The spreadsheet always starts well. Two columns, five rows, a link and a commission rate. Eighteen months later it has forty tabs, three colour-coding systems you no longer remember the meaning of, and a column called 'notes2'.
The failure is not the tool. It is that a spreadsheet has no opinion about what an affiliate link is, so every new situation adds a column instead of fitting an existing shape.
What is actually worth tracking
Five things, and not many more. The programme or offer. The network it runs through. The destination link and your tracking variant of it. Which project or site you are promoting it on. And the revenue it produced, by month.
Everything else people put in affiliate spreadsheets — creative variants, historical rate changes, contact names — is either derivable or belongs in your notes rather than your ledger.
The problem nobody plans for: link rot
Affiliate links die quietly. Programmes close, merchants migrate networks, offers expire, and your link keeps returning a page that no longer converts. Nothing tells you. Traffic keeps arriving and earning nothing.
This is the single strongest argument for tracking links somewhere structured rather than in a document: a system that knows what a link is can check whether it still resolves. A spreadsheet cannot, because to a spreadsheet your link is a string of text.
Organising by project, not by programme
Most people group links by merchant, which feels natural and is the wrong axis. You do not make decisions per merchant; you make them per site. The question you ask at the end of a month is 'is this site worth more of my time', not 'is this merchant worth it'.
Group by project, and each project's performance and links sit together. The merchant view is still available when you need it, but it stops being the primary structure.
A ten-minute monthly routine
Once a month, enter what each network actually paid — not what the dashboard estimated mid-month, since those figures get revised. Check anything flagged as a dead link. Look at which project earned most per hour of attention, not per click.
That last number is the only one that changes behaviour. Revenue per project tells you what happened; revenue against effort tells you what to do next.
On manual entry
Manual revenue entry sounds like a downgrade from an API integration, and for a large operation it is. For everyone else it is an advantage: networks report on different schedules, revise figures after the fact, and pay in different currencies. Typing a number once a month forces you to look at it, which is the actual point.