helban.dev

A company operating system in Notion: 17 areas, 8 databases, zero duplication

An early-stage brand wants one place for the whole company and asks how to lay it out so it grows instead of falling apart. Below is the architecture I would propose: how seventeen areas from the brief collapse into eight connected databases, why exactly this way, and how a dashboard comes out of it that shows, first thing in the morning, what needs attention.

Role: information architecture Tool: Notion · databases, relations, rollups Scope: company operating system
17 → 8
areas to track → core databases
1
source of truth per entity
0
retyped fields (relations, not copies)
The brief

What an early brand actually needs at 9am

An early-stage brand is holding a dozen areas at once: strategy, a product in development, suppliers, research, contracts, finance. The question in the brief isn't "where do I put this", it's "how do I connect it so that at 9am I see what needs attention and what stage every project is at". That is an architecture question, not a pretty-template question, and the client said as much: they want to understand the system, not just receive a workspace.

The one decision that matters

Databases, not nested pages

The most common mistake, and the hardest to undo later, is building the system out of nested pages. A page inside a page inside a page. It looks tidy at twenty entries and falls apart at two hundred: you can't filter it, sort it or count it, and the same fact ends up in three places. The foundation is a small set of databases. Every "page" is then a row you can query, not a folder you have to dig through.

The model

Eight databases, one source of truth

Seventeen things on the list collapse into eight databases. The rest is not separate databases, it is a "Type" property and relations. Suppliers, Manufacturers and Contractors are not three databases, they are one Partners database with a Type field. Strategy, research, brand docs, processes and contracts are not five sections, they are one Documents database with a Type and a relation to whatever they describe.

Entity model of the Notion workspace Projects sits at the centre, connected by relations to Tasks, Products, Partners, Contacts, Documents, Meetings and Finance. Products and Contacts also relate directly to Partners. Projects Tasks Products Partners Contacts Finance Documents Meetings
Primary relation (through Projects) Direct relation between databases

Projects is the operational spine the dashboard reads. Partners, Contacts, Products, Finance and Documents also link to each other directly, so the workspace is a graph, not a star. Each card below lists the key properties and what it relates to.

Projects

The operational spine. This is what the morning dashboard reads.

Key properties
StageOwnerPriorityDates
Relates to
TasksProductsDocumentsFinance

Tasks

Atomic to-dos. Everything actionable lives here, nowhere else.

Key properties
StatusAssigneeDuePriority
Relates to
Projects

Products

The catalogue of products in development, from concept to launch.

Key properties
StageCategoryOwner
Relates to
PartnersProjectsDocuments

Partners

Every external company in one place. Type separates supplier, manufacturer, contractor.

Key properties
TypeStatusCountry
Relates to
ContactsProductsFinanceDocuments

Contacts (CRM)

People. A person belongs to a company, so you never retype the company on the person.

Key properties
RoleEmailPhone
Relates to
PartnersProjects

Documents

The knowledge layer: strategy, research, brand, processes, legal. One Type field sorts them.

Key properties
TypeOwnerStatus
Relates to
ProjectsProductsPartners

Meetings

Meeting notes, chronological and linked to the people and project they belong to.

Key properties
DateAttendeesDecisions
Relates to
ContactsProjects

Finance

Budget and transactions, tied to who you pay and which project it costs.

Key properties
AmountTypeDate
Relates to
PartnersProjects
No duplication

Relations instead of copying

A product points to its manufacturer through a relation, not through a copied address. Change the manufacturer's details once and every product is instantly correct. This is the exact answer to "avoid duplicate information": a fact has one owner, and everywhere else is just a pointer to it. The same holds for people and companies, projects and their costs, meetings and the decisions they produced.

The mistake this prevents

Copy a supplier's address onto three product pages and you now have three versions the day they move. With a relation there is one supplier record; the products read from it. This is the difference the client will feel in month six, not week one, which is exactly why it has to be right at the start.

The morning view

Rollups build the dashboard

The home dashboard stores nothing. It reads. Projects rolls up its tasks: how many are open, the next due date, percent done. Finance rolls up cost per project. You open Notion at nine and the dashboard already answers "what needs my attention": tasks due today and overdue, projects laid out on a board by stage, recent meetings. You don't click into each project one by one, the view is assembled from filtered databases, not from new data.

  • "My tasks today" is a Tasks view filtered to due ≤ today and assigned to me.
  • "Projects by stage" is a board of Projects grouped by Stage, so status is one glance.
  • "Budget vs actual" is Finance rolled up per project, next to the work it pays for.
Scale

It grows with the company, not against it

The system grows by adding properties, not databases. A new product attribute is one column, not a new section. When people join the team you share top-level sections with permissions, and because the data isn't duplicated, access control stays clean: one database, one place to set who sees what. That is the difference between a system that grows with the company and one you have to rewrite after a year.

What you get

Architecture first, clicking second

I would start by designing the architecture, not by clicking. First the model: eight databases, their properties, the relations and rollups, and the reason behind each decision, so you understand the system instead of just receiving it. That can be billed as one closed phase. Then, if we fit, we build the workspace and I teach the decisions along the way.

Phase 1

Architecture blueprint

I map your areas onto the database model and hand you the schema, the relations and the "why". You build it, or I do.

Phase 2

Built with you

I set the workspace up in Notion and explain each decision live. It stays yours and you extend it yourselves.

I'll map the model to your areas

Tell me which areas you want in one place and I'll send back a first architecture sketch, before we talk price.

Get in touch