Hacker News Reader

Comment by jmacd | Hacker News Reader

I recently worked at a company that adopted Notion when it was a startup but has now grown to 800'ish people. I can say for certain that Notion does not scale.

The primary issue is very straightforward, but one that plagues so many companies: search.

Searching in Notion doesn't actually exist. It will perform some string matching based on some text you provide, but it is unable to provide any of the needed context in order to decide which result is the right one. It is also aloof of whatever structure you might be trying to impose, so when you are searching you truly are looking for a needle in the haystacks.

View on HN

Search is also obscenely slow. I type a few characters, and then have to wait several seconds to see the note I need. Compared to its competitors that frontload the entire graph, it’s unusable.

We've cut server-side search latency p50/p75/p95 in half compared to January, but still have a long way to go. On-device caching/indexing/searching is something we want to do more of, but currently we're focusing on server-side speed/reliability/relevance. We have a lot of work to do!

The fact that Notion’s search was so slow and not at all useful is part of what pushed me to Obsidian for knowledge management.

Obsidian's search is also bad: it doesn't prioritize file names or headings over content.

When you get past a certain number of notes, this becomes an issue. You might be trying to search for a concept you have a dedicated note of, but your search results are drowned out by normal usage of the word in other notes' contents.

Agreed, the irrelevancy of results has been pointed out several times. It just uses simple string matching, but the issue is that most of the tools do that :/

One nice thing about the Obsidian model is that it's pretty easy to build external interior with it. I previously wrote a search engine similar to fzf but for full-text search in a directory of Markdown files, and you can hook it up to open Obsidian: https://github.com/cgamesplay/pilikino/tree/v1

It’s not great but at least it’s fast and shows me more context in its results. It’s a more effective search for me than Notion.

I use a quick switcher alternative to ameliorate this exact issue: https://github.com/tadashi-aikawa/obsidian-another-quick-swi...

Makes obsidian so much better.

Every project management software grows until it has its own query language then promptly collapses under its own weight paving the way for a competitor.

I stopped using Notion as much when I discovered the magical solution that is Google Cloudsearch.

Any member of a Google Workspace team can navigate to cloudsearch.google.com and it will search across all personal + shared Drive documents, videos, slides, Gmail, etc. It has saved me countless hours trying to find documents.

This. This is the only reason why I don't use notion as my primary note taking platform.

The search is just not native and doesn't work in its current format as well as easier transition into various pages through it.

In what point it doesn't scale? When you get to 800 people?

Search in Notion is problem even without a lot of people.

My old job had a google search appliance that searched their internal documentation. It actually worked. I miss it.

Reply on news.ycombinator.com