Most of Building a Company Is Just Fixing the Plumbing
Welcome back to Founder Mode and Happy New Year, everyone!
As I look back this year at Pretty Good AI, the biggest lesson is not about strategy or big wins. It is about the small things that quietly make or break momentum.
Most progress did not come from breakthroughs. It came from fixing what was already there. Cleaner systems. Better data. Clear ownership. Fewer handoffs that break under pressure.
This year reinforced something I keep relearning. Momentum does not usually disappear in one dramatic moment. It leaks away through messy tools, unclear processes, ignored details, and decisions we put off because they feel boring.
That is why so much of building a company ends up being about plumbing.
Connecting ads to the CRM. Connecting the CRM to the bank account. Connecting the brand to the actual customer experience. And trying not to let real life, bad data, or unpaid invoices break the flow.
At Pretty Good AI, a lot of my time is spent fixing these invisible connections. Not because they are exciting, but because when they break, everything else breaks with them.
What follows are a few hard-earned lessons from the last stretch of building. None of these is theoretical. All of them came from things that broke, leaked, or quietly slowed us down until we fixed them.
Clean slates are underrated
There is a specific joy in starting clean.
New workspace. New logo. Old accounts deleted.
When you wipe the slate clean, something important happens. You remove legacy excuses.
No more “this is how it was set up before.” No more “we cannot change that because it might break something.” No more systems that no one owns.
We recently migrated to a new workspace and deleted all the old user accounts. We kept a small core team and rebuilt from there. Fresh design. Fresh permissions. Fresh expectations.
From that moment on, if something breaks, it is on us.
That accountability is uncomfortable, but it is powerful. It forces clarity. It forces ownership. It moves things forward fast.
Ad platforms are not your friend
Ad platforms are not neutral.
Their default settings are not designed to protect your budget. They are designed to spend it.
More than once, we have seen platforms quietly optimize toward the wrong audience. Wrong country. Wrong intent. Wrong quality. When we asked for help, the answer was always the same.
“Just give us blanket approval to make changes.”
Never do that.
One platform kept sending us leads from the wrong country. When we pointed it out, they asked for full control of the account instead of fixing the targeting.
We pulled back and fixed it ourselves.
Your job as a founder is not to trust platforms. It is to verify outcomes. Every setting should exist for a reason. Every dollar should have a job.
Micro-budgets are for testing plumbing, not growth
Spending ten dollars a day on ads will not grow your company.
It will not produce real lead volume. It will not prove that a channel works.
But it is very good for one thing.
Testing plumbing.
We run small campaigns just to make sure the tracking pixel fires, the click shows up, and the data hits the CRM correctly. Before we spend real money, we want proof that the system works end to end.
Too many teams skip this step. They turn on real budgets before the plumbing is solid. Then they argue about performance when the real issue is broken tracking.
Growth comes later. First, make sure the pipes do not leak.
Sales stages are not project management
Your sales pipeline should tell you one thing. Where deals stand.
It should not be a checklist of internal tasks.
At one point, we had sales stages like “send contract” and “provision account” mixed directly into the CRM. The result was confusion. Deals looked stuck when they were not. Forecasts became unreliable.
Sales stages are about buyer intent. Operations are about execution.
Once we separated those two, everything got clearer. Sales stayed focused on closing. Operations got their own system. The pipeline finally reflected reality.
If your CRM feels messy, it is usually doing too many jobs at once.
Forms should be idiot-proof
If you want three pieces of information, give the user three boxes.
Do not ask for everything in one big text field.
People will format answers differently. They will skip details. They will misunderstand instructions. That is not their fault. It is yours.
We once asked for multiple website URLs in a single field. Some people used commas. Others used spaces. Some added explanations.
The data became hard to use. Automations broke. Follow-ups slowed down.
We fixed it by making the form simpler. One box per input. Clear labels. No interpretation required.
Clear forms create clean data. Clean data creates leverage.
Cash flow leverage is real
Even large companies miss payments.
Not because they cannot pay, but because their internal systems are slow and messy. Invoices bounce between departments. Priorities shift. You fall down the list.
We had a major client who owed us a meaningful amount of money for months. The reason was always the same. Internal issues.
Eventually, we set a clear line. If payment did not come through, the service would pause.
That changed the conversation quickly.
This is not about being aggressive. It is about respecting your business. Cash flow is oxygen. You cannot let it drift.
The drive-by traffic opportunity
For a long time, I ignored traffic that did not convert right away.
If someone visited the site and left, I wrote it off as noise.
That was a mistake.
A lot of people are just passing through. They are comparing options. Learning the category. Getting familiar with the space.
Now we treat that traffic differently. Clear messaging. Strong positioning. Simple follow-ups. Not everything needs to convert on the first visit to be valuable.
Some of the best conversations start weeks after that first drive-by.
Five key takeaways
- Clean slates remove excuses and force ownership
- Ad platforms optimize for spending, not for your business
- Small budgets are for testing systems, not scaling growth
- Sales pipelines should reflect buyer intent, not internal tasks
- Clear systems protect your time, data, and cash flow
Final thoughts
Most founder pain does not come from big strategic mistakes.
It comes from small breakdowns that go unfixed. Leaky plumbing. Messy data. Unclear ownership. Delayed cash.
Fixing these things is not exciting. It is not what people post about. But it is what makes everything else work.
At Pretty Good AI, we focus on these details because they compound quietly. When the plumbing works, teams move faster. Decisions get easier. Growth feels calmer.
That is the work.
See you on Friday,
-kevin
Recent Social Posts
Recent Podcasts
Catch up on past emails here.
2810 N Church St #87205, Wilmington, DE 19802
Unsubscribe · Preferences
