A portfolio can span education, energy, food and economic inclusion while remaining exposed to one problem: every project receives payment from the same customer. Diversity of social causes does not remove economic concentration. For an impact investor, diversification by sources of risk means identifying events that could interrupt payments, destroy value, prevent social outcomes or harm beneficiaries. Only after identifying those events does it make sense to discuss how many projects or instruments the portfolio should contain.
The unit of diversification is the cause of loss
Start with a causal question: what must work for this project to meet each commitment? A school needs students, staff, premises and receipts. An energy business also depends on equipment, maintenance and contracts. A nonprofit program may rely on recurring donors and organizations that refer participants. Two apparently different projects can share an indispensable supplier, while one project can combine revenue streams with genuinely different dependencies. The useful comparison concerns what keeps each activity functioning.
Keep three outcomes separate: preservation or loss of financial capital, achievement of impact and protection of the people served. They can diverge. Aggressive collection might improve receipts while excluding the intended beneficiaries. A donation might produce meaningful social benefits without returning money to its donor. Diversification makes these relationships visible and helps manage exposure. It does not guarantee returns, liquidity or continued services, and success on one dimension does not establish success on the others.
Identify the legal and economic relationship before comparing projects
For a loan, examine the payment obligation, financial capacity, maturity and recovery options. For a receivables purchase, establish whether the claims exist and are eligible, who owes payment, when it falls due and what disputes could arise. An equity investment exposes the investor to business uncertainty, corporate rights, possible distributions and opportunities to exit. A convertible instrument adds conditions that may change the future relationship. These instruments cannot be understood through a single undifferentiated category labeled investment.
Donations, sponsorships and grants also require separate classifications. A donation to a nonprofit organization does not purchase equity or imply a promised financial return. Sponsorship may involve communication activities or other contractual benefits. A grant may have specific purposes, eligibility requirements and reporting duties. A Brazilian nonprofit does not offer ownership interests in the way a limited liability company, or LTDA, does. If a related company exists, rights in that company must be assessed separately from the nonprofit relationship.
Record who receives the funds, in what capacity, who assumes each obligation and which document establishes the relationship. A Brazilian individual microentrepreneur, or MEI, cannot admit equity partners while remaining a MEI. Public securities offerings and regulated financial activities require appropriate structures and providers; technology or a CNAE business activity classification is not authorization. This screening prevents false diversification through differently named contracts that leave rights unclear and expose the investor to the same underlying economic risk.
Map the dependencies that actually matter
Begin with the source of cash: households, businesses, public bodies, donors or product buyers. Identify payer concentration, collection periods, essential expenses and the need for additional capital. Add geography, infrastructure, critical suppliers, staff and regulatory dependencies. Describe an observable deterioration event for each item. Dependence on electricity is too broad; a prolonged outage that stops a facility with no operating alternative is a scenario that can guide a decision about exposure or continuity.
The social map should identify objectives, beneficiaries, indicators and the conditions required for results. A training program can report many enrollments yet fail if participants cannot complete its activities. Transport, accessibility, childcare and suitable schedules may determine participation. When several projects serve households in the same area, a flood or the closure of a major local employer can simultaneously damage operations and payment capacity. The same event may therefore affect financial performance, social delivery and beneficiary safety.
You do not need dozens of categories to begin. A matrix listing the project, loss event, exposure, available evidence, existing control and responsible person can reveal important weaknesses. Mark unknowns explicitly. If the information cannot distinguish risks, acknowledge the uncertainty and investigate instead of assigning precise-looking ratings. Each dependency should support a practical choice, such as reducing concentration, changing a maturity or funding a continuity option. A detailed map has little value if nobody uses it to revise decisions.
Investigate correlation through mechanisms as well as history
Small projects often have short histories, incomplete records and frequent operating changes. Statistical correlations calculated from a few months can create a misleading impression of independence. Begin with shared mechanisms: who pays, who supplies, where activities take place and what sustains demand. If three initiatives await transfers from the same Brazilian municipality, an absence of previous delays does not remove the possibility of a joint shock. Historical records help, but they do not replace a causal explanation.
Look for indirect exposures too. A business may sell to different companies that all depend on the same industry. A nonprofit may have many individual donors recruited almost entirely through one channel controlled by someone else. Two equipment suppliers may use the same imported component. Counting contracts can suggest dispersion while the source of vulnerability remains concentrated. Ask why each revenue stream would continue under different scenarios, and whether apparently separate counterparties would actually respond to the same pressure.
Place-based impact creates a particular trade-off. Concentrating effort in one community can improve coordination, trust and collective outcomes while increasing exposure to local events. The response need not be to leave that community. It could involve keeping resources available, developing alternative suppliers and diversifying other portfolio risks. Concentration can be a deliberate impact choice when its rationale, response capacity and financial consequences are explicit. Its social purpose does not make the resulting economic exposure disappear.
A numerical example: four relationships within one budget
Consider an entirely hypothetical example, with no market rates or personal allocation recommendation. An investor commits R$ 1 million to a portfolio initiative. Project A, a vocational education company, receives a R$ 230,000 loan with a single R$ 253,000 payment in month 18. Project B, an energy services company, sells existing receivables assessed as eligible for R$ 270,000. Their expected collections are twelve monthly installments of R$ 24,750 from month 4 through month 15, totaling R$ 297,000.
Project C, a company supporting economic inclusion, receives R$ 250,000 in exchange for equity. No distribution or sale of that interest is assumed during the first 24 months. Project D is a nonprofit running an accessibility program. It receives a committed R$ 200,000 donation in four R$ 50,000 installments in months 0, 3, 6 and 9, subject to agreed conditions. The remaining R$ 50,000 covers R$ 10,000 of initial costs, R$ 15,000 of monitoring and a R$ 25,000 contingency reserve.
The total is R$ 1 million: R$ 750,000 in relationships involving an expectation of financial return, R$ 200,000 in donations and R$ 50,000 for costs and reserves. Expected receipts from A and B total R$ 550,000. The R$ 50,000 difference over their combined R$ 500,000 outlay is gross and depends on full payment. It is not the return on the whole portfolio. C retains an uncertain value, and the R$ 200,000 donated to D is not principal awaiting recovery.
The example diversifies only if its dependencies differ
Suppose A receives most of its revenue from companies purchasing employee training. B's receivables arise from services already delivered to identified customers. C sells cooperative producers' products to private buyers, while D relies on the donation to deliver accessibility activities. These descriptions suggest different sources, but they are insufficient. Due diligence must establish whether A's customers, B's debtors and C's buyers belong to the same corporate group or depend on the same demand cycle.
Imagine that one corporate group appears across all three businesses. It accounts for an estimated R$ 80,000 of revenue needed for A's repayment, R$ 90,000 of B's purchased claims and a substantial share of C's expected sales. The R$ 80,000 represents estimated operating dependence; the R$ 90,000 is direct receivables exposure. C's exposure depends on effects on margins and business value. Adding them as equivalent losses would be incorrect, although the finding reveals concentration hidden by sector labels.
Possible responses include selecting other eligible debtors in B, reducing exposure to the group, revising maturities or requiring an alternative revenue source. Each has costs and limits. Replacing claims may reduce available volume; developing customers takes time. A guarantee dependent on the same group may add little protection. If a proposed change undermines social delivery or makes the project unviable, renegotiate it. Diversification should improve the structure of risk rather than simply push all uncertainty onto the entrepreneur.
Separate investor additionality from project results
Additionality asks what the investor's funding or involvement enables beyond what would probably happen without it. A socially useful project may already have access to equivalent financing on similar terms. In the example, A's loan might enable an accessible class that its current cash position cannot support. Purchasing B's claims might merely bring forward cash without expanding service capacity. These are different propositions requiring different evidence. A general statement of purpose cannot establish the contribution of either transaction.
Record the likely scenario without the transaction, the expected change, its mechanism and evidence that could contradict the hypothesis. For A, examine the budget, financing alternatives and cost of the additional class. For D, identify which adaptations and services require the donation and which already have funding. Disclose activities that would happen anyway. Contribution may involve scale, quality, earlier delivery or lower access barriers, without claiming that the investor is the sole cause of every observed benefit.
A portfolio can also diversify its impact hypotheses. One relying entirely on rapid expansion becomes vulnerable when expansion fails. Another might combine quality improvements, access for excluded groups and continuity of essential services. This does not justify adding incomparable effects into one total. Track the strength of the additionality hypothesis separately from financial execution. An additional social benefit does not automatically turn a weak contract into a sound investment, just as reliable repayment does not prove additional impact.
Choose metrics that reveal results, distribution and cost
A focused set of indicators should answer four questions: who was reached, what changed, for whom the change worked and what it cost. At A, enrollment measures activity; completion and acquired skills help assess outcomes. Employment outcomes require follow-up and care over attribution. At D, counting completed adaptations can help, but does not establish whether people used the service independently and safely. Select indicators that follow the project's stated objective, rather than the measurements easiest to display.
Define the unit, reference population, observation period, collection method and treatment of withdrawals. If a class starts with 100 participants and 70 finish, reporting only the finishers' performance hides part of the picture. Follow the original 100 where feasible, explain absences and examine relevant groups while protecting identities. Missing data mean neither failure nor success. Report how much is missing and whether missingness may relate to vulnerability, since the least visible participants may face the greatest barriers.
Measurement costs must fit the operation. Lengthy interviews can divert staff from services and burden beneficiaries. The example's R$ 15,000 monitoring budget needs a defined scope: which checks it funds and which questions remain unanswered. A few consistently measured outcomes are more useful than a broad dashboard nobody can verify. Standardize administrative definitions where useful, while retaining measures suited to each purpose. Comparisons should respect differences in populations, delivery conditions and the difficulty of achieving the stated result.
Protect beneficiaries before designing incentives
Contracts can change who receives a benefit. If A's payment depends only on completion, it may favor participants facing fewer barriers and exclude people needing extra support. If C is rewarded solely for sales volume, pressure may fall on producers with weaker bargaining power. Investigate these possibilities before linking disbursements, compensation or success assessments to targets. The risk comes from the incentive itself, even when no party intends harm. A favorable headline metric may conceal a worsening distribution of benefits.
Controls can include transparent eligibility criteria, monitoring of refusals, accessible complaints channels, participant consultation and procedures for dangerous situations. These mechanisms must be usable and must not make services conditional on sharing personal stories. Data collection needs a defined purpose, restricted access and retention suited to that purpose. Sensitive information about health, income or vulnerability should not circulate freely among funders because an initiative is described as impactful. Funding does not create an automatic entitlement to beneficiaries' personal information.
Protection must continue under adverse conditions. Agree how reduced activities will preserve referrals, communication and participant safety. This does not require funding an unviable service indefinitely; it requires planning a responsible interruption. At D, withholding an installment may affect already announced activities. At A, collection must not become a pretext for exposing students. Discuss financial rights alongside protection duties, identifying who will act and what resources will support the response. An unfunded continuity promise may offer little practical protection.
The cash calendar reveals risks hidden by the budget
Return to the hypothetical R$ 1 million. In month 0, the investor pays R$ 230,000 to A, R$ 270,000 to B, R$ 250,000 to C, D's first R$ 50,000 installment and R$ 10,000 in initial costs. Total outflow is R$ 810,000, leaving R$ 190,000 in cash. That balance already serves three purposes: R$ 150,000 for D's remaining installments, R$ 15,000 for monitoring and R$ 25,000 for contingencies. Money in the account is therefore not necessarily money available for a new commitment.
With no losses or delays, B generates R$ 297,000 between months 4 and 15, and A pays R$ 253,000 in month 18. After all donation installments and the R$ 25,000 of planned initial and monitoring costs, closing cash at month 24 would be R$ 575,000, assuming no reinvestment, taxes or other movements. This consists of R$ 550,000 received from A and B plus the unused R$ 25,000 reserve. C's equity interest remains outside cash, with no assumed liquidity.
The calendar explains why future receipts should not fund earlier commitments without an explicit bridge. D's installments are reserved from the outset and do not depend on B paying punctually. Investing that reserved money elsewhere would create a liquidity need through the investor's own decision. The reserve is also limited protection, not a guarantee: it can absorb only so much. Record mandatory outflows, conditional outflows and uncertain inflows separately, then update their timing and availability when facts change.
Treat joint losses and delays as different problems
In a hypothetical adverse scenario, A fails to make its R$ 253,000 payment and only R$ 138,000 is recovered in month 24. The loss against R$ 230,000 of principal is R$ 92,000, alongside R$ 23,000 of expected remuneration that never arrives. B's cumulative receipts fall from R$ 297,000 to R$ 198,000 by month 24. Its loss against the R$ 270,000 outlay is R$ 72,000, with another R$ 27,000 of expected receipts also missing. Principal loss and foregone remuneration are distinct components.
Combined receipts from A and B fall from R$ 550,000 to R$ 336,000, a R$ 214,000 difference. With other spending unchanged and the reserve intact, closing cash falls from R$ 575,000 to R$ 361,000. Spending another R$ 20,000 of contingency money on the response leaves R$ 341,000. Suppose C's equity is also estimated at R$ 125,000 against its R$ 250,000 cost. That estimate is neither cash received nor proof of a buyer. D's R$ 200,000 donation still carries no financial repayment obligation.
A delay requires a different interpretation. If B pays the same R$ 297,000 six months later, it has not produced the same nominal loss, but liquidity needs, exposure duration and uncertainty have increased. Delayed payments can also harm suppliers and beneficiaries. Test individual and joint shocks, explaining the mechanisms that could produce them. Applying an identical arbitrary reduction to every project is easy, but reveals little about which controls would help or when cash would actually become available.
Require evidence suited to the right and the hypothesis
For B, investigate documents supporting each underlying transaction, debtor identity, payment terms, disputes and signs of previous assignment or encumbrance. Hoped-for sales are not existing receivables. A commercial agreement may anticipate future orders without immediately creating the claim presented to an investor. Distinguish expected revenue, completed services, invoicing and an enforceable right, using professional support where needed. The distinction matters because buying a claim against work already performed involves different risks from financing sales that may never occur.
For A, statements and financial records help reconstruct inflows, expenses and commitments, but a balance snapshot does not establish repayment capacity in month 18. For C, projections must address margins, capital needs, governance and exit conditions. For D, the budget, team experience and service evidence support the donation assessment. Evidence is only as strong as its fit to the question: a testimonial may reveal an access barrier without establishing the average outcome of an entire program.
Classify information as documented fact, unconfirmed statement, estimate or hypothesis. Record its source, responsible person and limitations, including inconclusive documentation. Technology may improve traceability without fixing content quality. A blockchain record can preserve the integrity of an entry without proving truth, legal title, reserve solvency or payment certainty. Private documents and beneficiary data should not be placed on public blockchains to create an appearance of transparency. Integrity of a record does not validate the underlying claim or make confidential disclosure appropriate.
Negotiate protection the project can actually support
Contract terms should address the identified risk. For A, this may mean periodic information, limits on new obligations and notice of lost material contracts. For B, define eligibility, claim verification and dispute handling without assuming someone else covers every default. For C, consider information rights, reserved decisions, conflicts and exit rules. For D, specify permitted use, reporting and installment conditions while preserving the donation's nature. A contractual label should match the rights actually created, rather than imply protection the terms do not provide.
A common disagreement arises when an investor wants to stop funding after a missed target while the team attributes the shortfall to a change in participants. Separate execution failure, an unsuitable hypothesis and a legitimate contextual change. Responses may include a correction period, revised measurement or adjusted activities, documenting what changed and why. Renegotiation should preserve the original record. It must not erase the commitment or allow every unfavorable outcome to be relabeled as success simply because the parties prefer that interpretation.
Excessive protection has costs too. Weekly reports consume staff time; guarantees can tie up essential assets; short maturities can force damaging decisions. Agree what will be monitored, how often and with what consequences. Assess any reduction in investor risk alongside the risk created for the operation and its beneficiaries. Where acceptable conditions cannot support both sides, declining or reducing the transaction may be preferable to signing an unworkable contract. Formal rights are useful only within a realistic operating arrangement.
Assign responsibilities and address conflicts of interest
Monitoring needs named responsibilities. The project team produces operating information and reports deviations; the finance function reconciles cash and obligations; impact analysis checks indicator definitions and quality; the investor decides its exposure. Legal support examines rights and limits when needed. One person may perform several roles in a small organization, but that overlap should be visible. Verification of material information should not depend solely on the person who benefits from presenting it favorably. Responsibility for producing evidence differs from independent review.
Intermediaries may receive fees for origination, structuring, monitoring or other contracted work. Each arrangement creates incentives. Payment on completion may encourage faster decisions; assessing impact while selling consulting to the assessed organization may create another conflict. Identify who pays, how compensation works, which services are delivered and how parties are connected. Disclosure does not remove the conflict. It helps determine whether independent review, limits on involvement or exclusion from a decision is necessary, and makes the cost of intermediation easier to assess.
KEZS connects entrepreneurial projects and potential funders through technology and intermediation. A project-oriented approach can organize objectives, beneficiaries, documents and risk questions, while investors remain responsible for their own diligence and decisions. KEZS does not assign a credit score. The usefulness of organized information depends on its quality, freshness and verifiability, and on the services actually contracted. A structured presentation does not demonstrate approval, repayment, profitability, liquidity, protection of capital or authorization to conduct a regulated activity.
Correct classification errors before adding projects
One frequent mistake is calling a portfolio diversified because it contains many names but few ultimate payers. Reconstruct corporate groups and shared dependencies first. Another is presenting A and B's expected R$ 550,000 of receipts as the full return on the R$ 1 million budget. Separate returned principal, remuneration, donation, costs, reserve and illiquid equity. Without that breakdown, an attractive presentation can obscure the economic destination of each amount and the capital that remains exposed or is never intended to return.
Do not confuse a new funding round at C with a lasting ability to generate operating cash. Another investor's contribution may sustain operations temporarily and dilute the existing holding, depending on terms. It does not establish profitability or an available exit. On the social side, adding service counts across projects serving the same people can double-count outcomes. Use clear definitions and proportionate ways to identify overlap, without circulating complete personal databases between organizations merely for analytical convenience.
Another group of errors involves reacting too late. A portfolio may retain its original assumptions after a customer is lost, the team changes or the priority population shifts. Keep a decision record with the observed signal, expected consequence, responsible person and review deadline. A previous decision can have been reasonable on the available evidence and still require revision. Adding projects without correcting these weaknesses increases workload and cost while potentially preserving the very concentration that diversification was meant to address.
Turn the risk map into a decision discipline
Before assuming a new exposure, write a short conclusion that another responsible person can review. State the proposed relationship, the source of payment where applicable, the expected social change, who could be harmed and which dependencies already exist elsewhere in the portfolio. Add major uncertainties and conditions that would change the decision. Explain why the project serves the strategy and why the combined exposure remains manageable. The importance of the cause should not hide concentration or substitute for an assessment of financial capacity.
During execution, watch for events that change the thesis: loss of a payer, persistently higher costs, a different participant population, a beneficiary incident or less reliable data. Define which events require information, review or exercise of contractual rights. A private portfolio may offer no practical opportunity to sell positions for rebalancing. Adjustments may instead involve new disbursements, reserves, renegotiation or a decision not to increase exposure. Each response has constraints, including consent requirements, available cash and commitments already made.
At each evaluation cycle, compare expectations with reality without merging dimensions. What cash entered and left? What capital remains exposed? Which social outcomes were observed, with what limitations? Was additionality plausible and protection adequate? The example shows why one rate cannot answer everything. Diversifying by sources of risk means keeping decisions explainable and revisable in light of evidence, costs, conflicts and possible loss, while preserving the impact purpose without assuming that purpose resolves financial risk.
Frequently asked questions.
How many projects make a portfolio diversified?
No project count resolves the question by itself. Ten projects may depend on one payer, while a smaller group may have different cash sources and operating conditions. Assess concentration, exposure size, shared dependencies and monitoring capacity. Adding a transaction improves diversification only when it changes meaningful risks without creating unmanageable costs or complexity. Different sector labels are insufficient evidence. The question is whether a harmful event would affect several positions together, how severely it could affect them and what response resources are available.
Does a donation reduce the risk of an investment in another project?
Not automatically. A donation neither covers default nor creates rights over the recipient nonprofit's assets. It may finance a complementary activity and, in some structures, help remove an identified operating barrier. Demonstrate that relationship, document it where appropriate and assess conflicts and possible diversion from the donation's purpose. In the example, D's donation remains separate from rights in A, B and C despite sharing one impact budget. Any claimed protective effect requires its own explanation; proximity between activities does not establish financial protection.
Can a strong social outcome compensate for a financial loss?
That depends on the mandate and criteria established before the decision. An investor may accept a financial concession to pursue an impact objective, but should state that intention and the capacity to absorb losses. An unexpected financial loss should not be relabeled as success simply because people received services. Report financial performance and impact evidence separately, then assess the original commitment. This preserves an honest account of any trade-off and avoids retrospectively changing the objective to match whatever result happened.
How can projects with very different impact indicators be compared?
First compare the quality of the theory and evidence: clear objectives, intended population, mechanism of change, additionality and protection. Then assess outcomes in each project's context. Course completion and an accessibility adaptation are not interchangeable units. Cost per outcome can help compare similar activities when quality and delivery difficulty are considered. A single score can conceal value judgments and uncertainty. Portfolio decisions may still compare priorities, but should make those judgments visible instead of presenting unlike outcomes as an objective numerical ranking.
What if the project team considers data requirements excessive?
Identify the decision each item supports and the effort needed to produce it. Remove unused fields, use reliable operating records and agree a frequency proportionate to risk. If an important question remains unanswered, record the limitation and adjust the investment decision rather than demand a purely formal report. Personal data need additional care: financing does not automatically entitle a funder to detailed beneficiary information. Consider whether aggregated or less intrusive evidence can answer the question while preserving the ability to investigate material concerns.
Do existing receivables make a transaction safe?
Their existence is only a starting condition. Examine validity, ownership, eligibility, debtor identity, maturity, disputes and previous assignments or commitments. A debtor may pay late or default, and recovery can be costly. Understand who handles collection and how investors obtain information. Multiple debtors can remain correlated through the same economic activity or infrastructure. Existing eligible receivables must also be distinguished from hoped-for sales: a business forecast is not an established payment claim, however plausible its projected revenue may appear.
How should an opportunity presented by a connection platform be assessed?
Use the presentation to organize questions and locate documents. Establish who supplied each statement and what was actually checked. Examine the instrument, responsible parties, intermediary compensation and the structure required for the transaction. Platform presence does not replace diligence or guarantee an outcome. KEZS provides a technological and intermediation connection between entrepreneurial projects and potential funders; it does not assign credit scores or make the investor's decision. Evaluate the relevant rights, evidence and contracted services rather than treating organized information as an endorsement.
When can deliberate concentration make sense for an impact strategy?
A strategy focused on one place or problem may need concentration to achieve coordination and depth. Make clear which risks increase, what resources support a response to shocks and which beneficiary commitments need protection. Payers, suppliers or maturities may be diversified without abandoning the chosen purpose. The rationale weakens when concentration is simply unknown or relies on liquidity and additional capital nobody has committed. Deliberate concentration requires an explicit trade-off, realistic response capacity and a willingness to revise the decision when its assumptions no longer hold.
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.
