digital tools can organize permissions, versions and references, but cannot make entered information true. An intact file may contain false information. Blockchain integrity likewise does not establish impact, legal title, reserve solvency or future payment. Participants' private documents should not be placed on public blockchains. Access controls, suitable retention and the ability to correct records matter more than adopting technology for an appearance of security. Any integrity record should support the evidence process without being presented as a substitute for verification of the underlying facts.
Add a separate economic analysis when financial returns are expected
The worked donation creates no payment stream to the donor. A return-seeking opportunity requires additional analysis of the instrument and the debtor's or business's economic capacity. Revenue, margins, working capital, existing obligations, payment priority and loss scenarios become part of the decision. Social purpose cannot replace this work. A business may create meaningful impact while lacking enough cash to service debt. The memorandum should explain both the intended social contribution and the financial mechanism, including where either could fail.
In a separate hypothetical example, a business receives a R$ 100,000 loan and undertakes contractual payments totaling R$ 110,000. If projected cash available to service that obligation is R$ 125,000, the arithmetic cushion is R$ 15,000. If available cash falls to R$ 90,000, the shortfall is R$ 20,000. Timing still matters: a sufficient annual total can conceal shortages before intermediate due dates. These figures illustrate arithmetic, not a market rate or recommended structure, and the forecast does not guarantee repayment or recovery.
Equity does not create an equivalent contractual payment simply because a projection shows increased value. Returns depend on ownership rights and opportunities for distributions or exit, which may be unavailable. Convertible instruments require examination of conversion conditions and their effects. Public securities distribution and regulated financial activities require appropriate structures and providers in Brazil. Technology or a CNAE economic activity code does not confer permission. Banco Central and CVM materials are reference entry points for assessing classification, not evidence that KEZS is authorized or a specific transaction is valid.
Close with an actionable decision and verifiable learning
A practical conclusion might recommend a donation of up to R$ 240,000 in the three proposed installments, subject to formalizing the agreement, confirming signing authority and implementing the specified protection and information controls. Record 120 admissions, 96 completions and 60 employment outcomes as planning assumptions, preserve the cash schedule and identify who may authorize budget reallocations. Approval, disbursement, execution and results are separate events requiring separate records. Approval alone guarantees none of them and should not be reported as delivery.
Monitoring compares events with the reasoning that justified funding. If a report lists 98 completions but includes two duplicate participant records, correct the total to 96, explain the cause, review the affected dataset and communicate the revised version. Silently changing a chart is insufficient. Similarly, if a counterfactual assumption proves weak, update the interpretation while retaining the earlier reasoning. Useful learning identifies the control or premise to change next time, rather than merely explaining away an unfavorable number.
KEZS connects entrepreneurial projects and potential funders through technology and intermediation. Investors conduct their own diligence and make their own decisions; KEZS does not assign a credit score. Project-focused analysis and transparent documentation can help compare objectives, instruments and evidence without replacing independent assessment or guaranteeing approval, repayment, profitability, liquidity or capital preservation. At the close, the investor should understand what is being funded, the rights received, the losses accepted and the response to deviations. Those answers should remain accessible to whoever oversees implementation.
Frequently asked questions.
Can a memorandum recommend a donation within an impact portfolio?
Yes, provided it clearly identifies the decision as a donation. A donation has its own purpose and conditions and should not be treated as an asset carrying promised financial returns. Explain why the absence of payment to the donor fits this contribution's objective, how use of funds will be monitored and what responses are available if conditions are breached. A strong impact assessment does not turn a nonprofit association into an equity company or create ownership rights. This distinction concerns the nature of the transaction, rather than any recommendation about personal asset allocation.
How do a target, an observed outcome and additional impact differ?
A target is an expectation used to guide delivery. An observed outcome is what was measured under established criteria, such as 60 participants in sustained employment. Additional impact is the change attributable to the intervention compared with what would have happened without it. That conclusion requires further evidence and assumptions. Keep the three categories separate so an approved target is not presented as delivery and an observed result is not presented as proven causality. In the example, changing the assumed outcome without intervention changes estimated additionality even when observed employment remains the same.
Should an investor require every beneficiary's individual records?
Information requests should be proportionate to monitoring purposes and applicable data-processing requirements. Many assessments can use aggregate information and restricted sample-based verification. Requesting complete identity documents or sensitive personal accounts merely to create a stronger feeling of control can introduce unnecessary risk. Specify the necessary data, authorized users, purpose, storage arrangements and disposal process. An audit need does not authorize indiscriminate disclosure, and small aggregated groups may still be identifiable. The memorandum should address that possibility when deciding what the funder receives.
Should missing a target automatically block the next installment?
Not necessarily. Examine the cause, severity, the organization's control over the outcome and the consequences for participants. Delayed referrals may require operational adjustment; deliberate false reporting or serious harm may require suspension and investigation. Disbursement conditions should be established in the agreement and linked to concrete risks. In the example, withholding funds without considering cash availability could interrupt services before staff can address the problem. A review should distinguish a justified protective pause from a condition that inadvertently makes recovery harder, while never assuming that additional funding will be available.
How should projects with very different costs per beneficiary be compared?
First check whether services and denominators are comparable. Brief admission support, intensive follow-up and a sustained outcome are different deliveries. Then examine participant circumstances, service intensity, duration, protection and evidence of additionality. A lower cost can reflect efficiency, but it can also reflect selecting people with fewer barriers or omitting necessary expenses. Present these differences before concluding that one proposal uses resources better. Comparing planned costs with another project's verified actual costs also requires explicit qualification; superficially similar ratios may rest on different evidence.
What should be recorded if the committee approves despite a major objection?
Record the objection, supporting evidence, response considered and reason the majority accepted the risk. If approval depends on mitigation, assign a responsible person and verification criterion. Disagreement need not disappear for a decision to be well governed. It can guide monitoring: concern about financial dependence, for example, supports checking future commitments and continuity. The record should later show whether the concern materialized and how the organization responded. Preserving a dissent also helps distinguish a consciously accepted uncertainty from a problem the committee failed to examine.
Which rights need explanation when financial returns are expected?
Identify the instrument, obligated party, expected source of payments, due dates or relevant economic events, priority of rights and consequences of default or loss. For equity, explain governance, dilution and exit possibilities; for convertibles, explain conversion conditions and effects. Any agreed security or guarantee also requires analysis of validity and enforceability. Receivables analysis must distinguish existing eligible claims from hoped-for sales. These arrangements create different rights and risks, and none assures recovery, profitability, liquidity or preservation of capital. Social benefit does not remove the need to examine the financial obligation.
Can a well-organized digital record replace legal and financial review?
No. Document organization helps locate evidence, preserve versions and track conditions, but does not establish signing authority, legal classification or repayment capacity. Those matters require analysis appropriate to the transaction. Public guidance from the CVM, Banco Central, Portal do Empreendedor and ANPD can help identify relevant questions and rules. These sources are reference entry points, not proof of KEZS authorization, an applicant's compliance or a contribution's legal suitability. Likewise, a digital integrity check can show that a record remained unchanged without establishing that its original contents were accurate.
References and context
Institutional reference points for checking applicable rules and guidance. Calculations and scenarios are original examples, not quotes, historical results or individualized recommendations.
- Banco Central — correspondentes no País ↗
- CVM — Resolução 88, texto e alterações ↗
- CVM — Resolução 175, fundos de investimento ↗
- Portal do Empreendedor — perguntas sobre MEI ↗
- ANPD — orientações e perguntas frequentes ↗
AI-assisted educational content. Decisions require verification of the information, instruments and providers in each actual case.
