Community Action Agency Software: One Household, Five Funding Streams, One Record

Why one household turns into five separate files

It is the second week of November and the LIHEAP application season just opened. There is a line at the intake desk that starts before the building does. A family of four sits down with a folder: four weeks of pay stubs, a shut off notice from the gas utility, a Social Security award letter for grandma who lives in the back bedroom, and a lease. Your intake worker opens CAP60 for the CSBG file. Then a separate state LIHEAP portal in another browser tab, because the state runs its own system and will not take an upload. Then a spreadsheet called WX_WAITLIST_2026 for weatherization, because that family also has a furnace that cycles every eleven minutes. Then a paper form for the food pantry, because the pantry has always used paper.

The same pay stubs get keyed four times against four different income tests measured over different periods. The household composition gets described four different ways, because CSBG counts grandma, the LIHEAP application counts the heating account holder, and weatherization counts the dwelling unit. Nothing reconciles. In June, when the CSBG Annual Report is due and the state monitor schedules a visit, your compliance officer builds the numbers by exporting all four and matching on last name and date of birth. She finds 30 to 60 households that appear twice, some that appear once but were served three times, and a handful whose eligibility documents cannot be located at all.

Across the community action projects Digital Heroes has delivered, the recurring cost is 12 to 20 hours a week of duplicate data entry across the agency, plus two to four weeks of a senior person's year lost to report reconciliation. The harder cost is the finding letter. Agencies do not lose funding over one duplicate. They lose it over a pattern that says the file cannot be trusted.

Problem 1: every program counts income differently, and the difference is the whole job

CSBG eligibility is commonly set at 125 percent of the federal poverty guidelines, with states permitted to raise it. LIHEAP commonly runs to 150 percent of the poverty guidelines or 60 percent of state median income, whichever your state plan chose. Weatherization commonly runs to 200 percent of the poverty guidelines, with categorical eligibility if the household already received LIHEAP or SSI. Confirm your own numbers against your state plan, because they move and they are not the point. The point is that three programs measure three different things over three different windows: last 30 days annualized for one, prior 12 months for another, and a point in time snapshot for a third. Self employment gets treated differently in each. Zero income households need an attestation in one and a third party verification in another.

CAP60 knows CSBG well and models it properly. It was built for community action and that shows. What it does not do is hold one household object that satisfies four eligibility engines at once from a single set of documents. CharityTracker is a referral and service tracking tool, and a good one for network coordination, but it does not calculate a benefit. So the workaround is your intake worker, doing arithmetic in her head against a laminated income chart taped to the monitor.

What a custom build does: model the household once, with people, income sources, addresses and documents as first class objects with effective dates. Then write each program's eligibility as a versioned rule set tied to a program year, so the 2026 rules do not overwrite the 2025 rules that a monitor will ask about next spring. Intake runs every rule set at once and returns a screen that says what this family qualifies for today, what they would qualify for with one more document, and exactly which document.

Problem 2: LIHEAP is a payment system pretending to be a case file

Regular LIHEAP benefits are usually a matrix lookup: income band by household size by heating fuel type by dwelling type, sometimes with a vulnerability bump for elderly, disabled or under five. Crisis assistance is a different animal, with federal expectations that a household in crisis gets resolution within 48 hours, and 18 hours where the situation is life threatening. That is not a case note. That is a clock, and somebody needs to be watching it on a Friday afternoon in January.

Then the money has to move. Benefits are paid to vendors, not clients, which means utility account numbers must validate before the file goes out, fuel dealers need delivery authorizations, and every payment file has a format that your gas utility, your electric co-op and your propane dealer each define differently. Credits have to be confirmed as posted. When a client moves mid season, the credit has to be recovered or transferred. Generic case management tools have no concept of a vendor payment file at all, so this ends up in a spreadsheet emailed to the fiscal office.

What a custom build does: treat a benefit as an obligation that becomes a payment that becomes a confirmed credit, each with its own status and audit trail. Vendor agreements live in the system with their file formats. Account numbers validate against the vendor's own check digit rules before anything is sent, which alone eliminates most of the returned payments. Crisis cases carry a countdown that escalates to a supervisor before it breaches, not after.

Problem 3: weatherization is a construction business inside a social services agency

Weatherization is where off the shelf case tools stop pretending. A weatherization unit is a job: an energy audit produces a measure list, each measure has a savings to investment ratio that must justify its inclusion, the whole package must come in under the average cost per unit your state allows, health and safety costs sit in their own budget category, and some homes get deferred for mold, knob and tube wiring or a failing roof before any work happens. Then contractors invoice, a quality control inspector signs off, and pre and post blower door readings have to sit in the client file forever.

What a custom build does: the dwelling unit is the object, not the household, because units get re weatherized and households move. Every unit carries audit inputs, an approved measure list, contractor purchase orders, actual costs against the cap in real time, deferral reasons, inspection sign offs and photo evidence. The average cost per unit is a live number the program manager watches, not a June surprise. Photos are worth calling out separately, because a monitor asking for pre and post images of an attic in a home you did 14 months ago is the single most common reason a file fails.

Problem 4: cost allocation is where the audit actually goes wrong

Your intake worker spent Tuesday morning on CSBG, Tuesday afternoon on LIHEAP, and forty minutes on a weatherization referral. Under Uniform Guidance, that time has to land on the right grants with documentation that supports it. Most agencies do this with a monthly percentage estimate that has not been revisited in three years. The single audit finds it eventually.

What a custom build does: because staff are already working inside the system on program tagged activities, time allocation can be derived from actual case activity rather than reconstructed from memory. You still keep a certification step, because auditors want a signature, but the underlying numbers stop being fiction. This is also the feature that finally lets a finance director answer the board's question about what a served household actually costs by program.

Problem 5: the monitor asks for the file, not the number

State and federal monitoring visits sample households and ask for evidence. Not the count. The file. Missing signature pages, an income calculation nobody can reproduce, a ROMA outcome recorded with no supporting note. CSBG Organizational Standards work the same way, as evidence attached to an assertion.

What a custom build does: an append only activity log so nothing can be quietly edited after the fact, document capture at intake rather than at audit time, and a monitoring packet generator that assembles a sampled household's complete file as a single PDF in about a minute. Document extraction is where AI genuinely helps and nowhere else: reading pay stubs, award letters and shut off notices to pull employer, gross amount, pay period, utility account number, arrears balance and service address for the worker to confirm. In our builds that takes document handling from roughly six minutes per household to under two, and it flags a utility bill address that does not match the lease.

What this costs and how long it takes

A first release covering unified household intake, a multi program eligibility engine, document capture and CSBG outcome tracking runs $55,000 to $120,000 and ships in 12 to 16 weeks in our delivery experience. A full platform adding LIHEAP benefit calculation with vendor payment files, weatherization job costing with contractor management, cost allocation, and report generation for CSBG, DOE and state formats runs $140,000 to $350,000 phased across 6 to 12 months.

What drives the number up: the count of distinct funding streams, because each one is its own rule set and its own report. Head Start, if you run it, because it carries an entirely separate data and reporting world and should usually be scoped as its own phase or left alone. Integration with a state LIHEAP portal, if the state offers an interface at all, and many do not, which means the honest answer is a controlled export and a human. Multi county operations with different local rules. And the size of your paper backlog, because migrating five years of client files is a real project with a real cost that nobody puts in the budget.

Build versus buy, and when buying is the right call

Buy, and be happy about it, if you are a single county agency running CSBG plus one or two ancillary programs on a budget under roughly $4M. CAP60 was built for community action agencies, it knows the CSBG report, and paying for it beats building a worse version. CharityTracker is the right tool if your real need is coordinated referral across a network of local providers rather than benefit calculation inside your own walls. There is no prize for building software you did not need.

Build when three or more of these are true. You administer four or more funding streams with different eligibility tests. You deliver weatherization with in house crews or a contractor panel. You operate across multiple counties with different local requirements. Your last monitoring visit produced a finding related to documentation or duplicate service. Your fiscal office reconciles program numbers by hand every quarter. Above that line, the coordination between programs is the actual operation, and no product built for a generic nonprofit models it.

How to choose a developer for community action software

Ask them to explain how they would handle a rule change mid program year. The right answer involves versioned rule sets with effective dates and the ability to reproduce a two year old determination exactly as it was made. The wrong answer is that they will update the code. If they update the code, your monitor cannot verify a past decision and neither can you.

Ask how they handle document retention and access control. A weatherization file and a domestic violence shelter note cannot carry the same permissions, and if you run a Head Start program, its records have their own rules. A developer who proposes one flat staff role has not thought about it.

Ask who owns the code and get it written down before kickoff. You should own the repository, the hosting accounts and the right to hire anyone else. At Digital Heroes the client owns the code from the first commit. An agency dependent on a vendor who holds the repository is one contract dispute away from losing its intake system in November.