Eight Gaps Where People Named the Product They Want to Leave

GapMine tracks 1,083 open gaps right now. Only 115 of them have four or more distinct people publicly asking for the same missing thing: 90 with four to seven askers, 7 with eight to eleven, 18 with twelve or more. We pulled that top slice and read the actual posts. In eight of them the person never describes a feature. They name the product they already use, and ask for a way off it.

A feature request tells you what to build. An exit request tells you who is paying today, what broke the deal, and what must be imported on day one.

YouTube video editing as a service pulled 18 people and 16 rivals detected. One of them drew the boundary precisely: "I don't need to upload my footage, and be locked to their model, I can accept tool lock-in, but model lock-in, no." That is a buyer telling you exactly which lock-in they will sign for.

The trigger behind private photo storage is mundane. 6 people, 25 rivals detected, and the post opens with "I currently have 1.76 GB of space occupied in Google Photos." Nobody there is shopping for a photo app. They want a library moved.

Simple note taking tools shows 6 people and 15 rivals detected. The ask is one sentence: "I don't want something where am paying monthly subscription." Note-taking is about as crowded as consumer software gets, and the request is still open, because the request is about billing.

A vendor changing terms produced the NAS cloud sync ask: 5 people, 17 rivals detected. "I've used Sync's Unlimited cloud storage for a few years for my video production business and it's served me very well, but with them ending unlimited I'm up against a wall to backup my data." A happy customer became a shopper overnight.

Webflow alternative hosting has 5 people and 12 rivals detected. The reasoning is new: "Webflow in the age of Claude and Codex seems super expensive to manage and is limiting our output to keep it updated all the time." Coding agents changed what a site builder is worth to this team.

Two of the asks target one company. A Google Play alternative has 5 people and 7 rivals detected: "I'm pretty much at the stage where I need a new phone, and pretty much at the stage where i want to eject google out of my OS."

And a search engine without AI answers has 4 people and 8 rivals detected. "I want to move away from Google so my search engine is the first step. I tried Kagi but found out it's paid." Note the second half. The alternative was found, then rejected on price.

The pattern

Six of these eight sit in our crowded bucket, where our scan finds more than ten rivals. Pool-wide that bucket holds 734 of the 1,083 gaps. Crowding is the normal state of this dataset, which we covered in crowded does not mean solved. What makes an exit ask different is that the product already exists and the switch still does not happen. The blocker is import, export, pricing shape, or trust. That is the first thing to build, before any feature comparison. Several of these overlap with the storage and privacy demand we covered in the state of self-hosted software demand.

Honest limits

Our rival scan has a recall ceiling. A rival count of eight means our scan found eight. It does not mean eight is all there is, and we never read a low count as an empty market. The people count is distinct public posters. Four or five of them asking is worth a week of research. It is no forecast of revenue. Each quote is one post, picked because it states the reason plainly, and one person does not speak for the rest of the thread. We pulled only gaps with four or more askers, so this sample is narrow on purpose. Dates are when we recorded the gap, not when the poster wrote.

All numbers from GapMine's live dataset, same pool the site renders, as of 2026-09-25.