Irrational Labs / B2B Home Care Research

The referral is just the beginning

A Place for Mom connects families to home care every day. But the moment a referral is made, the context disappears. This research asks: where in the post-referral journey should APFM show up next?

Scroll to explore

Executive summary

What this research found

Scheduling and hiring are pressure points on the same system
Demand is volatile, so agencies can't guarantee stable hours. Caregivers hedge by working for multiple agencies. Agencies overhire to have coverage. The system is inefficient but stable: each side's rational behavior reinforces the other's. The information that could break this cycle doesn't travel.
Three opportunity spaces
Referral-to-care conversion: increasing demand. Better referrals that convert into clients, supported through the full journey to a scheduled home visit. The caregiver layer: reducing overhead. A shared bench where caregiver context travels across agencies. The family relationship arc: longer-term. Giving families visibility across providers. Depends on building the relationship through the referral journey first.
APFM sits at the intersection
APFM is the demand signal (referrals) and has a large network of agencies from which caregiver context could travel. Individual agencies can't see across their own walls.

The approach

Three lenses on the same question

We used three distinct research methods to understand the post-referral journey. Each answers a different question. Together, they triangulate where the real opportunities are.

22
In-depth interviews

90-minute conversations with agencies across 13 states. From 4-person startups to 500+ caregiver operations.

How agencies actually operate. Where systems break. 33 workarounds documented across 19 of 22 agencies.

246
Survey responses

Agencies across 38 states. Franchise (60%), independent (35%), corporate (4%). From 10-person shops to 500+ operations.

Which pains are widespread and which are niche. The magnitude at scale.

80+
Tools evaluated

Competitive landscape scan across 15 operational areas. Tool density, market penetration, entrenchment, and adoption patterns.

Where the market has invested, where it hasn't, and what existing tools actually fail to do.

In-depth interviews (n=22)

In 22 interviews, agencies described pain across 15 operational areas. Seven stood out.

Across three tracks, every area has distinct pain points, workarounds, and broken handoffs. But not all of them carry equal weight. Seven areas consistently surfaced with the strongest signal: the biggest operational drains, the deepest structural fragility, or both.

Descriptions show the core pain point in each area. Highlighted areas are the seven biggest drains on agency time.
Client onboarding
Intake calls
Between first call and first shift, families repeat themselves to each new person they talk to.
Home assessment
Context captured on paper gets compressed in re-entry, and the family's situation keeps changing after the visit.
Care plan creation
Supposed to be the source of truth. In practice, a compliance artifact.
Payment setup
Payer source shapes every downstream workflow. Re-entry adds friction.
Caregiver matching
Tools handle the basics. The real matching decisions rely on personal knowledge of each caregiver.
Caregiver management
Hiring
Continuous, regardless of demand. High investment, high drop-off.
CG Onboarding
Same compliance process regardless of experience. Highest drop-off point.
Performance mgmt
Informal everywhere. Quality knowledge lives in schedulers' heads.
Retention
Programs designed inside-out. What actually retains: hours and respect.
Payroll
Mostly functional. EVV compliance is the main friction point.
Care delivery
First shift
Make-or-break moment. Most agencies have no structured feedback from this visit.
Scheduling
Availability does not equal willingness. Call-outs dominate every day.
Care Tracking
Families call for a person. Portals go unused. Texts bypass the agency.
Feedback
Compliance-driven, not quality-driven. Caregiver notes are the real signal.
End of care
Agencies close cases administratively. What was learned about the family doesn't transfer.
19 of 22 agencies built workarounds. See examples
Spreadsheets taped to monitors. Group texts for scheduling. Paper binders that travel with caregivers (in some states, a regulatory requirement, not just a workaround). 33 distinct workarounds to compensate for tools that don't work.

Survey findings (n=246)

What agencies rank as their biggest pains, and what they've stopped noticing

246 agencies confirmed what the interviews surfaced: scheduling and hiring dominate at every size. But the pain ranking alone doesn't tell the full story. Several areas that agencies live with every day rank low, not because they're solved, but because agencies have normalised the fragility.

0%
of agencies are always hiring
0%
lost caregivers in the past 3 months
0%
rank scheduling in their top 3 (highest of all 11 options)
0%
rely on personal knowledge for caregiver matching
What consumes the most time and effort
% selecting in top 3 (n=246)
Top 3 Middle tier Lower tier
Notice what's not at the top
Matching (15%) and retention (17%) rank low. But 81% of agencies rely on personal knowledge for matching, and 74% lost caregivers in the past three months. Agencies don't experience these as operational pain because they've normalised the fragility: matching feels like expertise, not a gap. Turnover feels like the industry, not a solvable problem. The areas agencies rank highest are the ones they feel every day. The areas they've stopped noticing tell a different part of the story.
Why caregivers leave
Top reasons agencies cite for caregiver turnover (n=183, select all that apply)
The top two drivers are hours instability and multi-agency work. Pay rate is fourth at 26%. Hours instability is itself a financial problem: unpredictable hours means unpredictable income. The industry often frames turnover as a rate issue and tries to solve it by raising pay. This data suggests the financial instability of irregular hours is at least as large a factor.
Note: this is agency-reported. We have not yet spoken directly to caregivers. Whether caregivers would describe the same reasons for leaving is an open question and a primary goal of the next research phase.

How agencies differ

Scale predicts agency needs better than franchise vs. independent

APFM currently segments agencies by type: franchise, independent, corporate. When we cut the data by agency size instead, the patterns were consistently clearer. Three distinct profiles showed up.

Small (1-25 CGs)
The capacity trap
Can't grow without caregivers, can't retain caregivers without clients. Owner knows everyone by name. Personal knowledge matching works (71%) because the scale allows it.
#1 painHiring (48%)
#1 churn driverHours instability
Always hiring61%
Mid (26-100 CGs)
Where problems converge
Enough volume to feel every gap. Not enough organizational power to fix root causes. Highest personal matching reliance (86%). Most dissatisfied with their tools (12% very dissatisfied).
#1 painScheduling (61%)
#1 churn driverMulti-agency work
Always hiring83%
Large (101+ CGs)
Structural crisis
Scheduling dominates. Recruiting shifts to geographic coverage (they've given up on matching at hiring). Personal knowledge matching persists (81%) despite more software adoption (16%).
#1 painScheduling (69%)
#1 churn driverPersonal/family
Always hiring93%
Top pain points shift with scale
% selecting in top 3, by agency size
Small: Hiring leads
Mid: Scheduling takes over
Large: Scheduling dominates
Software handles availability and credentials. The match decision stays human.
% of agencies where personal knowledge is involved in matching
71%
Small (1-25)
86%
Mid (26-100)
81%
Large (101+)
Personal knowledge is involved in matching at every size, and it doesn't decrease with scale. The soft-fit decision, personality, communication style, reliability patterns, stays personal whether an agency has 10 caregivers or 300. This works until a coordinator leaves, or the agency outgrows what one person can hold in their head.
Small (1-25)
Mostly personal: 37%
Equal mix: 19%
Entirely personal: 15%
Hire per client: 14%
Mostly software: 5%
Entirely software: 8%
Mid (26-100)
Mostly personal: 37%
Equal mix: 35%
Entirely personal: 14%
Mostly software: 7%
Hire per client: 1%
Entirely software: 1%
Large (101+)
Equal mix: 34%
Mostly personal: 26%
Entirely personal: 21%
Mostly software: 16%
Hire per client: 2%
Entirely software: 2%
The #1 reason caregivers leave shifts with scale
Top reasons for leaving, by agency size (% of agencies reporting)
Small (1-25)
Hours instability
54%
Multi-agency 41%
Personal 30%
Pay 11%
Mid (26-100)
Multi-agency work
49%
Hours 48%
Personal 47%
Pay 26%
Large (101+)
Personal/family
51%
Multi-agency 42%
Hours 42%
Pay 32%
At mid-size agencies, multi-agency work is the #1 reason caregivers leave.
Mid-size agencies are where problems converge.
They have enough volume to feel every gap but not enough organizational power to fix root causes. They're the most dissatisfied with their tools, the most reliant on personal knowledge for matching, and the segment where multi-agency work is the #1 reason caregivers leave.

Competitive landscape (80+ tools)

Where tools exist and where they don't

We evaluated 80+ tools across 15 operational areas, assessing how many vendors exist and how entrenched they are: platform integration, funding, and market penetration.

Served and entrenched
Scheduling Payroll / Billing Retention Feedback
Tools address the core operational need and are deeply embedded. Scheduling has 6+ vendors including Phoebe and Zingage (WellSky-native, 500+ agencies, $12.5M Bessemer). ADP and UKG own payroll. Zingage and Activated Insights are building retention moats. Sensi.AI ($98M+) serves 80%+ of the largest networks for feedback. High switching costs across the board.
Tools exist, miss the core need
Hiring Onboarding Matching Care plans
Entrenched tools solve one dimension, but the pain agencies described is a different problem entirely. Indeed dominates job sourcing (39% cite as #1) but the core problem is onboarding overhead and portable credentials, not pipeline volume. CareAcademy owns compliance training but nobody addresses ghost rates or redundant paperwork. Three ERP matching tools handle credentials and availability, but 81% of agencies rely on personal knowledge for the actual match decision; 10 of 22 bypass their tools entirely. ERP modules (WellSky, AxisCare) handle care plan creation, but the plans are compliance artifacts: daily care runs off task lists and notes. No tool enriches care plans with context from the referral source. The market is crowded for the operational basics but the gap between what tools do and what agencies actually struggle with is wide.
Open but structurally constrained
Intake Home assessment First shift Care Tracking End of care
The gaps exist but entry faces structural barriers. ERPs handle the operational basics (shift logistics, discharge admin, lead capture). The relational and strategic layer on top is largely unaddressed, but building there requires navigating ERP integration, trust deficits, or behavior change. CRMs capture leads but context dies at handoff. WellSky Scribe is deploying for assessment digitization, but the context-at-handoff gap (referral information reaching the assessor) is a different problem. Family portals exist in ERPs, but agencies that mentioned them described low engagement; families prefer to call. The supply-side opportunity is open on the dimension that matters, but the entrenched operational layer creates friction for entry.
Wide open
Performance mgmt
No standardized outcome measures for non-medical care. Care quality lives in schedulers' heads. Wide open, but no clear path to the data makes it hard to enter despite the lack of competition.

Each area contains specific unmet needs, and the entrenchment varies by need. Hiring has dozens of job boards, but the onboarding overhead and portable caregiver context gap is largely unaddressed. The categories above reflect how open the specific gap is, not the overall competitive density of the area.

Methodology: Each operational area contains multiple unmet needs. Some needs are well-served by entrenched tools. Others are wide open. We assessed each area on two dimensions: how entrenched the existing tools are, and how well those tools actually resolve the unmet needs agencies described. A crowded area can have a wide-open gap if the tools are solving the wrong problem. Based on competitive product analysis (80+ tools), vendor-reported penetration data, survey tool adoption (n=246), and interview validation (n=22).

Area Market openness What's served (entrenched) What's open (the gap)
Client onboarding
Intake Open but structurally constrained Lead capture: Moderate adoption (WellSky CRM, HubSpot, Zoho, GoHighLevel). Fragmented, no dominant player. Follow-up cadence: CRMs handle this. Context preservation at handoff: Not addressed by any tool we evaluated. Sales stack doesn't map into ERP ops record. Context captured at intake dies at the handoff. Lead qualification accuracy: Partially open. 8/22 cite APFM lead quality as a trust issue.
Assessment Open but structurally constrained Paper-to-digital: Entrenched. WellSky SkySense Scribe (voice AI, 60%+ installed base). CareSwitch Looper (voice dictation). Technology moat forming. Context at handoff from referral: Not addressed by any tool we evaluated. Different problem from digitization. No tool pre-populates assessment context from the referral source. Assessment transcription: Emerging. WellSky Scribe entering but some clients uncomfortable with recording.
Care plans Tools exist, miss the core need Plan creation tool: ERP modules widely available (WellSky, AxisCare). Moderate lock-in. Agencies print because apps are clunky. Input quality from referral chain: Not addressed by any tool we evaluated. No tool enriches care plans with context from the referral source. Cross-provider context: Not addressed. One agency describes care plans as multi-provider "operating system" but no tool supports this.
Matching Tools exist, miss the core need Hard constraints: Contested. WellSky CareFinder, AxisCare Intelligence, CareQB. Filter on credentials, availability, geography. Soft-fit matching: Not addressed by tools we evaluated. 81% rely on personal knowledge. CareMatch (early-stage) attempts it but depends on coordinator data entry. Cross-agency match outcomes: Not addressed. No tool learns from outcomes across agencies.
Payment setup Served and entrenched ERP billing modules, clearinghouses (Waystar, Change Healthcare). Configuration step within universal tools. No meaningful unmet need. Functional.
Caregiver management
Hiring Tools exist, miss the core need Pipeline volume: Dense. Indeed (39% #1 source, 88% turnover), Hireology (8,000+), myCNAjobs, Apploi. Crowded but no lock-in. Hire-to-retention connection: Not addressed by any tool we evaluated. No tool connects hire source to 90-day retention outcomes. Portable credentials: Not addressed. Background checks reset at every agency. Kare is the closest to cross-agency caregiver context but narrow.
Onboarding Tools exist, miss the core need Compliance training: Entrenched. CareAcademy ($5-12/CG), Nevvon, Relias. Strong franchise penetration. Ghost rate reduction: Not addressed by any tool we evaluated. No tool addresses the drop-off between offer and first shift. Portable onboarding credentials: Not addressed. Redundant paperwork resets at every agency.
Retention Served and entrenched Turnover prediction: Contested. Zingage ($12.5M Bessemer, WellSky-native, 500+ agencies). Activated Insights (top 100). Moats forming. Hours stability: Open but structural (tied to scheduling and client mix, not solvable by a standalone tool). Cross-agency visibility: Not addressed. 83% know CGs work elsewhere but no data flows.
Performance mgmt Wide open Aaniie, CareSmartz360 (negligible share). No dominant player. Care quality assessment: Mostly open. No standardized outcome measures for non-medical care. Broad market not addressed. Family feedback as quality signal: Not addressed. Could arrive as byproduct of a Care Tracking play.
Payroll Served and entrenched ADP, UKG, Workday, Paychex, QuickBooks. Near-universal adoption. High switching costs. No meaningful unmet need. Solved.
Care delivery
First shift Open but structurally constrained ERP handles shift logistics (clock-in/out, task lists). No dedicated tools identified for the introduction or feedback layer. Caregiver-family introduction: Not addressed by any tool we evaluated. 11/22 built their own protocols. Brightwheel validates category in childcare (8M users). First-shift feedback: Not addressed. No mechanism captures whether the match is working.
Scheduling Served and entrenched Shift fill optimization: Entrenched. 6+ funded AI overlays (Phoebe, Zingage/Casey, ShiftMatch.AI). Deep ERP integration. Availability tracking: Emerging (AI overlays). Cross-agency supply exchange: Not addressed, but blocked by trust. Agencies view each other as competitors. Cross-office sharing works within one organization but nobody enables cross-agency.
Care Tracking Open but structurally constrained Family-agency messaging: Entrenched in ERPs and phone/text. WellSky Family Room, AxisCare portal exist but agencies described low engagement; families prefer to call. Cross-provider family visibility: Not addressed by any tool we evaluated. Families work with multiple providers, and no tool gives a view across them. Requires trust repair and ERP integration. Care transition signals: Not addressed. No dedicated tool identified for knowing when to move to senior living.
Feedback Served and entrenched Family satisfaction: Contested. Sensi.AI ($98M+), Activated Insights (500K+ surveys/yr), Zingage pulse surveys. Three funded players. Feedback-to-action loop: Not addressed by any tool we evaluated. Signal doesn't trigger operational change. A low NPS score doesn't change Monday's scheduling.
End of care Open but structurally constrained ERP discharge workflow (billing status change, reason codes). Sensi Solo Care monitors seniors without caregivers (adjacent). Re-referral to senior living: Not addressed by any tool we evaluated. End of care = beginning of new referral. No agency has incentive to flag this. Knowledge transfer: Not addressed. What was learned about the family doesn't travel.

The market built where penalties are immediate: miss a visit, lose revenue. Skip compliance, lose your license. But the areas with the highest relational stakes, the moments that determine whether care actually works for families, have the least supply-side competition. The tools that exist capture data into structured fields. The information that matters most stays in notes, texts, and people's heads.

The system underneath

The biggest pains share a root system.

The underlying dynamic is a supply-demand matching problem, complicated by low trust and fragmented information.

Demand side
Families need care, often urgently. Demand is volatile and unpredictable: timing, location, hours, care type all vary. When there's no signal to match that demand to available supply, referrals fail or get declined. For many agencies, the core challenge isn't finding caregivers. It's having enough of the right clients to keep them.
Supply side
Caregivers want financial stability, but no single agency can guarantee stable hours because demand is volatile. So caregivers work for multiple agencies (83% of agencies know this). Agencies overhire and under-invest: they build a bench for coverage but don't invest deeply because they expect people to leave. When demand doesn't match the bench they built, those caregivers sit idle, drift to other agencies, and the hiring and onboarding overhead is wasted.

Each side's rational response makes the other side's problem worse

78%
always hiring
49%
say CGs leave over hours instability
83%
say CGs work for multiple agencies
74%
lost CGs in past 3 months
Demand is volatile Can't guarantee stable hours CGs work for multiple agencies Agencies overhire for coverage Demand doesn't match → bench drifts away

Demand volatility means agencies can't guarantee stable hours. Caregivers hedge by working for multiple agencies. Agencies respond by overhiring and under-investing because they expect people to leave. When demand doesn't match the bench they built, those caregivers sit idle and drift away. The hiring and onboarding overhead resets every time. The signals that could reduce this friction don't travel. On the demand side: where are families looking for care, what hours, what care types. On the supply side: who is reliable, who works well with which clients, who is actually available. None of it moves. It stays in one scheduler's head at one agency.


The convergence

Where pain, market gap, and APFM advantage overlap

We assessed 15 operational areas on three dimensions: agency pain (from survey and interviews), market openness (how well existing tools resolve the unmet need), and APFM advantage (where APFM can add unique value). The areas where all three converge are where APFM should play.

Where APFM should play

Three opportunity spaces where APFM is well-positioned to win

Each space has direct value for APFM, independent of the others. Together, they address the root system behind the biggest agency pains.

1
Referral-to-care conversion
Increasing demand
8 of 22 agencies described lead quality as eroding trust in APFM. Family context degrades across handoffs and evolves in ways no one captures. The journey from first contact to care start has multiple moments where APFM can add value. Demand is what most agencies need most.
What this could look like: Capturing richer family context so agencies start with a more complete picture, and maintaining that context as family needs evolve. Matching family needs to agency capacity and location so referrals convert at higher rates (demand-capacity matching). Owning more of the intake process so context doesn't degrade at the handoff. Supporting agencies through the home assessment so the lead becomes a client. The endpoint: a scheduled home visit, not a referral.
2
The caregiver layer
Reducing overhead
83% of agencies know caregivers work for multiple agencies, but each agency builds its own bench independently. Hiring is the #2 pain: agencies overhire to have coverage, but no one tracks who's reliable or available across the network.
What this could look like: A shared caregiver pool across APFM's large agency network where context travels with the caregiver. Agencies that participate get access to caregivers who aren't starting from zero. Over time, that context accumulates (reliability patterns, cross-agency feedback) and the pool becomes more useful for everyone. The exact shape of this, what agencies would share, what caregivers would opt into, needs more research.
3
The family relationship arc
Making care quality visible · longer-term
Families often work with multiple providers simultaneously. No information is shared between them. When care ends or an agency changes, the family starts over. APFM sees the family before any single agency relationship, but today that contact ends at the referral.
What this could look like: Helping families stay informed across providers. Episodic touchpoints at care transitions. Reducing duplicate intake. This depends on building a deeper family relationship through the referral-to-care journey first. Requires ERP integration and a level of trust APFM hasn't established yet.

Deep dives

The evidence behind each opportunity

Each opportunity space is anchored in converging evidence: agency stories, survey data, competitive gaps, and behavioral mechanisms.

Referral-to-care conversion
The family's story starts incomplete and keeps evolving. Give agencies a better starting picture, and build the thread that keeps it current.
8 of 22 agencies
described lead quality as eroding their trust in APFM
Context starts incomplete
then degrades across handoffs and keeps evolving after care starts
No dedicated tools we evaluated
support the post-referral journey from intake to first shift

Some families enter the process under-informed and not yet "decision-ready"

Holly has been coordinating clients at a franchise agency in rural Maryland for six years. She describes a recent caller: a woman who hadn't slept in over a week, whose voice was breaking on the phone. By the time that woman's information reaches Holly, it has already been sent to three other agencies. By the third call, the woman is frustrated. By the fourth, hostile. Holly starts from zero, asking questions the woman already answered, to someone who no longer wants to talk.

0:15
Holly, Assisting Hands, Maryland

Holly is not unusual.

8 of 22 agencies described lead quality as eroding their trust in APFM. Families fill out one form and receive four or more calls within minutes. The agency that calls first wins, not the best-matched one. About half of agencies absorb unpaid family education: what home care is, what it costs, what Medicare covers. None of this is tracked.

In Chicagoland, Jeremy spent $4,200 on APFM leads when he was starting out. One converted. His trust in APFM is "fully broken."

Lead quality is a known problem with work underway. But the post-referral journey, what happens between the lead arriving and care actually starting, is what this research surfaced.

The core problem
Holly starts from zero, asking questions the woman already answered, to someone who no longer wants to talk.
22 agencies interviewed
8 described lead quality as eroding their trust in APFM
The economics
$4,200 on APFM leads when starting out. One converted. Trust is "fully broken."
246 survey responses
40% mentioned lead quality unprompted, the #1 unasked pain

The home assessment is doing too many jobs at once

The home assessment juggles data capture, rapport-building, expectation setting, and scope-of-care education all at once. Agencies choose paper over screens deliberately. Paper signals active listening and presence. Agencies choose paper because it builds trust, not because they lack technology.

From the interviews
"We do everything on paper when we're in the house. That is how they feel most comfortable and it helps us build trust."
Emily, Home Instead, Northern Indiana (~280 CGs)
0:41

The cost: 10-60+ minutes of re-entry per assessment, done in parking lots and coffee shops. The care plan, supposed to be the source of truth, becomes a compliance document. Detailed understanding gets compressed into structured fields. Daily care runs off notes, not the plan.

In their words
"We actually do everything on paper when we're in the house... because that's how they feel most comfortable and it helps us build trust."
Emily, Home Instead, Indiana (~280 CGs)
The rework
10-60+ minutes of re-entry after every paper assessment. Data entry in parking lots and coffee shops.

The family's story is incomplete at intake and keeps changing after care starts

The story APFM captures is the family's understanding of their situation, which is often partial. Many families call pre-decision, not yet knowing what they need. That incomplete picture then gets compressed across handoffs: what the intake coordinator heard, what the assessor observed on paper, reformatted and stripped of nuance at each step. The caregiver arrives with a task list. Not a story.

And the story doesn't stop evolving. Mom's condition changes. The daughter's work schedule shifts. The care plan that was right in week one isn't right in week six. Today, nobody maintains this thread. APFM goes blind after the referral. The agency captures snapshots but not the arc. The caregiver sees changes in real time but has no way to surface them.

The incomplete story
The family's initial story is partial: what they think they need, not the full clinical picture. Even that partial context gets compressed across handoffs.
The evolving story
The family's situation keeps changing after care starts. Nobody maintains the thread. APFM goes blind. The agency captures snapshots. The caregiver sees it in real time but can't surface it.
Client acquisition was the #1 unasked pain in the survey.
40% of the 58 meaningful open-text responses mentioned lead quality or client acquisition unprompted. It wasn't even one of the ranked options. 5 of 6 direct APFM mentions were negative.
What this means
  • Families often don't know what they need when they first call. About half of agencies absorb unpaid education: what home care is, what it costs, what Medicare covers. That burden could sit with APFM instead.
  • The family's context degrades across handoffs and keeps evolving after care starts. Nobody maintains the thread. APFM has the earliest and most complete picture of the family, and a position to maintain it.
  • The journey extends beyond intake. The home assessment is where referral quality materializes: better starting context means assessors focus on rapport and observation, not re-gathering. The care plan gets more detailed as a result.
  • The core of this (richer context, demand-capacity matching, owning more of the intake) does not require ERP integration. Longer-term extensions would.
What this could look like: family profiles that travel with the referral and keep updating. Demand-capacity matching: connecting referrals to agencies based on available capacity. Owning more of the intake so context doesn't degrade. The endpoint: a scheduled home visit, not a referral.
Next
The caregiver layer
The caregiver layer
Each agency builds its own caregiver bench independently. A shared pool where caregiver context travels across agencies could reduce the overhead that resets every time.
78% are always hiring
Spending thousands on redundant background checks and onboarding each time
High ghost rates in onboarding
Multiple agencies described new hires simply stopping showing up before onboarding is complete
83% know CGs work elsewhere
But no system captures caregiver context across agencies

The trust gap drives hiring overhead

Daryl runs an agency with about 50 caregivers in Clearwater, Florida. His operations coordinator's goal is to hire three people per week. "We invite a whole bunch of people into the office and we do a training session and mass paperwork," he says. "And then we call them in three days for work, and they're like, nah, I already got. So we waste a lot of our time."

Daryl's agency is not an outlier. 78% of agencies are continuously hiring. 74% lost caregivers just in the past three months. Each time a caregiver leaves, onboarding resets to zero: new background check, new paperwork, new probation period.

Agencies have adapted to this. A VP at a 300-caregiver operation across California and Nevada told us her workforce splits into two populations: those who self-select out within 90 days, and a long-term core. "A person who's new to caregiving is going to decide within the first 90 days if this is for them or it is not." The constant hiring engine exists to feed the first 90 days. Agencies don't invest because they expect most new hires to leave. And because agencies don't invest, new hires leave.

In their words
"Monica, who's really great, wants 40 hours. We call her this week and Monica took another job on Friday."
Daryl, franchise agency, Clearwater FL (~50 CGs)
78%
always hiring
74%
lost CGs in past 3 mo
The 90-day expectation
Agencies expect most new caregivers to leave within 90 days. So they don't invest. The hiring engine exists to feed the first 90 days, not to build a workforce.
From the interviews
"We invite a whole bunch of people into the office and we do a training session and mass paperwork. And then we call them in three days for work, and they're like, nah, I already got."
Daryl, franchise agency, Clearwater FL (~50 CGs)
0:35
Why caregivers leave
Among agencies with recent turnover (n=183)
"Always hiring" scales with size
% of agencies that are continuously hiring, by size (n=246)

The top two reasons caregivers leave are hours instability and multi-agency work, not pay. 83% of agencies know caregivers work for multiple agencies. They estimate 38% of their workforce is shared. But nobody coordinates this. Every agency starts from scratch.

Small agencies (1-25): Drop-off is diffuse. 19% say "no significant drop-off." Interview (19%) and background check (18%) are the main exit points. The funnel is short enough that most people make it through.

Mid-size (26-100): Drop-off concentrates at background check/onboarding (19%) and within the first 90 days (18%). The funnel gets longer and more people fall out at each stage.

Large agencies (101+): Interview (28%) and first 90 days (21%) dominate. Larger agencies invest more in hiring but lose people later, after more sunk cost. This reinforces the underinvestment cycle: they've seen this enough times to stop investing early.

The same missing context makes matching fragile

The lack of portable caregiver context doesn't just drive hiring overhead. It also means agencies rely on one or two people's memory to make the decisions that most determine care quality.

At an agency with 450 active caregivers in Los Angeles, one person holds the matching knowledge for the entire operation. "It's kind of in our head," Aisha says. Which caregiver has the patience for an anxious client, whose custody schedule affects their weekend availability, who actually shows up when they say they will. This is the caregiver's context. It exists, but it's not portable.

20 of 22 agencies rely on knowledge in one to three people's heads for matching. 81% rely at least partly on personal knowledge. Only 12% use software. Three matching tools exist, but 10 of 22 agencies bypass them because they solve credentials, not fit.

In their words
"We can look at resumes all day long, and that's not going to tell you how good somebody is."
22 agencies interviewed
20 rely on tacit knowledge for matching
246 survey responses
81% rely at least partly on personal knowledge for matching
From the interviews
"I kind of go with my gut. I don't know if that's scalable."
Nerissa, A Personal Touch Care, Southern Maryland (6-7 CGs)
Nerissa draws on years of nursing and management experience to match caregivers by affect, personality, and background. She knows the model breaks at scale. She can see the limit but has no way around it.
0:43

What makes a caregiver good at their job goes beyond credentials

Holly coordinates clients at a franchise agency in rural Maryland, working alongside a counterpart who manages the caregivers. Six years of rotating through recruiting, scheduling, and coordination accumulated layers of knowledge about the 60-80 caregivers she places. "A lot of that information just kind of resided in my brain," she says. When colleagues cover for her, they ask questions she answers instantly but can't look up: "How am I supposed to know that this caregiver has custody of her children every other weekend?"

Availability does not equal willingness

Caregivers state availability, then routinely refuse assignments within those parameters or accept ones outside them. This is systematic: caregivers maintain optionality. Stated availability is a soft signal, not a hard constraint.

Tacit knowledge
Six years of accumulated knowledge about the caregivers she places. None of it is in the system. When Holly is out, her colleagues can't look it up.
In their words
"They'll say I'm not available on Thursdays. And we'll call them and they'll work a Thursday. And other times it'll say like, oh, yeah, I'm open eight to eight on a Wednesday. And they'll be like, I can't work Wednesdays."
Emily, Home Instead, Indiana (~280 CGs)
How agencies match caregivers
Matching method breakdown (n=246)
Small agencies (1-25 CGs)
Can't hire without clients, can't take clients without caregivers. One loss is catastrophic. Churn looks lower (47%) because the pool is so small.
Mid-size agencies (26-100 CGs)
Multi-agency work is the #1 reason caregivers leave (49%). Personal matching reliance peaks at 86%. Both symptoms of the same missing thing.
What the survey tells us about verification

We asked agencies: if a trusted third party had already verified a caregiver's background check and credentials, would you accept that instead of running your own?

31%
likely or very likely
33%
neutral
37%
unlikely or very unlikely

Among those who said no, the top reasons were control ("my business is on the line"), compliance requirements, and accuracy concerns. Small agencies were most open (41% positive). Mid-size agencies were most resistant (23%).

This is an even three-way split, not a clear rejection. The resistance may itself be a trust issue: agencies say they wouldn't accept third-party verification, but that could change if the system proved reliable. There may also be a say-do gap. We need more research to understand what agencies would actually adopt and what "caregiver context" means to them in practice. The broader opportunity (a system that helps agencies trust caregivers faster) is supported by the rest of the data, but the specific shape of that system is an open question.

Competitive note: Kare is the closest existing tool to a cross-agency caregiver layer model. Both sides rate each other after each shift, reliability scores accumulate, and agencies can see a caregiver's track record across placements. It's early-stage but validates that the market is moving in this direction.

Both problems trace to the same missing thing

Hiring overhead and matching fragility are both consequences of the same gap: no caregiver context follows them across agencies. The trust cycle drives the hiring loop. The knowledge gap drives matching failure. A shared caregiver pool addresses both.

THE TRUST CYCLE New CG arrives as unknown Agency doesn't invest CG doesn't commit Churn (74%) Hire replacement (78%) No context follows

The entry point is onboarding: solve the overhead and agencies have a reason to participate. As caregivers move between agencies, their context builds across the network. That context helps agencies trust faster, which reduces churn and improves matching. What exactly that caregiver layer looks like needs more research.

What this means
  • 78% are always hiring. The top two reasons caregivers leave are hours instability and multi-agency work, not pay
  • Each agency builds its own bench independently, with no visibility into caregivers' history at other agencies. The overhead of onboarding, screening, and learning who someone is resets every time
  • A shared caregiver pool across APFM's network, where context travels with the caregiver, could reduce this overhead. Agencies that participate get access to caregivers who aren't starting from zero
  • The survey shows a three-way split on third-party verification (31% positive, 33% neutral, 37% negative). What agencies would share, and what caregivers would opt into, needs more research
A shared caregiver pool across APFM's large agency network. Agencies that participate get access to caregivers with context that travels: who they are, where they've worked, how they've performed. Over time, that context compounds as more agencies participate. The exact shape needs more research.
Next
The family relationship arc
The family relationship arc
Families often work with multiple providers at once, and no one holds the full picture. The family is the only entity with context across providers, and they're managing it manually. APFM sees the family before they enter any single agency relationship.
Families split care across agencies
Medicaid for weekday hours, private pay for weekends, home health or hospice overlapping. Each agency does a full intake. No information is shared.
Family portals exist but families prefer to call
Agencies that mentioned family portals in their ERP described low engagement. Families want a person on the other end, not a portal to check.
Closure without transfer
Agencies close cases administratively, but what was learned doesn't travel. APFM benefits from knowing when and why care ends: families often transition to senior living.

Families work with multiple providers, and no one holds the full picture

Agencies described a pattern we heard across interviews: families split care across multiple providers. One agency covers Medicaid hours during the week. A second, private-pay agency fills evenings or weekends. Home health or hospice overlaps with home care as the client's condition changes. Each agency does its own full intake. Each gathers the same information independently.

Agencies view each other as competition, so information doesn't flow between them. One operator described getting clients' records through MyChart because the other agency "won't give us anything." The family ends up as the only thread connecting their providers, managing the complexity manually.

Important caveat: This evidence comes from the agency side. We haven't spoken to families directly about their experience coordinating across providers. Whether families would use or trust a cross-provider visibility tool from APFM is an open question. This opportunity also depends on ERP integration, a real technical dependency.
In their words
"That's real. Actually, that's pretty often. The way Medicaid works is they only get a certain amount of hours a week. And if they need more hours with us whose private pay. So they kind of split that time between the two."
Michelle, independent agency, California
In their words
"We won't have information from other companies because again, they view us as competition. So they're not going to give us anything."
Michelle, independent agency, California

The highest-risk moments are also the moments when families need the most context

At a small Acticare franchise in Georgia, Myrna has a rule: no caregiver walks into a client's home alone on their first day. She arrives first. The caregiver waits outside. They walk in together. She's already sent the client a photo and bio of the caregiver. She stays 30-40 minutes, walks through the care plan, and lets the conversation settle before leaving. She built this protocol herself. It's not in the franchise playbook. No technology supports it.

Myrna is one of the few operators who has formalized what most do by instinct.

11 of 22 agencies described structured first-shift introductions. No dedicated tool supports this. Agencies do it with phone calls, printed photos, and personal presence. Brightwheel validates the category: 8M users in childcare with exactly this kind of handoff.

No technology supports this
Myrna built her own first-visit protocol: photos, bios, personal introductions, 30-40 minutes of handoff. It's not in the franchise playbook.
22 agencies interviewed
11 described structured first-shift introductions

Family communication is central, but it mostly happens outside systems

Agencies that mentioned family portals described low engagement. Families want a person on the other end, not a portal to check.

Part of the gap is generational. Many seniors receiving care are uncomfortable with technology. Agencies describe clients who resist digital forms, are uneasy about location tracking, and prefer paper because it feels less formal. Any family-facing product has to meet people where they are, not where the technology is.

From the interviews
"We do everything on paper when we're in the house. That is how they feel most comfortable and it helps us build trust."
Emily, Home Instead, Northern Indiana (~280 CGs)
They piloted tablets and went back to paper. The carbon-copy service agreement lets families keep a copy immediately. The trade-off: 10-30 minutes of data re-entry into WellSky, done in the car at Wendy's or Starbucks.
0:41

At a 280-caregiver franchise in Indiana, caregivers call clients directly to cancel shifts. Clients cancel their own care based on the caregiver's message, bypassing the agency. The concern is caregiver protection, not convenience: preventing unpaid work requests, boundary violations, and hostile after-hours messages.

When care ends, agencies close the case, but the family's journey continues

Agencies do have structured closure processes: billing close-out, system updates, portal deactivation, family notification. Some offer bereavement support and feedback surveys. But what they capture is administrative: that care ended, not what was learned or what comes next. APFM uniquely benefits from knowing when care ends: families often transition to senior living. Agencies have no incentive to flag this. APFM has every incentive.

In their words
"We have the family room. They can go in. They can put appointments on the calendar. There's a lot they can do." [pause] "Usually it's kind of the opposite. Everyone calls us to just talk, to have a person on the other end."
Samantha, Right at Home, Colorado
In their words
"The generation that needs care at this time, a lot of them aren't real tech savvy. A lot of them are intimidated by that."
Tiffany (TC), home care agency
In their words
"Whoever can come up with something real good on that is going to do very well."
Kimberly, VP, corporate agency (~300 CGs), on the back-and-forth between families, caregivers, and agencies
The missing signal
Agencies close cases in their systems, but what was learned about the family doesn't transfer. APFM uniquely benefits: families often transition to senior living.

Family context includes more than tasks and notes: fit, trust, change, and concern signals

What agencies track in their systems is operational: scheduled visits, hours, tasks completed. What actually matters to families is softer: is the caregiver a good fit? Is my parent comfortable? Is the situation changing? These signals live in phone calls, texts, and the coordinator's intuition. They don't make it into the ERP.

Family and caregiver communication is largely handled manually via phone, text, and email, with the ERP used primarily as a documentation repository. The 30-day goal framework during assessment serves a dual purpose: anchoring family expectations and providing a diagnostic tool if complaints emerge. But once care is underway, these signals fragment across channels that the agency can't search, aggregate, or act on.

Beyond tasks and notes
What matters to families: fit, trust, change, concern signals. These live in phone calls and coordinator intuition, not the ERP.
In their words
"Family and caregiver communication is largely handled manually via phone, text, and email, with the ERP used primarily as a documentation repository."
What this means
  • Families work with multiple providers simultaneously, and no information flows between them. The family manages this complexity manually
  • Giving families visibility across their providers and reducing the need to repeat themselves every time a new provider enters the picture
  • APFM sees the family before any single agency relationship begins. That structural position exists. Whether it extends into ongoing trust needs validation
  • First shift, end of care, and care transitions are episodic moments where this cross-provider context matters most
  • This opportunity depends on building a deeper family relationship through the referral-to-care journey first
What this could look like: first-shift check-in tools (like Brightwheel in childcare), cross-provider family context that reduces duplicate intake, care transition signals when families move to senior living. Ongoing care tracking is a longer-term play that depends on ERP integration and family adoption.

From where to how

Emerging design principles

The research doesn't just point to where APFM should invest. It reveals how agencies work, what they protect, and what they won't adopt. These three principles should guide every product decision that follows.

Give families ownership, not onus
Any product for families should feel like simple visibility, reassurance, and continuity. Not a heavy portal or a burden, but something that adds value rather than replacing the relationship.
Preserve the humanity of core relationship moments
Agencies deliberately use paper in the home because it preserves eye contact and signals care. The opportunity is not "digitize the visit" but "reduce the rework after the visit."
Build to augment, not replace, operator expertise
Agencies take pride in their matching ability. The goal should be to help scale their knowledge while preserving craft, not override it with an algorithm.

What comes next

From a referral platform to a trust platform

APFM currently connects families to agencies. The research points to a bigger role: the trusted layer in a three-sided marketplace that makes the entire care ecosystem work better.

Platform currency: trust + context
Trust is currently broken on both sides of the marketplace. Whether and how APFM can become trusted with context (for both families and agencies) is the underlying question these opportunities rest on. APFM's network makes the cross-agency layer possible, but trust repair comes first.
Hover over the dotted lines between nodes to explore each relationship
A Place for Mom TRUST PLATFORM Families Agencies Caregivers Context + trust Data + referrals Portable context
Families ↔ Agencies
Referral-to-care conversion
Increasing demand. APFM owns the referral, so the opportunity is demand-capacity matching: scheduled visits, not just leads. Agencies don't trust APFM leads today. Repairing that trust is the precondition.
Families ↔ Caregivers
Family relationship arc
Longer-term. Giving families visibility across providers, from first shift through end of care. APFM has no visibility here today. Depends on building the family relationship through the referral-to-care journey first.
Agencies ↔ Caregivers
The caregiver layer
Reducing overhead. A shared bench where caregiver context travels across agencies, so every hire doesn't start from zero. Portable credentials, reduced onboarding duplication. The exact shape needs more research, but the problem is clear.

Immediate next steps

1
Caregiver research
We've heard from agencies. The next step is hearing directly from caregivers: what they experience across agencies, what makes them stay or leave, and where portable context would change their experience.
2
Concept development
Build concepts against the three opportunity spaces and emerging design principles. What does a referral context handoff look like? What does a first-shift check-in tool look like? What does the caregiver layer look like?
3
Concept testing with agencies and caregivers
Our recommendation: test concepts with both agencies and caregivers. For agencies, start with the workaround-rich middle (26-100 caregivers), where every problem converges. For caregivers, test whether portable caregiver context changes their experience across employers. Validate willingness to adopt, identify deal-breakers, and refine before building.
Research team
M
Monica
Irrational Labs
D
Deena
Irrational Labs
E
Eva
Irrational Labs
T
Taryn
A Place for Mom
A
Allison
A Place for Mom
Irrational Labs / A Place for Mom / Q1 2026

33 workarounds across 19 agencies

When tools don't work, people build workarounds. These are the systems agencies created themselves, categorized by what they compensate for.

Scheduling and availability
Group texts for shift coverage Spreadsheets tracking real availability (not stated) Personal phone logs of who said yes/no Sticky notes on desk for call-out patterns
Caregiver knowledge and matching
Mental maps of caregiver personalities Free-text notes in ERP ("our strongest CG") Personal notebooks with caregiver strengths Informal peer recommendations between coordinators
Client intake and assessment
Paper forms for home assessments Parking lot data entry after visits Printed client summaries for caregivers Photos of client homes sent to caregivers
Communication and handoffs
Personal mobile numbers for family contact Printed caregiver bios and photos for families Manual email summaries after each shift Paper binders that travel with caregivers
Hiring and onboarding
Spreadsheets tracking applicant pipeline Manual reference check logs Whiteboards tracking onboarding status Personal follow-up calls to prevent drop-off
Documentation and compliance
Spreadsheets taped to monitors (coverage tracking) Duplicate entry across ERP and paper Manual EVV override logs Personal checklists for care plan updates