Scaling Without Breaking

You do not wake up one morning and discover your practice has outgrown its systems. Growth happens gradually, then all at once.

The pattern is familiar. One location schedules surgical consultations one way, another does it differently, and the newest acquisition follows a process no one can quite explain. A report that should take 20 minutes to complete takes 2 days and still feels unreliable. Clinical leads spend their mornings putting out fires that did not exist 6 months ago. The onboarding process depends on whichever manager a new hire lands under. Decisions are made by instinct because the data you need are not available in a usable form.

If several of those statements hit close to home, you are not dealing with a growth problem but with a systems problem that growth has made impossible to ignore.

Ridge Eye Care grew from three locations and roughly 40 staff members to 17 sites, 22 physicians, and more than 225 employees across Northern California. Our experience taught us that practices do not scale on capital, ambition, or technology alone. They scale when they build the right infrastructure at the right stage. This article shares the framework we wish someone had handed us when we began.

THE SCALING STAIRCASE 

Growth becomes dangerous when complexity starts to look normal. One new location, one more physician, or one more service line may feel manageable in isolation. Together, however, those additions can outpace the systems holding the practice together. 

Every stage of growth eventually breaks the systems that worked at the previous stage. The practices that scale well are not the ones that avoid those symptoms forever but the ones that recognize them early and build deliberately in response. That process starts with understanding which pillar your current stage requires.

FOUR PILLARS, BUILT IN ORDER 

Most practices that decide to fix their systems make the same mistake: they try to fix everything at once with a new electronic health record system, a rebrand, a leadership retreat, a consultant, etc. Six months later, the team is exhausted, the initiatives are half-built, and the original problems remain and are now competing with the debt of unfinished projects. The problem is not a lack of effort but a lack of sequence.

After living through every stage ourselves, including a rapid expansion phase that taught us lessons we would have preferred to learn differently, we landed on a framework of four pillars that must be built in order. Each creates a foundation for the next, something we call the scaling staircase. Trying to build governance before data, data before standardized processes, or standardized processes before centralized functions is like trying to build the second floor of a building before the first is load-bearing.

No. 1: Centralization

The instinct when building operational infrastructure is to bring everything in-house. For certain functions, that instinct is exactly right. A useful question, however, is not whether to centralize but which functions are strategic enough to own and which would be better served by a trusted outside partner. We have used both strategies deliberately. 

Bringing human resources in-house gave us insight into our people systems (hiring, onboarding, and performance management) that compounded in value every year it stayed inside the organization. Building an internal information technology (IT) function was a different calculation. When we were expanding rapidly across multiple locations, the speed and cost of outsourced IT could not keep pace. Quoting, negotiating, and delivery were too slow and too expensive. Owning our IT meant we could move at the speed the organization required.

When we needed to build a call center, we chose a different model—a global staffing partner that provides the personnel while we retain management and operational control. This function is neither fully outsourced nor fully in-house. It is the structure that best serves our practice’s needs.

Centralization at scale is not a binary choice between outsourcing and building but a deliberate, ongoing evaluation of what is required for each function at the practice’s current stage. The right answer for a three-location practice is rarely the right answer for a 15-location enterprise, so it is important to build in the flexibility to change.

No. 2: Standardization

The second pillar is where most scaling practices are most exposed and damage accumulates quietly. When Ridge Eye Care entered its rapid expansion phase, we added seven clinics in 3 years. What we did not add quickly enough was the repeatable infrastructure to run them consistently.

The temptation when confronting a gap like that is to standardize everything. Resist it. Overstandardization is its own failure mode. It can produce bureaucracy that suffocates the local culture and manager autonomy that make individual locations work. The goal is not uniformity but rather to identify the processes that are of highest risk when inconsistent and standardize those first.

What belongs on that list differs for every practice. For some, the critical few are the people systems that determine whether staff members feel positioned to succeed or left to sink or swim. For other practices, the priority is clinical and revenue standardization (charge capture processes, appointment types, premium IOL packages, usual and customary pricing, and optical frame and lens selection). For multisite practices, inconsistent scheduling templates and appointment-type definitions can quietly corrupt every productivity metric in the building and produce patient experiences that vary by location for reasons unrelated to clinical quality.

Before a single standard operating procedure is built, it is important to identify where in the organization one person’s departure would reveal that there was no system, only a person. That is where to start. The processes most dependent on individual knowledge are the most urgent ones to standardize. Everything else can wait. 

No. 3: Data and Automation

The third pillar transforms how an organization makes decisions but only if this pillar is built on top of the first two. This is the mistake we see most often. The sequence matters. Practices invest in sophisticated reporting before they fix the processes generating the data. A dashboard built on inconsistent scheduling inputs or charge capture variations is worse than no dashboard because it produces false confidence. 

For us, the foundation is our practice management/electronic health record system, our source of truth. Everything flows from it, which means everything depends on how well we have configured the system to work for us rather than the other way around. We customized our system to match our workflows by reducing unnecessary steps, eliminating redundant clicks, and enabling our physicians to post their own charges directly. The result has been hundreds of thousands fewer clicks per person each year across our organization. That is not a technological win; it is an operational win that technology made possible.

From that foundation, we moved from static reporting to interactive dashboards, a shift that reflects a specific management philosophy—trust but verify. Static reports answer the questions we already know to ask. Dashboards help us find the questions we did not know we have.

The metrics that matter most reveal where an organization is working as intended and where it is not. Year-over-year totals for kept appointments, charges, and payments by location and provider show growth or erosion. Appointment utilization by physician, region, and appointment type shows where capacity is being left on the table. Detailed Current Procedural Terminology analysis with real-time payment trends by payer, location, and provider shows where revenue assumptions are drifting from reality.

Some of the most valuable data, however, lie in the questions practices rarely think to ask. Lost patient analysis (where and why patients leave the practice) is as much a care continuity question as it is a revenue question. Referral pattern tracking, both internal and external, reveals whether physicians are functioning as a coordinated system or a collection of independent practices sharing a roof. Patient population data by diagnosis, cross-referenced against clinical utilization, allow leadership to ask whether care delivery is consistent with the practice’s own standards and to find variation before it becomes a problem.

When data are trustworthy, the quality of every conversation in the room changes. Decisions shift from “I think” and “it feels like” to “the data show this, and here is what we are going to do about it.”

No. 4: Leadership and Governance

Every practice, regardless of size, has team members who are capable of more than their current role requires. The question is whether the organization is paying attention.

Leadership development is not a luxury for large or well-resourced enterprises. It is a discipline every scaling practice either invests in deliberately or pays for eventually. The medical assistant who handles difficult patient interactions with composure, the front desk lead on whom colleagues rely when something goes wrong, and the technician who asks the questions managers are not asking are already demonstrating leadership. The organization’s job is to recognize and develop that capacity before another practice does.

The framework we have found most useful is simple. For any team member in whom we are considering investing, we ask whether they are capable of growing into what the next stage will require. Do they understand the role’s objectives and how their work connects to the larger practice? Do they want the actual job, including its hardest and least glamorous parts? Do they have, or can they develop, the skills the organization will need next?

The key is to compare the answers to those questions against the practice’s vision for the next 5 years, not the experience of the previous 5 years. The person who was exactly right for where the practice was may not be the right person for where the practice is going. The person who seems too junior today may be exactly who is needed in 3 years if their development begins now. The practice’s next senior leader is probably already on the payroll. The practices that find them are the ones actively looking. 

These observations apply to physician leaders as well. Medical training produces exceptional clinicians, but it does not automatically produce operational leaders. Most associate physicians and emerging clinical leaders do not get meaningful exposure to revenue cycle, staffing models, financial statements, governance, or practice operations until they are already in a role that requires that knowledge. An earlier introduction builds the shared language that makes clinical and administrative leadership more collaborative.

A technician developed into a lead today can become a clinic manager in 3 years. An associate physician introduced to operational leadership today can become a partner who can lead in 5 years. Organizations that build this pipeline continuously, not in response to a vacancy but as a permanent operating rhythm, can scale with less dependence on external hiring for every key role.

Culture lives here, too. At 40 people, culture is the founders. At 225 people across 17 locations, culture depends on the systems built to reinforce it, including communication channels, recognition structures, and the way leaders model behavior when no one is watching. Silence, not structure, kills culture. 

BUILD OR BREAK 

Sustainable growth requires recognizing which system the practice has outgrown and building the next one before the strain reaches patients, the staff, and the physicians.

Centralization creates the visibility that standardization requires. Standardization creates the process integrity on which data depend. Data create the clarity that leadership development needs. Leadership development creates the human capacity that makes everything else possible.

The staircase only goes up. A practice either builds the next step before it is needed or discovers it after the fall. 

An AI language model (Claude, Sonnet 4.6, Anthropic) was used to assist with drafting and editorial development of this manuscript. The AI tool served as a writing collaborator under direct author supervision: all intellectual content, professional positions, frameworks, organizational examples, and factual claims originated with the authors. The authors directed all structural decisions, reviewed and revised all generated content, verified the accuracy of all claims and examples, and assume full responsibility for the accuracy, originality, and integrity of the submitted manuscript. The AI tool was not used for data analysis, interpretation, or drawing clinical or scientific conclusions. No AI-generated citations are included.