What we build

Six areas. One layer.

An Operating Partner builds the infrastructure a company needs to operate properly as it grows. Below is what that actually means, in the order the questions usually come up.

The principle

These are not six products. They are components of one operating layer, and a company rarely needs all of them.

Every company that outgrows its systems outgrows them unevenly. One has clean financials and a CRM full of fiction. Another has a disciplined sales process and no idea what anything costs. A third has both and no way to tell who is responsible for what.

So the first job is not building. It is working out which three of the six below are the constraint this year, and which three are a distraction dressed up as a best practice. Most of the value in the first two weeks is the things that come off the list.

We do not build infrastructure a company this size does not need. If a spreadsheet is the right answer for another year, the right answer is a spreadsheet.

01 / Data

One version of the numbers, owned by the company.

Most companies do not lack data. They lack a single place where it has been reconciled, defined, and assigned an owner, so every question turns into a negotiation between three systems.

Data architecture
Where each number lives, which system is authoritative for it, and how the others get it. Written down, so a new hire can follow it.
Reporting
The reports the business actually runs on, built once, produced the same way every period, reconciled to the accounting file.
Dashboards
A small number of views people genuinely open. Not a wall of charts nobody has looked at since the week it was built.
Data ownership
A named person responsible for each set of numbers being right. Without this, everything else decays inside two quarters.
Management information
What leadership sees monthly: the figures, the definitions, and the commentary that turns them into a decision.
Done looks like

Somebody asks what a customer, a job, or a product line actually earned, and the answer takes ten minutes and nobody argues with it.

02 / Systems

Five subscriptions behaving like one system.

Each tool was bought to solve a real problem. Nobody was responsible for how they fit together, so the company now employs people to move information between them by hand.

CRM
Configured to how the company actually sells, with stages that mean one thing, required fields that earn their place, and data good enough to forecast from.
ERP and core systems
Selection, implementation, or the far more common job of making the system you already bought do what you bought it for.
Business applications
The working tools: ticketing, scheduling, field service, project management, inventory. Chosen for fit, not for feature count.
Integrations
The connections between them, built so a record entered once does not need re-entering twice more.
Technology stack
What you run, what it costs, what it overlaps with, and what should be turned off. Frequently this step pays for the engagement.
Done looks like

A record is entered once. The systems that need it have it. Nobody is maintaining a spreadsheet to bridge two tools you already pay for.

03 / Operations

How work moves through the company, written down.

In most growing companies the process exists. It lives in the heads of four people who have been there longest, and it is different depending on which of them you ask.

Workflows
The actual path a job, order, patient, ticket, or deal takes, mapped as it is, then designed as it should be.
SOPs
Written procedures for work that has to happen the same way every time. The test: could someone who has never done it do it to standard tomorrow, from what is written?
Accountability
Who owns each step, who approves, and who gets told. Named people, not departments.
Handoffs
The seams between teams, which is where almost everything is actually lost. Defined, with a trigger and a definition of complete.
Operating cadence
The meetings that matter, on a fixed rhythm, with a fixed agenda and a page in front of everyone. And the meetings that should stop.
Done looks like

A new hire is productive in weeks instead of months, and a good person leaving is a problem rather than a crisis.

04 / Commercial

Pipeline you can see. A forecast that holds up.

The most common version of this problem: the CRM is technically working, the reports run, and nobody in the leadership team believes the number at the bottom.

Pipeline infrastructure
Stages defined by buyer behaviour rather than seller optimism, with exit criteria that make a stage change mean something.
Forecasting
A method, applied consistently, that can be checked against what actually happened. Accuracy matters less at first than the fact that it is measurable.
Sales process
How the company actually wins, written down and made repeatable, so performance is not entirely a function of who is on the call.
Customer data
Clean records, deduplicated, with the fields that drive decisions actually populated. Unglamorous and load-bearing.
Revenue reporting
What came in, from where, at what margin, and what it cost to get. Reconciled to finance, not a separate sales version of reality.
Done looks like

The forecast is a piece of information rather than a negotiation, and the leadership team plans off it without hedging.

05 / Management

The structure that turns a month into a decision.

A company can have good data and still not manage off it. The gap is the structure: what gets reviewed, by whom, on what rhythm, and what happens when a number is off.

Scorecards
A single page per function. Few enough numbers to hold in your head, each with an owner and a threshold.
KPIs
Measures chosen because a person can change them, not because they are easy to pull. Leading where possible, lagging where honest.
Management reporting
A fixed monthly pack in the same format every month, so the shape of a bad month is visible before anyone reads the commentary.
Meeting structure
Weekly, monthly, quarterly, each with a distinct job. Most companies run one meeting three times and wonder why nothing gets decided.
Decision systems
Who decides what, at what threshold, and what gets escalated. This unblocks a founder more than anything else on this page.
Done looks like

The founder is not in every decision, and the decisions made without them are the ones they would have made.

06 / Automation

Remove the manual work that pays to remove.

Automation is last on this list on purpose. Automating a broken process produces a faster broken process, and a good deal of what companies want to automate should simply stop being done.

Workflow automation
The routing, the notifications, the status changes, the re-keying. Boring, high-volume work that costs real hours.
AI where economically justified
We will tell you where it earns its cost and where it does not. A great deal of what gets sold as AI in this market is a report that could have been scheduled.
Eliminating manual rework
The work that only exists because two systems disagree. Usually cheaper to fix the disagreement than to automate the reconciliation.
System handoffs
Passing a record between systems without a person in the middle copying it.
Reporting that builds itself
The monthly pack assembled from source systems rather than by somebody's Sunday evening.
Done looks like

Hours come back to people who were doing work a system should have been doing, and the case for each one was made in dollars before it was built.

Sequence

The order matters more than the list.

Built in the wrong order, each of these undermines the next. This is the sequence most engagements follow, and the reason we architect before we build.

01 First

Make the numbers true

Data and reporting come first, because every other decision on this page gets made off them. Building a management cadence on numbers nobody trusts produces a confident, regular, wrong answer.

02 Then

Make the work visible

Workflows, ownership, and handoffs. Once you can see how work moves, you can see where it stops, and most of the systems decisions become obvious rather than debatable.

03 Then

Fit the systems to the work

CRM, core systems, integrations, and the stack. Configured to the process you just designed, rather than the process being bent around a default configuration.

04 Then

Install the management layer

Scorecards, cadence, decision rights. This is the part that makes the first three stick, and the part most companies skip.

05 Last

Automate what is left

Now the repetitive work is visible, stable, and worth removing. Automating earlier is how companies end up paying to speed up a process they were about to change.

Limits

What we do not build.

We do not build custom software products. If the answer is a piece of software the market already sells, we will tell you which one and configure it properly.

We do not build infrastructure ahead of the company needing it. Enterprise architecture inside a fifteen million dollar business is a cost with a story attached.

We do not build anything that requires us to stay for it to keep working. If your team cannot run it, it is not finished, and we do not invoice for it as though it were.

We do not run your company. We do not dispatch your trucks, manage your floor, close your month forever, or sit in your management seat. If we did, the capability would leave when we did, which is the failure this whole model is designed to avoid.

Next

Which three of the six are your constraint?

One call is usually enough to tell. We will ask what systems you run and where the reporting breaks, and if it is not a fit we will say so on that call.

Talk to an Operating Partner