Lean Data

An outcome-measurement pattern that uses short, customer-centered surveys and rapid feedback loops to test whether an impact claim matches the lived experience of affected people.

Also known as: customer-centric impact measurement, beneficiary feedback, right-fit evidence, 60 Decibels Lean Data.

The name borrows “lean” from the Lean Startup’s build-measure-learn loop, not from “thin” or “cheap.” Acumen coined it for data that is lean the way a startup is lean: collect the smallest amount that changes a decision, collect it fast, and act on it before the question goes stale. A Lean Data study is not a smaller version of a formal evaluation. It is a different instrument with a different job — testing whether an impact claim holds with the people the claim names, in time to do something about the answer.

Understand This First

• Theory of Change — the assumptions the survey is meant to test.
• IRIS+ Metric Selection — the standardized metric set that Lean Data may supplement with direct-user evidence.
• The Five Dimensions of Impact — the claim frame that turns customer responses into What, Who, How Much, Contribution, and Risk evidence.

Context

The discipline is simple to state: ask affected people a small number of decision-useful questions, collect the answers through low-cost channels, and use the findings quickly enough to change the investment or program. Acumen built it for social enterprises where traditional evaluations were too slow, too expensive, or too disconnected from operating decisions.

In family-office practice, the pattern belongs in the Impact Measurement and Management stack after the office has a Theory of Change. The theory says what should change. Lean Data asks whether customers, workers, patients, tenants, smallholder farmers, or other affected people are actually experiencing that change.

The pattern is especially useful for consumer-facing and service-delivery investments. Off-grid energy, childcare, small-business lending, agricultural inputs, health access, workforce training, disability products, and financial-inclusion businesses often have reachable end users and measurable experience. Lean Data is weaker for infrastructure, liquid public markets, fund-of-funds exposure, and capital-markets instruments where the affected person is too far from the office’s capital to answer a short survey honestly.

Problem

Impact reports often stop at outputs because outputs are easy to count. The office financed 40 clinics, reached 18,000 customers, lent to 600 enterprises, installed 9,000 solar home systems, or trained 1,200 workers. Those numbers may be true. They don’t tell the investment committee whether people are better off.

Traditional evaluation can answer deeper questions, but it often arrives too late for management. A randomized evaluation, quasi-experimental study, or deep ethnographic assessment may be the right tool for a large policy question. It is rarely the right tool for a $3M recoverable grant, a $7M PRI, or a portfolio company that needs to know next quarter whether customers understand a product, whether women are using it differently than men, or whether a repayment schedule is creating stress.

The family-office failure mode is familiar: staff either overbuy rigor they can’t use, or underbuy evidence and accept the manager’s story. Lean Data sits between those poles. It doesn’t pretend to prove every causal claim. It gives the office a fast, disciplined way to hear from affected people before the annual report turns weak evidence into polished language.

Forces

• Rigor versus usefulness. The office needs credible evidence, but evidence that arrives after the decision window has closed won’t change the investment.
• Low cost versus honest sampling. Short phone or SMS surveys are cheaper than field studies, but the sample still has to represent the people the claim names.
• Standardization versus local fit. Repeated questions help comparison, while each enterprise’s theory of change needs questions that fit its product, population, and geography.
• Manager convenience versus respondent dignity. The data plan has to serve investment decisions without extracting time or personal information from people who see no benefit.
• Customer voice versus attribution. A customer’s answer can show experienced change; it doesn’t by itself prove that the office’s capital caused the change.

Solution

Use Lean Data as a right-fit evidence layer for direct-user outcomes, not as a universal proof engine.

Start with the decision. The office should be able to name what it will do differently based on the answers: approve a second tranche, change repayment terms, fund technical assistance, revise a metric, narrow a public claim, or ask the manager to fix an operating issue. If no decision changes, don’t run the survey.

Then translate the theory of change into five to ten questions. Keep the instrument short. Ask about the outcome the memo named, the baseline condition, the user’s alternatives, the depth of change, the parts of the service that helped or failed, and any negative effects the manager may not see. Use standardized questions where they work, especially for comparable indicators such as first-time access, quality of life, customer satisfaction, affordability, and household resilience. Add custom questions when the local theory needs them.

Choose the channel that fits the affected people, not the office’s convenience. SMS can work for short, simple questions where literacy and phone access are high. Interactive voice response can help in lower-literacy settings but limits open-ended answers. Phone calls cost more but allow probing, consent checks, and clarification. In-person or partner-assisted collection may be necessary when phone ownership, language, disability access, or safety concerns make remote collection weak.

Finally, build the feedback loop into governance. A Lean Data survey should feed the investment memo, quarterly review, technical-assistance budget, manager covenant, or family council dashboard. It should also name what it cannot prove. If the survey reaches only current customers, it may miss excluded non-customers. If the survey is self-reported, it may overstate income or health changes. If the office lacks a comparison group, it should describe the evidence as experienced-outcome evidence, not as causal proof.

How It Plays Out

Consider a $900M single-family office with a $120M foundation and a $70M impact sleeve. The foundation is considering a $6M seven-year PRI into a lender that finances childcare centers in three counties. The Theory of Change says longer-tenor capital will let centers add seats, which should help hourly workers keep jobs because care is closer, more affordable, and more reliable.

The manager’s first dashboard reports the easy numbers: 28 centers financed, $18.4M in loans outstanding, 1,940 childcare seats created or preserved, and 64% of seats serving households below the office’s income threshold. The numbers are useful. They still leave the family council’s real question unanswered: are parents and caregivers experiencing the employment stability the PRI was meant to support?

The office funds a $95,000 Lean Data study instead of asking for a full academic evaluation. The survey design has four rules:

The results change the decision. Parents report shorter commutes and fewer missed-work days in two counties, but the third county shows a different pattern: seats were created, yet families still missed shifts because centers could not keep classrooms staffed after 3 p.m. The survey also finds that Spanish-speaking respondents report more billing confusion and lower satisfaction, even though enrollment counts look strong.

The committee approves the second $3M tranche, but it changes the terms. A $250,000 technical-assistance grant shifts from generic reporting support to extended-hours staffing and bilingual billing help. The manager’s next report has to segment results by county and language. The family council’s annual note says the PRI has evidence of improved care access and fewer missed-work days in two counties, with staffing and billing risks still under review. That’s a narrower claim than the first draft. It is also a claim the office can defend.

A failure case is easy to picture. The same office could survey only center directors, report “1,940 children reached,” and call the investment successful. That would miss the parents whose schedules still don’t match the care offered, the families who left because billing was confusing, and the county where the capital solved the building problem but not the workforce problem. Lean Data earns its place when it finds exactly those frictions before the report is written.

Consequences

The benefit is decision speed with a closer connection to affected people. Lean Data lets the office test whether the outcome pathway is working while there is still time to revise terms, technical assistance, manager reporting, or public language. It also gives the family council evidence that is easier to understand than a 70-page evaluation appendix: direct answers from the people the claim names.

The pattern also improves metric discipline. IRIS+ Metric Selection keeps the reporting package comparable. Lean Data fills the gap where standardized operating metrics don’t capture experience, exclusion, depth, or negative effects. A small survey can show whether a product was first-time access, whether affordability improved, whether quality of life changed, or whether a supposed solution created new burdens.

The liabilities are real. Lean Data depends on reachable respondents, careful consent, language access, good sampling, and questions the respondents can answer. It can overrepresent current customers and miss people the product failed to reach. It can also become survey theater if no one changes the investment after the answers arrive.

The mature use is bounded. Use Lean Data where direct-user feedback can change a live decision. Pair it with theory of change, standardized metrics, additionality review, and verification when the claim is material. Then write the report in the same bounded language the evidence can carry.

Related Articles

Sources

• Acumen, Innovations in Impact Measurement, 2015 — the Acumen and Root Capital report that describes Lean Data as mobile-enabled social-performance measurement, including Acumen’s shift from compliance reporting toward decision-centric customer evidence.
• Acumen, Acumen Launches 60 Decibels to Make Lean Data an Impact Measurement Standard for Impact Investing, 2019 — the official spinout announcement naming Sasha Dichter and Tom Adams, the 85,000-customer / 33-country evidence base, and 60 Decibels’ role carrying the methodology forward.
• Innovations for Poverty Action, Goldilocks Toolkit, current access 2026 — the right-fit evidence lineage organized around credible, actionable, responsible, and transportable data systems.
• 60 Decibels, A Simpler Way to Measure Impact, 2019 — the white-paper landing page connecting Lean Data to customer listening, impact benchmarks, and social-performance comparison.

This entry describes a structural pattern and is not legal, tax, or investment advice. Consult qualified counsel and tax advisors licensed in your jurisdiction before adopting any structure described here.