Centralize your data without a data engineer

The problem: data is everywhere, so it's nowhere

After a few months, the entrepreneur has numbers in ten places: analytics in GA4 or Plausible, sales in Stripe, prospects in a CRM or a spreadsheet, subscribers in their emailing tool, expenses in their accounting, audience on social media. Each tool has its own dashboard. To answer a simple question like "how much does a customer cost me and how much do they bring in?", you have to open three tabs and do the math by hand. That's exactly the kind of friction that means you never look at your numbers.

Centralization means gathering the key numbers in one place, ideally automatically. You're not trying to import everything — only the indicators that matter.

The legitimate level zero: the spreadsheet

Before talking about sophisticated tools, let's be clear: for an entrepreneur just starting out, a simple Google Sheet (or Excel) is an excellent centralization point. Each week, by hand or semi-automatically, you log the five or six key numbers: visitors, prospects, sales, revenue, marketing spend. In fifteen minutes a week, you get a historical record most competitors don't have.

The spreadsheet has one flaw — manual entry is time-consuming and error-prone — but one decisive virtue: it's free, universal, and forces you to choose the important numbers. Don't look down on it. Many profitable businesses still run on a single well-kept Google Sheet.

The intermediate level: no-code databases

When the spreadsheet saturates, no-code databases take over without any coding:

Airtable: a supercharged spreadsheet that behaves like a database. You centralize prospects, customers, content, projects, with views, filters, and automations. Decent free plan, then paid.
Notion: mainly a documentation workspace, but its databases work for light tracking combined with documentation.
Baserow or NocoDB: open-source alternatives to Airtable, self-hostable if privacy is paramount.

These tools shine when data is also operational (managing prospects) and not just analytical (counting prospects).

Flow automation: Make, Zapier, n8n

The real time saving comes from automating collection. Rather than copying numbers, you connect tools so they feed your centralization point:

Zapier: the best known, an immense catalog of connectors, very simple. Example: "for every new Stripe sale, add a row to my tracking Google Sheet".
Make (formerly Integromat): more visual and often cheaper by volume, ideal for slightly complex scenarios.
n8n: open source, self-hostable, the most economical at scale for those who accept a bit of technical work.

A typical entrepreneur scenario: Stripe → Google Sheets for sales, form → Airtable for prospects, and a weekly summary sent by email. The whole thing is set up in an hour and then runs on its own.

The advanced level: a lightweight data warehouse

When volumes grow and calculations become heavy for a spreadsheet, you move to a real data warehouse where all sources converge:

BigQuery (Google): cloud warehouse, generous free tier, natively connected to GA4 and Looker Studio.
ETL tools like Fivetran, Airbyte (open source), or Stitch: they automatically replicate your sources (Stripe, database, marketing tools) into the warehouse.

This is a level most entrepreneurs won't reach for a long time, and that's perfectly fine. You come to it when the spreadsheet and no-code show their limits — not before. Mentioning it mainly serves to know the step exists when the need comes.

The progression rule

graph LR
    A[Google Sheets<br/>manual] --> B[Sheets + Make/Zapier<br/>semi-automatic]
    B --> C[Airtable<br/>no-code database]
    C --> D[BigQuery + ETL<br/>warehouse]

The right trajectory is to start as simply as possible and only move up a notch under the pressure of a real need. Building a data warehouse for a site doing ten sales a month means spending weeks on infrastructure instead of looking for customers. Centralizing a Google Sheet in fifteen minutes already gives you a view most don't have.

The deciding criterion: "do I actually look at it?"

Whatever the level, a centralization point is only worth it if it's consulted. A beautiful Airtable never opened is worthless; an ugly Google Sheet looked at every Monday morning is gold. Before choosing a tool, choose the ritual: which day, how often, who looks and decides. The tool serves the ritual, never the other way around.