Last edition, the experts gave me the discipline I was missing. I finally had a proper way to build.
It fixed how I built. It did nothing for what the system kept quietly forgetting.
The promise that vanished
Here is the moment I realised there was a whole class of problem my shiny database could not touch.
A partner — let us call him Jim — told me he would touch base in about three months. I replied, said something warm, and moved on. Normal. The kind of exchange that happens ten times a week.
And then the system lost it. Completely.
No reminder three months later. No record that a promise had even been made. No little nudge saying "Jim said he'd circle back — has he?" The commitment simply evaporated the moment the email thread went quiet. Nothing happened, because as far as my system was concerned, nothing had ever been said.
That should not have been possible. I had just spent months building a database that held everything.
Why the filing cabinet could not help
Remember the filing cabinet from a few editions ago? Postgres. The proper database, the single source of truth. It is brilliant at one thing: storing facts. This contact. That conversation. This decision, on that date.
But a promise is not a fact you file in a drawer. A promise is a connection. It links a person, to a thing they said, to a moment in the future when it comes due. And it only means anything if you can see all three at once, joined together.
My filing cabinet had every fact about Jim. His details in one drawer. Our conversation in another. It just had no way to hold the thread between them — the line that says "this person made this commitment and it matters in September."
I had built something that knew everything and understood nothing. It had all the dots. It could not draw a single line between them.
A map instead of a cabinet
So I added a different kind of database. One built not for facts, but for relationships.
(A normal database stores things in lists. A graph database stores the connections themselves as the main event — people, conversations and commitments as points, and the lines between them as real, first-class things you can follow. Less like a filing cabinet, more like a map with every road drawn in.)
Now the system does not just store that Jim exists and that we spoke. It holds the line between them: Jim → committed to → touching base → due in three months. A promise it can actually follow, not a sentence that scrolls off the screen and dies.
That graph now quietly feeds the things I look at every day — my morning briefing, who my outreach should be chasing. The connections are finally visible, because I finally gave them somewhere to live.
The lesson
Here is the thing I had genuinely not understood.
Data is not the same as relationships. You can know every fact about a person — their company, their history, every email you have exchanged — and still have no grip on the one thing that actually matters: what passes between you. The promises. The threads. The follow-ups that quietly define whether a relationship is alive or dead.
My foundation held the facts beautifully. It took a completely different tool to hold the meaning between them.
And that is oddly human, when you think about it. The people who are good with people are rarely the ones who remember the most facts. They are the ones who never drop the thread!