ABA Practice KPIs: The Weekly Scorecard for Owners

Which ABA practice KPIs matter most

The ABA practice KPIs that matter most are the few that tell an owner, every Monday, whether families are getting into care, whether scheduled care is being delivered, and whether the center is getting paid for it. Most centers track far more than that. Their practice management system offers dozens of reports, and someone exports a few of them into a spreadsheet each month. Nobody acts on most of it, because nobody owns most of it.

This guide proposes a different approach: eight numbers, one owner each, reviewed every Monday in 30 minutes. We call it the Monday Eight. Each number has a formula, an owner, a threshold that turns it red, and a first action when it does. The thresholds are yours to set; we explain how.

The guide covers:

The Monday Eight scorecard, with a copyable template
How to run the weekly review so it leads to action
Which numbers belong in a monthly review instead
Standardizing KPIs across locations
Building the operations team behind the numbers
Adopting software and AI responsibly, including privacy basics

It is written for owners, CEOs and operations directors of US ABA centers, from one location to many.

Demand is rarely the constraint. CDC's monitoring network estimates that about 1 in 31 children aged 8 has been identified with autism, based on 2022 data. The constraint is the operation: how quickly families move from referral to care, and how reliably the center delivers and bills what it schedules. The Monday Eight is built around those two questions.

The Monday Eight

#	KPI	Formula	Owner	Turns red when…	First action when red
1	Referral to first session	Median calendar days, for clients who started in the last 4 weeks	Intake lead	Above your target for 2 weeks running	Review files over target by stage; fix the slowest stage
2	Ready-to-start families waiting	Count of families with authorization and availability confirmed, plus the oldest wait	Scheduling lead	Count rising 2 weeks running, or oldest wait above target	Match open recurring blocks to the ready list
3	Sessions delivered vs. scheduled	Delivered ÷ scheduled, last week	Operations lead	Below target	Read cancellations by reason; check the fill protocol
4	Cancellation fill rate	Cancelled slots refilled ÷ cancelled slots, last week	Scheduling lead	Below target	Review the fill-priority list and short-notice families
5	Hours delivered vs. authorized	Delivered ÷ authorized to date, across active clients	Clinical director	Below target	List clients furthest behind pace; agree actions
6	Renewals at risk	Authorizations ending within 30 days with no renewal submitted	Billing lead	Any above zero	Escalate each one today
7	Billing lag	Median days from session to claim submitted	Billing lead	Above target	Find the blocking step: notes, review or submission
8	Supervision delivered vs. required	RBTs on track to meet monthly supervision ÷ all RBTs	Clinical director	Any RBT off track after mid-month	Book the missing supervision this week

Why these eight

Each one answers a question an owner cannot afford to get wrong:

1 and 2: Are families getting into care, and how long do they wait?
3 and 4: Is scheduled care actually delivered?
5 and 6: Is the care families were approved for being used, and will it continue?
7: Are we getting paid promptly for what we delivered?
8: Are we meeting a certification requirement that also protects quality?

Number 8 uses the BACB's RBT rule as its floor: each RBT must be supervised for at least 5% of their monthly service hours, including two face-to-face contacts. It is listed weekly because missed supervision is easy to fix mid-month and hard to fix after the month ends.

How to set thresholds

Do not borrow thresholds from a blog post, including this one. Published benchmarks for ABA operations vary widely, and most come from vendors or consultants without a stated method. Instead:

Measure for four to six weeks without setting any targets.
Set the first target slightly better than your current median. A target you can reach builds the habit; a target you cannot reach teaches the team to ignore the scorecard.
Review targets every quarter. Tighten the ones you consistently meet.

A copyable template

Use one row per KPI and one column per week. Keep the last eight weeks visible so trends are obvious.

KPI	Owner	Target	Wk 1	Wk 2	Wk 3	Wk 4	Trend	Action this week
Referral to first session (days)
Ready-to-start waiting (count / oldest)
Delivered vs. scheduled (%)
Cancellation fill rate (%)
Delivered vs. authorized (%)
Renewals at risk (count)
Billing lag (days)
Supervision on track (%)

A spreadsheet is fine to start. What matters is that the numbers are ready before the meeting, calculated the same way every week.

A short weekly scorecard where every number has an owner, a red line and a first action.

A worked example: one red week

Here is an illustrative Monday at a two-location center. The numbers are made up to show how the scorecard leads to action.

KPI	Target	Last week	Status
Referral to first session	45 days	58 days	Red, second week
Ready-to-start waiting	6 or fewer	11, oldest 34 days	Red
Delivered vs. scheduled	90%	91%	Green
Cancellation fill rate	40%	22%	Red
Delivered vs. authorized	85%	86%	Green, watch
Renewals at risk	0	2	Red
Billing lag	3 days	3 days	Green
Supervision on track	100%	100%	Green

Four numbers are red, and they are connected. The intake lead finds that referral-to-first-session rose because 11 families are ready to start but waiting for a staffing match, not because intake slowed down. The scheduling lead notes the fill rate dropped because the short-notice list has not been updated in two months, so cancelled slots went unfilled while ready families waited.

The team leaves with three actions:

Scheduling lead: refresh the short-notice list and offer this week's cancellations to the 11 ready families first. Due Wednesday.
Operations lead: check whether two RBTs at the second location can each take one new recurring block. Due Friday.
Billing lead: escalate the two renewals at risk with the payers today and tell both families what is happening. Due today.

Next Monday, the first review item is whether those three actions happened. Without the scorecard, each of those problems would have been noticed separately, probably later, and probably by a frustrated family first.

In the worked example, four connected red numbers turn into three actions, each with an owner and a date.

Running the Monday review

A scorecard only helps if the review leads to action. Keep the meeting short and the format fixed.

A 30-minute agenda

Green numbers (5 minutes). Read them out. No discussion.
Red numbers (15 minutes). For each, the owner explains the cause in one or two sentences and proposes one action with a date.
Last week's actions (5 minutes). Done, not done, or no longer needed.
One improvement (5 minutes). Pick one pattern to fix this week, not five.

Rules that keep it useful

The owner speaks for their number. Not the person who exported it.
Every red number leaves with an action and a date.
No new metrics in the meeting. Propose them for the monthly review instead.
Write down decisions in the same document as the scorecard, so next week starts where this week ended.

The Monday review keeps a fixed 30-minute agenda so it always ends in actions.

What the Monday Eight deliberately leaves out

The scorecard is an operations tool. It deliberately leaves out three things that still matter.

Clinical quality and client progress. How clients are progressing, and whether treatment plans are working, are clinical questions. They belong in a review led by your clinical director, with measures the clinical team chooses. An operations scorecard should support that review, for example by making sure supervision happens and authorized hours are delivered, not replace it.

Family experience. Numbers 1 to 4 tell you a lot about how families experience your center, but not everything. Consider a short, regular check-in with families about communication and scheduling, reviewed monthly alongside complaints and compliments.

Staff experience. Turnover is a lagging signal: by the time it moves, people have already left. Regular one-to-one conversations, exit interviews and a simple pulse survey give earlier warning. Review what you learn monthly, and connect it to scheduling patterns such as fragmented days or missed breaks.

Keeping these out of the Monday review is not a judgment that they matter less. It keeps each review focused on decisions its attendees can make.

What belongs in the monthly review instead

Some numbers move too slowly, or are too noisy, to review weekly. Look at these once a month, alongside the four Monday Eight trends:

Monthly KPI	Why monthly
Staff turnover, by role	Too few events to read weekly
Open positions and time to fill	Hiring cycles are longer than a week
Claim denial rate, by reason	Needs enough claims to show a pattern
Collections vs. billed	Payment timing makes weekly numbers noisy
Client discharges, by reason	Needs context and clinical review
Clinic-level revenue and cost	Accounting closes monthly
Utilization, by role and location	Weekly swings mislead; trends matter

Practice management vendors publish longer KPI lists that include most of these, such as attendance, hours delivered against authorized, utilization, time to onboard, turnover, authorization expiry, billing lag, denial rate and discharge rate. They are all useful. The point of separating weekly and monthly is attention: the weekly review should only contain numbers someone can act on this week.

Standardizing KPIs across locations

With more than one location, the scorecard is only useful if every site calculates every number the same way. Otherwise you compare definitions, not performance.

Write a one-page definition for each KPI

For each of the Monday Eight, write:

The exact formula, including what counts and what does not
The data source and who pulls it
When it is calculated (for example, Monday 8 a.m., covering the previous Monday to Sunday)
The owner at each site

One source of truth

Pull numbers from one system wherever you can. When a number has to come from a spreadsheet, keep that spreadsheet in one shared place with one owner. Two copies of a waitlist will disagree within a week.

Compare sites, then visit

When one site is consistently red on a number, look at its process before its people. A site with a long referral-to-first-session time usually has a stage that stalls, often document collection or first contact. Our guide to the ABA intake process from referral to first session explains how to find it.

Building the operations team behind the numbers

Every KPI needs an owner, and as a center grows, the owner of most numbers can no longer be the founder. A common path:

First location: the owner or clinical director owns most numbers, with an intake coordinator owning intake.
Growing caseload: a dedicated scheduler takes numbers 2–4, and billing takes 6–7.
Several locations: an operations director owns the scorecard itself, and each site has local owners for the same eight numbers.
Larger groups: specialist roles appear, such as an authorization coordinator, a credentialing lead and a data or reporting lead.

Hire for the number that is red most often and has no clear owner. That is usually where the next role belongs.

Before hiring, check whether the number is red because of capacity or because of process. If a step has no written definition of done, adding a person to it often adds another way of doing it. Write the process down first, measure it for a month, and then decide whether the work needs another person, a tool, or a better handoff. The scorecard makes that decision easier, because the baseline is already there. For the scheduling side, see ABA scheduling best practices; for authorizations, see ABA prior authorization tracking.

Adopting software and AI responsibly

Software and AI can take a lot of manual work out of these numbers: pulling data, chasing documents, sending status updates, flagging renewals. They also introduce new risks. Treat adoption as an operations project with clear rules.

Start from the scorecard

Pick the red number that costs you most, map the workflow behind it, and measure it for a month. Then evaluate tools against that workflow. A tool that cannot show you how it moves that number is not the right tool yet.

Protect patient information

Any vendor that creates, receives, stores or transmits protected health information on your behalf is a business associate under HIPAA, and you need a business associate agreement with them before data flows. HHS publishes sample business associate agreement provisions you can use as a reference. Also ask each vendor:

Where data is stored and whether it is encrypted in transit and at rest
Who can access it, and whether access is logged
Whether your data is used to train models or for any other purpose
What happens to your data when the contract ends

No vendor can be "HIPAA certified"; there is no such certification. Ask what the vendor does, not what badge it shows.

Keep people in charge of decisions

AI tools work best on well-defined tasks with a person approving the result: drafting a request for missing records, preparing a status update, or flagging an authorization that is behind pace. Clinical decisions, medical necessity and coverage decisions should stay with qualified people. Write down which actions a tool may take on its own and which need approval.

Use a framework for AI risk

The NIST AI Risk Management Framework, released in January 2023, is a voluntary framework organized around four functions: Govern, Map, Measure and Manage. For a small center, a lightweight version is enough: decide who is accountable for each AI tool (Govern), write down what it does and could get wrong (Map), check its output against your scorecard and spot-check its work (Measure), and act on what you find (Manage).

Measure the result

Adopt one tool at a time, against one number, with a baseline. After a month, compare. If the number did not move, find out why before adding the next tool.

Common scorecard mistakes

Mistake	What happens	Fix
Too many metrics	Nobody reads them; nothing changes	Eight weekly numbers, the rest monthly
No owner per number	Red numbers stay red	One named owner who speaks for it
Averages instead of medians	One unusual case distorts the picture	Use medians for time-based numbers
Borrowed targets	Targets feel arbitrary and get ignored	Measure first, then set your own
Different definitions per site	Sites can't be compared	One written definition for each KPI
Numbers pulled during the meeting	The meeting becomes data entry	Numbers ready before the meeting starts
Discussing every number	The meeting runs long and loses focus	Discuss red numbers only

Most of these are habits, not tools. A center with a simple spreadsheet and a disciplined Monday review usually learns more than a center with an expensive dashboard nobody opens.

Where to start on Monday

Copy the Monday Eight template and name an owner for each row.
Write a one-line formula for each KPI and agree it with the owners.
Pull last week's numbers without targets, and keep pulling them for four to six weeks.
Book a recurring 30-minute Monday review with the owners.
After a month, set first targets slightly better than your current medians.

This article is for informational purposes and is not legal, billing or clinical advice.

Questions

Which KPIs should an ABA clinic owner review every week?

A short list the team can act on that week: referral to first session, ready-to-start families waiting, sessions delivered vs. scheduled, cancellation fill rate, hours delivered vs. authorized, renewals at risk, billing lag, and RBT supervision on track. Give each one an owner and a red threshold.

How do I standardize operations across multiple ABA locations?

Write one definition for each KPI and each core process, such as intake stages, the cancellation fill protocol and the renewal countdown, and use them at every site. Pull numbers from one system, compare sites weekly, and when a site is consistently red, look at its process before its people.

How can an ABA practice use AI for operations without risking HIPAA?

Sign a business associate agreement before any vendor handles protected health information, and ask where data is stored, who can access it, whether it is encrypted and whether it is used for training. Keep people approving decisions, especially clinical, medical necessity and coverage decisions, and adopt one tool at a time against a measured baseline.

What targets should I set for ABA operations KPIs?

Your own. Published benchmarks vary widely and rarely state their method. Measure each KPI for four to six weeks, set the first target slightly better than your current median, and tighten targets each quarter as you meet them.

How long should a weekly KPI review take?

About 30 minutes if the numbers are ready beforehand: five minutes on green numbers, fifteen on red numbers with an owner and action for each, five on last week's actions and five on one improvement to make this week.

ABA Practice KPIs: The Weekly Scorecard for Clinic Owners