Friction Log: Finding My Next Product Without Building

I'm not building anything for 60 days.

Instead I write down every time a tool doesn't do what I want. On day 60 I read the list. Whatever keeps hurting is the candidate for my next product.

This is the method, with real entries from my log.

Why a friction log

My last SaaS taught me one thing: I built first and asked later. You can read the numbers in the post-mortem and why I got no feedback in the interviews post.

So this time I want the idea to come from something that actually costs me time or money, and that I can write down with a date.

A friction log is the simplest version of that. One line per annoyance. No brainstorming.

How it works

Every entry has four parts:

Date: when it happened.
Friction: what did I want a tool to do that it didn't?
Workaround: what did I do instead?
Severity: 1 is a small annoyance, 3 means I lost real time or money.

Rules:

One line per entry. Don't solve it, just write it down.
Log it when it happens, not at the end of the week. If it happens again, log it again with the new date.
No product decisions before day 60. Small scripts for my own workflow are allowed. A product is not.

My real entries

My 90-day challenge started on October 5. I started the log a few days earlier. Here are five entries, shortened:

Date	Friction	Workaround	Severity
Oct 1	Where do I capture content ideas quickly from my phone?	Buffer's Ideas board, synced to my AI	2
Oct 5	Making a non-template X banner and avatar takes trial and error	Canva blank page, many iterations	1
Oct 5	Instagram deactivated a new account after video verification	Appeal, wait, retry carefully	3
Oct 6	AI chats lose context between conversations	A folder of notes the AI reads first	2
Oct 9	No single tool for posting text, image and video to all platforms via AI	Buffer + Canva + a public file bucket	2

By October 9, when I wrote this, I had 11 entries.

Most of them are about my own content workflow. That makes sense, because that's what I do all day right now. It also means the log only sees one kind of problem.

The filter: my own pain is a data point of one

A log of my own annoyances is a start. It's not proof that anyone else would pay.

So every entry that keeps coming back goes through three more questions:

Who is the buyer, and do they already pay for something like it? If nobody pays for anything nearby, the pain is probably not big enough.
Is the market already crowded? Some competition proves demand. Twenty competitors means I'd be fighting on price.
Do the buyers make money with it, or do they hope to? People who earn with a tool pay for it. People who hope to earn mostly look for free versions. My last SaaS had a lot of the second kind.

On October 9 I tested this on the last entry in the table: one tool to post everything everywhere through AI. I checked the market the same day. There are more than 20 scheduling tools, and most of them already let AI agents post for you.

Crowded market, low potential. I wrote down the result and moved on. That is what the log is for: killing ideas cheaply.

The gate on day 60

Day 60 is December 4. Then I read the list. To move on to a product I need two kinds of signals:

5 entries with severity 2 or 3 that came up more than once
2 or 3 "I'd pay for that" signals, from other people, not from me

The second part is the hard one for me. It means talking to people, which I avoided last time. I'll start in writing: replies and DMs on X.

If I don't have both, I don't build. I keep logging until the end of the challenge on January 3. If there's still nothing then, I keep logging. No signal means no product.

Why this is hard for me

I have a habit of researching instead of publishing. I could spend the 60 days reading about markets and call it work.

So the log has a time limit, and the posts come out weekly. The research I do has to end up as an entry or a post.

I'm doing this with a friend. Every week we show each other our numbers. If you want to follow the log and the numbers too, the signup is below.