Clott Operations

Infrastructure

One source of truth, and the systems to run the business on

Your CRM, your billing, your project data and your spreadsheets all say something slightly different, and you are the thing connecting them. I build the layer that makes them agree, automate the handoffs, and document it so it runs without you.

Project based, fixed price per milestone. The first hour is free.

Growth breaks the way you were working

Six systems, six versions of the truth

The CRM says one number, the billing system says another, the project tracker says a third, and the board deck is assembled by hand from all of them the night before. Nobody can answer a simple question without a spreadsheet and a favour.

You are the integration layer

The handoffs only happen because a person remembers to do them. Copy the deal into the invoice. Move the hours into the report. Every new client makes that person's week longer, and that person is usually the owner.

The default answer is to hire

The usual fix is an operations hire at $90,000 a year to do by hand what should be running by itself. That works, and it also means the process lives in someone's head instead of in a system.

None of this shows up as a crisis. It shows up as a week that is four days too long, every week, and a set of numbers nobody is willing to bet the hiring plan on.

What gets built

Every project is different, but it is almost always some combination of these four.

One source of truth

Pick the system that should win for each kind of record, define exactly what a customer, a deal and an invoice mean, then make the rest of the estate agree with it. The output is that two people looking at the same question get the same number.

  • A written data model, so the rules survive you changing tools
  • History migrated and reconciled rather than abandoned
  • One definition per metric, agreed with whoever reports on it

The integrations nobody wants to build

The CRM to the billing system. The project tracker to the reporting. The website and the chat to the customer record. Most of the wasted hours in a growing company are sitting in these gaps, and each one is a small build with a lot of leverage.

  • Two-way where it needs to be, one-way where it should be
  • Failures surface as alerts, not as a quiet gap someone finds in a month
  • Idempotent, so a retry does not double-charge or duplicate a record

Reporting that assembles itself

The weekly number, the pipeline view, the utilization report, the board pack. If a person is rebuilding it every month, it is a system problem. I move it to a scheduled job that writes itself and tells you when it is stale.

  • Scheduled and self-filing, so the report arrives rather than being chased
  • One place where the numbers are defined, so two reports cannot disagree
  • Honest about gaps instead of quietly printing a zero

The plumbing underneath

The part that is invisible until it is missing: where it is hosted, who can reach it, what happens when it breaks, what happens when a key expires, and how the data is backed up. Projects that skip this become a support burden six weeks after handover.

  • Access scoped so a shared link is not the security model
  • Encrypted off-site backups with a restore that has actually been read back
  • Monitoring that alerts someone before a customer notices
  • Written handover, so it does not become only I can touch it

What it has done before

Forecast error, about 50% to under 15%

I rebuilt the forecast for a company of over a hundred people, working alongside the COO, CFO and CRO. The problem was not the model. It was that sales, marketing and customer data lived in different places and nobody could reconcile them. Connecting the data came first, the forecast came second, and the number became something the leadership team could plan against.

260 hours a year back, from one build

A client's reporting was a two-person manual job before every meeting. It became one click. The work was not a dashboard on top of the mess, it was replacing the chain of exports and copy-paste underneath it so the dashboard had something true to read.

A CRM carrying real operational weight

For Venn Digital I built the layer behind their practice research: a scoring engine that audits any US dental practice and returns a personalised gap report in under a minute. That replaced a consultant doing the same research by hand, one practice at a time, and it is the difference between a scalable sales motion and a queue.

How a project runs

1

Audit

What systems exist, what each one actually owns, where the handoffs are manual, and what the numbers are worth. One hour free to see if there is a project here at all, then a scoped audit if there is.

2

Architecture

A written plan: the data model, which system wins for which record, the integration list in priority order, and what gets built in which milestone. You approve it before anything is built, and it is yours to keep either way.

3

Build in milestones

Each milestone is a working thing you can use, priced before it starts, and paid on delivery rather than on hours. If a milestone takes longer than I estimated, that is my problem, not an extra invoice.

4

Prove and hand over

Tested against your real data, not a demo. Documentation and a walkthrough so the next person can pick it up. Then the choice: you run it, or I stay on to run and extend it.

This is for you if

  • You have more than one system holding real company data
  • Someone rebuilds the same report every week or month
  • A handoff between two tools only happens because a person remembers
  • You are making hiring decisions based on numbers you are not sure about
  • Growth means the current process stops working, and you can feel it coming

This is not for you if

  • Everything genuinely lives in one system that does the job
  • You already have engineering or an IT function owning this
  • You want one small fix and nothing around it, which is a smaller conversation

If you are a one or two person business with nothing to connect, this is the wrong page. The AI employee is the better starting point.

What it costs

Priced per project. There is no hourly rate on this page and no package price, because a number quoted before the audit is a guess that somebody ends up paying for.

What moves the number:

  • How many systems have to agree with each other
  • How much history has to move, and how clean it is
  • Whether any custom software has to be written
  • Whether it has to keep running afterwards, and who runs it

You get a fixed price against named milestones in writing before anything is built, and you pay per milestone rather than up front.

One hour, free, capped in writing. You leave with a written list of what can come off your plate, and you keep the plan whether or not you hire me.

30 days to change your mind, and the monthly stops when you can run it

Run it for 30 days. If it is not taking work off your plate, tell me before day 30 and I refund the setup fee in full and collect the machine. No form, no argument. After 30 days the machine and the data are yours, and the monthly ends the moment you can run it without me. No lock-in, no hostage data, no exit fee. If you fire me, it keeps running.

Questions owners ask

What does it cost?

It is priced per project, not per hour and not off a menu. The number comes from how many systems have to agree, how much history has to move, whether custom software is needed, and whether it has to keep running afterwards. You get a fixed price against named milestones in writing before anything is built, and you pay per milestone rather than up front.

How is this different from hiring an operations person?

An ops hire does the handoffs by hand and keeps the process in their head. This puts the handoffs in a system and writes the rules down. You may still want the person, but you want them deciding things rather than copying records between four tabs.

We already have a CRM. Are you replacing it?

Usually not. Most of the time the tools are fine and the problem is that nothing connects them and nobody agreed what the fields mean. I would rather make what you own agree with itself than sell you a migration you did not need.

Who owns the work when it is finished?

You do. Your accounts, your data, your code, and a written handover so it is not a black box. If you want me to keep running it, that is a separate and smaller arrangement.

How long does a project take?

Depends on the systems. A single integration with clean data can be days. A full consolidation with history, reporting and hosting is months. You get the milestone plan before you commit, so you are never guessing at the end date.

What if it does not work?

30 days to change your mind, and the monthly stops when you can run it. One hour, free, capped in writing. You leave with a written list of what can come off your plate, and you keep the plan whether or not you hire me.

Is this only for companies with an IT team?

No, and it is usually the opposite. The companies that need this most are 10 to 100 people with no IT function at all, where the founder or the ops lead has quietly become the integration layer between every system.

The business grows either way

The only question is whether it grows on systems that hold the weight, or on one more person remembering to move data between tabs. That is what an hour of your time settles.