Portfolio intelligence connects each investee or grantee's agreed plan, reporting, risks, and evidence—then rolls up the portfolio without losing the source behind any number.
Portfolio intelligence is a way to keep every investee or grantee's plan, reporting, evidence, risks, and results connected—then view the whole portfolio without losing the source behind any number. A spreadsheet can list organizations and a dashboard can display totals. Portfolio intelligence helps a team understand whether agreed plans are being followed, which results can be compared, where evidence is missing, and what the portfolio can honestly claim.
Watch · 2 min
Portfolio reporting from every source
Every source in, every investor's report out. CRM, diligence, investee surveys, documents and email — one connected record, personalized reporting per recipient.
Key takeaways
Portfolio intelligence connects the evidence from each funded relationship to the decisions made across the whole portfolio. For an impact investor, that relationship may begin in an application or diligence file and continue through the term sheet, impact agreement, quarterly reports, annual review, and exit. For a foundation, it may begin with a grant application and continue through onboarding, progress reports, board reporting, renewal, and close.
The aim is not to force every organization to use identical metrics. It is to make the differences explicit and governed. Each organization can have an agreed plan that fits its work. Shared definitions then make selected results comparable, while source links preserve the context that aggregation usually removes.
When a grant or investment closes, the evidence that justified it often becomes a folder handed to another team. Someone rereads the diligence, writes an impact plan in a document, creates a separate collection form, and later tries to reconnect reported results to the original case for funding.
A better approach carries the evidence forward. Candidate outcomes and indicators can be drafted from the application, diligence memo, baseline, contract, and site notes. Each candidate keeps its source. The fund or foundation and the funded organization then agree on what matters, add what the documents missed, remove what is not useful, and set a realistic cadence.
The image is not a promise that software can design the plan alone. It shows the governance that matters: source, author, definition, reporting frequency, and agreement are visible. The organization being measured has a voice in the plan.
| Question | Tracker or spreadsheet | Portfolio intelligence |
|---|---|---|
| What is each organization expected to report? | A metric list or document link | An agreed plan with definitions, owners, cadence, and source context |
| Who is late or incomplete? | Manual status updates | Cadence and missing evidence surfaced against each agreed plan |
| Can results be compared? | Columns with similar labels | Compatibility checked through shared definitions and mapping |
| What is the portfolio result? | Copied and summed figures | A governed roll-up that distinguishes measured, estimated, and missing data |
| Where did the number come from? | Search the files | Open the portfolio line back to the organization and reporting period |
A data dictionary gives important measures one agreed meaning. It records the outcome or indicator name, definition, unit, calculation, population, time period, source, owner, and any standards mapping. This prevents two organizations from reporting “jobs created” while one counts temporary placements and the other counts sustained employment.
Portfolio teams need both levels. The organization-specific plan protects relevance. The shared dictionary identifies which measures can be rolled up and how. Where a funder requires IRIS+, the Five Dimensions, SDG, GRI, or a custom framework, the same governed evidence can be mapped without silently changing its meaning. See how to map portfolio data to standards and how to build a practical data dictionary.
Suppose a donor asks, “How many roles has the portfolio created for women this year?” A credible answer is not only a total. It shows which organizations are included, which agreed indicators contributed, which periods are covered, which organizations are missing, and what definition of a role was used.
That last point matters. Excluding an organization with an open report should be explicit. Otherwise, a clean-looking total can create false confidence. Portfolio intelligence makes the boundary visible so the reader understands what the number does—and does not—represent.
Illustrative example. An impact fund has thirty investees working in food systems, health, and livelihoods. Their agreements contain different outcomes and reporting schedules. One funder asks how many sustained income opportunities were created for women. The portfolio team should not search thirty PDFs and add every number containing the word “job.”
First, the team identifies the governed definition: paid roles or contracted income opportunities, held for the required period, with gender and reporting dates specified. Next, it maps compatible investee indicators to that definition. It excludes measures that count training attendance or temporary event work, flags investees whose current report is missing, and retains the narrative evidence that explains unexpected changes. The final answer includes the total, contributing organizations, definition, reporting window, excluded records, and source links.
This workflow is developed further in onboarding a portfolio and tracking results, rolling up a grant portfolio, and producing an LP or board impact report.
SROI and other valuation methods can help a portfolio express social value in monetary terms, but the ratio is only as credible as its evidence. Every outcome needs a quantity, financial proxy, source, and rationale. Deadweight, attribution, displacement, drop-off, discounting, and the reporting period must be documented. Sensitivity scenarios should show how the result changes when assumptions change.
Most importantly, measured results must stay separate from benchmarked estimates. If an investee did not report the required evidence, mark it missing. Do not turn the gap into a precise-looking number. The Academy chapter on blended portfolio SROI explains the calculation and its limits.
Portfolio reporting often becomes a race to standardize numbers. Yet interviews, report narratives, participant comments, site notes, and explanations from investees show why results changed and whether the same indicator means the same thing in practice. They can reveal implementation barriers, negative effects, and groups who are not benefiting.
The goal is not to replace quantitative evidence with stories. It is to read both together. A reported increase should stay connected to the definition, calculation, organization, period, and the field evidence that helps a reviewer interpret it.
AI can read inherited files, suggest candidate outcomes, classify evidence, check missing reports, map definitions, and prepare a source-linked response. Those actions save time only when people can inspect them. The team and funded organization must agree on the plan. Portfolio staff decide when to follow up, whether measures are comparable, and what the funder or board is told.
Access controls, data minimization, version history, and human review are essential. Sensitive investee or participant information should not be exposed simply because a portfolio query is convenient. A reliable answer shows its source and boundaries; it does not ask the reader to trust a fluent paragraph.
Choose one portfolio question that matters to a real decision. Take three funded organizations and try to answer it. Write down where definitions differ, where a reporting period is missing, and which figures cannot be opened back to their source. That small test will show whether the problem is collection, definitions, follow-up, or roll-up.
Portfolio intelligence connects each investee or grantee's agreed plan and evidence to a whole-portfolio view while preserving the source behind every result.
Portfolio management software often tracks organizations, capital, tasks, and documents. Portfolio intelligence governs and interprets outcome evidence across those relationships and can complement an existing system.
No. Each organization can keep relevant measures. A shared dictionary identifies which measures are compatible and how they map to portfolio outcomes or standards.
Mark it missing, show which organization and period are affected, follow up at the right cadence, and keep the gap visible in any aggregate answer.
Themes can be analyzed across reports, but quotes and interpretations should remain connected to their sources and should not be presented as counts without a defined method.
A clear definition, compatible contributing records, a stated reporting period, source links, documented exclusions, and visible assumptions or limitations.
Yes, when proxy sources, quantities, adjustments, attribution logic, discounting, and sensitivity assumptions are documented and measured results remain separate from estimates.
Start with one decision question and three organizations. Test whether the answer can be produced with consistent definitions and opened back to evidence.
Continue the course: Onboard a portfolio and track results → Then learn to monitor portfolio risk, ask the whole portfolio a question, and produce an LP or board impact report.
ChatGPT, Claude, and Gemini are fine for a quick test — but not for an answer you'll put in front of a funder or board. When it has to hold up, run it in Sopact Sense.
See portfolio management in practice →
