This one comes from the 30-Day AI Challenge. On 26 October, the false start, I gave ChatGPT the first version of the kickoff prompt and it came back with three business ideas. All three went through Roast My Plan, the tool I built to pick apart ideas before anyone spends money on them. This is the third of the three.
Why it sounds good
Affiliate marketing does work. There's no stock and no customer service, and if you're useful, the commission follows. Doing it honestly sounds like the decent version of something that's often done badly.
1It describes the end, not the hard part
The hard part of affiliate marketing is getting people to trust you. Any niche that pays decent commission already has established review sites, YouTube channels and long-time forum regulars who've earned that trust over years.
2Search engines are against thin review sites
Google has spent the last few years pushing down sites that exist mainly to earn commission. Most beginners pick something like "best productivity tools for remote workers" and find thousands of pages already there.
3Readers can tell
If you don't really know the field, your comparisons read like reworded spec sheets, and people can smell the commission. "Honest" only means something when you're one of the people you're writing for.
4You're building on someone else's land
Expect months of content before anything comes in. Then a search update can wipe out your traffic overnight, and affiliate programmes can cut their rates or close without warning.
The £0 smoke test
Go very narrow. Not "project management tools", but something like "time-tracking for freelance translators who bill by the word". Ideally it's a group you're already part of.
Write one genuinely useful comparison and share it where that group already talks. If people reply, ask questions or share it on, there's something there. If nobody bites when it's free and in front of them, a search engine won't do any better.
It's worth doing if you know a narrow field well, can live with six to twelve months of low earnings, and have a way to reach people beyond search. Otherwise you're building on rented, shaky land.
The unedited Roast My Plan result is here: raw roast. The rest of the race: the 30-Day AI Challenge.
ChatGPT Bonus Roast 3
Identify a specific audience with a recurring problem and help them choose relevant tools or products through honest comparisons, tutorials and recommendations.
- Discovery: Search-friendly answers, comparison content and platform-native posts.
- Offer: Relevant third-party products with legitimate affiliate programmes.
- Revenue: Commission when qualifying referrals convert under the programme's terms.
Appraisal Verdict
This is affiliate marketing dressed up as utility. The model works, but you're describing the endgame without the hard part: earning trust in a saturated space.
Every niche with decent affiliate payouts is already covered by established sites, YouTube channels, and Reddit veterans who've built credibility over years. You'll be SEO-fighting The Wirecutter clones and dodging Google's helpful content filter, which actively punishes thin affiliate plays. Most beginners pick something like "best productivity tools for remote workers" — forgetting that 10,000 others already rank for that.
The only way this works is extreme specificity: not "project management tools" but "time-tracking for EU-based freelance translators billing by the word." You need an audience you already understand deeply, ideally one you're part of. Otherwise your comparisons sound like rewritten spec sheets and people smell the commission motive immediately.
Real risks: months of content before traction, algorithm changes killing your traffic overnight, affiliate programs cutting rates or shutting down.
Worth doing if: You have genuine expertise in a narrow vertical, patience for 6-12 months of low earnings, and a content distribution plan beyond just SEO. Otherwise you're building on rented, shaky land.