In this guide, you’ll learn how to turn Notion into a simple, but structured, CRM using three connected databases: People, Companies, and Notes. This setup avoids the most common Notion CRM pitfalls and gives you a system that can grow with your team.

Notion’s strength and weakness lie in its flexibility. It is infinitely configurable, but that also means that it is not necessarily ideally suited to use as a Customer Relationship Management database (CRM) “out of the box”.

This setup is designed for teams running a relationship-driven process (sales, fundraising, partnerships, vendor management). It is not designed for product-led growth or complex multi-sided marketplaces, which require more specialized systems.

Overview: The Relational Database model

I suggest you follow the principles of most CRM platforms, with separate databases for each of the key types of thing you’re tracking. Database professionals will usually call these “objects”. In Notion you’ll set up a standalone database for each and then use the “relational” field to link between them. Understanding how to setup and manage a series of related tables is important to making this work, but can feel counter-intuitive at first.

Why single tables always fail (eventually)

You probably started using Notion with a single database, perhaps titled “Potential Customers” or “Discovery calls” or “Meetings”. It’s easy to continue adding to this structure, but there is a very real risk that as you add more and more structure, things begin to break. If you’re reading this document you probably realized you need something more structured, but it’s worth considering why this approach doesn’t work.

The first place things start to break is when a single table needs to accommodate multiple types of data. For example – if you started by tracking companies you meet what happens when you meet more than one person from the same company? Do all meeting notes go into the same row? Perhaps you make new rows, but that causes other problems – it looks like you have more prospect companies than you do in reality. If you start tracking people, what if a person changes jobs and moves to a new company? What if they are connected to multiple companies?

Other problems occur as you and your team use the flexible structure of Notion to “get work done”. Adding columns titled “Dinner in SF Feb 3rd" or “Potential investor? Yes / no?” is easy to do, but very quickly a simple table becomes hard to navigate, with stale information clogging things up.

The three non-negotiable objects

People, Companies and Notes databases

My suggestion is to structure things before you get started, with objects for People, Companies and Notes (i.e. documents, meeting notes, etc.)

Classic CRMs might also include “Deals” (since you can potentially sign multiple deals with a single company) or “Interactions” (e.g. the meetings you had). Some may also allow you to link your email address and see all related emails for each person or company. For now I think these 3 objects are a great starting point without overly complicating things, if you really need all of the capabilities here you might already be considering buying a full blown CRM product. 

Each row in each table should represent 1 unique and atomic thing – a person, company or document. What that means in plain english is: 

  • An atomic record is the smallest possible record – so we’re looking for a row that refers to exactly one real-world entity (i.e. NOT “Thomas George and Noah Zoschke” or “The Founders”, and not “James Smith” if there are really 2 different James Smiths)

  • A unique record means that the same real-world thing does not appear in multiple rows. So all of the records for Thomas George should be linked to the same Person record, not split across multiple rows.

The best way of doing this is for each table to have 1 master column which you can use to flag duplicates. Choosing what data to use is important - I’ll make recommendations below. 

Setting up the tables:

You’re going to set up 3 different databases, to make it easier you can also look at the example CRM I set up here (with demo data). Set these up one by one as follows:

PEOPLE

Suggested columns:

Email address

Name

Labels

This is your “key column” since email addresses are unique (and names are not)

Pretty self-explanatory

Multi-select

Use a single column rather than adding new columns for each individual use case to keep things simple.

Optional columns:

Role

LinkedIn URL

Location

This could be either Select or Multi-Select, depending on whether you want to be able to record more than 1 role per person

You might want to consider some validation to ensure you have valid LinkedIn URLs here

I suggest thinking about how you’ll use this column and select the right level of specificity. You might only care about country, but if you’ll be organizing in person events then city might be better.

COMPANY

Suggested columns:

Company website

Name

Status or Pipeline status

Labels

This is your “key column” since URLs are unique values

Try to write the name of the company as you’d naturally write it. So “Housecat”, not “Housecat, Inc.” That way you’ll be able to use this data as the input into emails etc.

Set up the options you want for the stages in your pipeline e.g. Screening, Qualified, Demo, Won, Lost 

As above, use this as a catch-all, while avoiding cluttering your CRM with lots of additional fields.

Optional columns:

Sector / Industry

Stage / size

Again think about what you need this for. Does the sector inform how you segment opportunities amongst your sales team? Does it drive unique content marketing?

This is a good example of data that can quickly become stale as companies grow. If this matters a lot to you there are external data vendors, or you could set up an automations to keep your research up to date.

NOTES 

Suggested columns:

Note Name

Summary

Type

Labels

this is your “key column” and ideally is in a useful format like “YYYY-MM-DD Note Title” e.g. “2025-01-24 Phone call with James Jameson”

It’s helpful to have a 1 line summary of the meeting to aid in finding content quickly. You could set up an AI automation to produce this.

Multi-select e.g. Granola notes, Proposal, Customer feedback etc.

As above

Setting up the Relations:

Now you’ll want to set up the relations between these tables, creating a mapping between all 3 fields like so:

Here’s a step by step guide for how to set up your first two way relation between People and Companies: 

  • In People add a new Company column

  • Set the Type to Relation and link to your Companies table

  • Turn on 2 way relation and name the column as “People”

  • Now navigate to your Companies database, a new People column should have appeared. The data in these two columns will be kept in sync. 

  • Test it works by adding a company to a Person. You should see the person recorded in the company record. 

Add a new column and change the type to Relation

Select your Companies table in the Related to field

Turn on Two-way relation

Now repeat for the remaining links: Link People to Notes, and Link Companies and Notes as well. Now your Notion CRM is fully up and running!

User guide

  • Avoid duplicate records: When creating new entries make sure to avoid adding duplicates. Your key data field is your friend here. You can set up rules in Notion to flag duplicates, or periodically review your data to ensure you combine duplicates / near duplicates.

  • Set up custom views using Filters and Kanban view: For example create a “My Pipeline” view for each owner using the Kanban view. Filter by Owner = Me and Status ≠ Won/Lost so each person sees only the companies they’re actively responsible for.

  • Use filtered views instead of adding more columns: Use the label column for ad-hoc filters and views, rather than adding new columns like “Hot lead?” or “Investor maybe”, this way your CRM will stay simple and clean going forwards.

  • Automatically log meeting notes into your CRM: Tools like Housecat, Zapier, or n8n can take notes from call recorders like Granola or Gemini and create a new entry in your Notes database, linked to the right Person and Company. Because you have the concept of People and Companies you can review meetings notes or other docs that don’t have owners on a regular cadence to ensure they are logged appropriately. 

  • Use page bodies for narrative: Use the page body for more detailed notes, context, and anything that doesn't need to be filtered or sorted. Keep structured data in columns, narrative in the page.

  • Rollups for at-a-glance info: Add a rollup column to Companies showing count of linked Documents or date of most recent Document. This gives you a quick sense of relationship activity without opening each record.

You can find a full example set of databases here 

Keep Reading