Fibery Product Report: October 2024

We’ve made five releases in October (overall, weekly schedule is very consistent now).

Total Highlights Score: 405 (for comparison, it was 770 in September).

Main areas we focused on in October:

  • Permissions
  • New pricing and payment system (out of scope of this report)
  • Space Setup
  • Search
  • Navigation

Permissions

Done

We’ve released the first version of Introduce Guests limited to per-Entity access. It took exactly one month since the kickstart which is great for a Feature involving so many people and services.

Now we’ll see if Guests in their current form are enough to solve the majority of external collaboration use cases.

Next

We have leftovers from the Features we’ve just released that we’d like to cover:

  • Finish the Guests Feature: fix non-critical Bugs left and finish the small sub-Features.
  • Convert Assignments into regular People Fields to erase confusion about why some Fields behave differently than others.
  • Improve Guests user experience: in particular, how sidebar navigation works for them.

The next big thing is Share Entity with Group: Andrew has already started his part, and we’ll do a proper kickstart this week. Share Database with Group would follow.

Adoption

Almost as many workspaces use automatic access for assignees as those who share manually. 50+ workspaces have granted DB access:

Simplicity & Usability

Creator usability improvements

We’ve mostly focused on problems related to Fibery workspace setup and streamlining the creator user experience. Major change was new Space Setup design. It caused some problems, but it seems we found good solutions to all of them, so in November we will polish Space Setup and make it shine.

Done

  • Space Setup 2.0
  • Convert relation Field: to-many → to-one
  • People Field
  • Settings Page: Integrations
  • Move Folders between Spaces

Next

  • Oleg decided to not move forward with AI for automation rules, but give AI Q&A a second chance (he will try to change the approach, so we will see how it goes.)
  • New Space Setup: Fields screen
  • Other Space Setup improvements around navigation and relation screen.

Rich text improvements

Sergey decided to make rich text better. First, he improved rich text loading time by 10-30%. Now the goal is to remove jumps in rich text fields and documents and preload all mentions: Avoid jumps when loading mentions in Fibery Documents. It should make rich text experience much better.

Adding and updating data

We’ve decided to fix all major problems related to adding and updating data in Fibery. In October we’ve released Quick Add. Other major features will include Required Fields and Batch Updates. I hope we will be able to complete them by the end of the year.

Done

  • Create anything from sidebar / quick add

Next

  • Required Fields
  • Improve Entity creation when formula is used for name (or some fields are required)
  • Batch edit Entities from Views other than Table

Sidebar

Suddenly we decided to fix Fibery Sidebar. First, we made it completely collapsible and added Cmd + \ hotkey, this is just a start. Main idea is to introduce custom Sections, so it will be possible to move Spaces to these Sections. Usual users will NOT be able to re-order sections in the left sidebar, but just hide them. As a result, it will be easier to enforce a proper structure of the sidebar for creators, but end users will have slightly less control on the order.

To compensate that, we will improve Favorites section: it will be possible to create folders in Favories and also add Spaces to Favorites. This any curious and savvy user will be able to fine-tune Favorites and just hide everything else. Less curious users may be happy with the default sidebar configured by creator.

Done

  • Completely collapse left sidebar

Next

  • Quick Filter Views in all Spaces
  • Collapse all spaces with a single click
  • Folders in Favorites
  • Add Spaces to Favorites
  • Sections (Special, Custom, Smart) and make Spaces order in the left menu predictable for all users

Search

We did some improvements for search and tried to merge semantic and keyword search results. It affected performance and now search feels slower for 2+ keywords. We also played with re-ranking algorithm and results were better, but it added ~0.5 seconds to results display time, and so far we disabled it.

  • Search in basic text/email/phone Fields
  • Semantic Search Results mixed with keyword search results

Next (maybe in December)

We are making a pause in Search improvements to think, but possible next search features are:

  • Search exact match using quotes
  • Semantic search setup simplification
  • Find duplicates for target (or any?) entity

Email/CRM Theme

This theme should be completed on October (at least MVP). We did two small features, and are working on a major feature that will allow to sync several email accounts into a single Email database. We also will implement some integration with data augmentation service, so it will be able to enrich Leads/Accounts with some information automatically.

  • Email Integration: Cleanup email body
  • Add Email URL Field to Email integration

Next

  • Integration with some data enrichment system
  • Multiple accounts for the same integration

Whiteboard

It appeared that bringing ProseMirror into Whiteboard is a very challenging task. We already spent 6 weeks on it and it is not done yet (but close).

In the meantime, some small additions were made here and there to the Whiteboard.

Done

  • New shapes: Pill, Quote and Attention
  • Holding Shift modifier key while moving an item on the board, should keep the item sticky to the X or Y axis
  • Whiteboard: Copying text and style to apply formatting to objects
  • When a whiteboard is opened and the user’s viewport is empty, automatically adjust the view to fit the screen if there is any content on the whiteboard
  • Preserve selected database in search popup for current Whiteboard

Next

We are still fixing basic things to make Whiteboard good. It is already much better and people noticed that.

  • Whiteboard: Text Editor floating Menu
  • Adding and displaying external text links on the whiteboard
  • Comments in Whiteboard (maybe in December)

P.S. Your questions and ideas are welcome, as always.
With :orange_heart:, Fibery Team.

11 Likes

Looking forward to many of these improvements.

Definitely noticed the reduction in speed. One thought that might help is ability to search by just one or more “Spaces” and select multiple “Databases”. This way, it doesn’t search the entire workspace every time but you still have the flexibility of searching more than one database.

Great report!

The consistent enhancements to our daily workflows across Fibery’s recent product releases leave no doubt that we chose the optimal solution for our company.

This conviction is further strengthened by the reality that the majority of planned future improvements directly align with the suggestions and ideas our team had previously proposed as valuable enhancements.

When the Fibery product roadmap so closely mirrors our own assessed needs and recommendations, it unequivocally demonstrates the development team possesses a comprehensive understanding of our requirements and is resolutely committed to addressing them.

Thank you to the entire Fibery team for your responsiveness and user-centric approach. We greatly appreciate the continued effort to evolve the product in lock-step with our evolving needs.

3 Likes

This is not a problem, major problem is general slowness of AI search. Now we wait till both searches return results and mix them.

1 Like

Possible idea: Split search results into two “panes” horizontally (one above the other). Each one is titled, e.g. “Fast Search” or “Basic Search” and “AI Enhanced” or whatever (you have better ideas, I’m sure, hah). Then show top 5 results from “traditional” search on top, immediately (as soon as they are ready), and populate the AI search below it as soon as those results are ready (show some indicator that it is loading in the meantime, e.g. spinner or whatever). Don’t mix them. At the bottom of each Top 5 you have “show more” or “expand” that will expand that specific set of results into the current infinite scrolling list (and at the bottom it still shows the “AI Enhanced” bar below it, clicking that will expand those results to top 5 bringing it back to default state maybe). Perhaps also figure out some way to optionally enable “mixed” results, e.g. a default-collapsed “Combined Search” section at the top that you can click to expand, and hopefully by the time you click it the AI results are ready. :smile: I guess maybe there are complications from interaction with keyboard shortcuts or other issues, but maybe it’s the seed of an idea.

Btw Claude is great for quick mockups! 60 seconds to create the prompt (based on my description here), provide a screenshot of current search box UI, a little directed refinement, and within 2-3 minutes boom:

Bottom line for me is that the recent “enhancements” to date have not been an overall improvement. I am hopeful of the “best of both worlds” because I didn’t use AI search very often before, partly because it was a whole different “mode” that I had to enable. If I could see the top 5 results of each fairly quickly I think that would be quite helpful. The main benefit of AI search is to bring up a relevant result much quicker, higher in the sort order (when it works, for particular searches), and I can usually evaluate that with the top 5 or so results. If the AI’s top 5 is not clearly closer to what I want, the traditional search results are more likely to be worth scrolling through beyond top 5 because the AI’s results tend to digress further the more you scroll (in my experience). Plus the AI results are loading while I am evaluating the relevance of the top 5 faster traditional search results. Similar principle as loading a skeleton of the UI before you are able to load data, makes it feel faster, and in this case genuinely is in a functional and not just psychological way.

2 Likes

Oh yeahh :sunglasses:

I was actually thinking this!

Or maybe split horizontally.

1 Like

I would have suggested horizontal except I am biased from being on 4k all the time and I’m not sure the majority’s screens would ideally support the width. But if so, great! Side-by-side would be better for UI/UX IMO.

1 Like

I’d opt for “responsive” :sweat_smile:

2 Likes

CSS please. :face_holding_back_tears:

1 Like

A lot of our people are remote and mobile, working on laptops, so a vertical split would be better for most of our team, whereas a side-by-side view would create too much clutter and would make it too difficult to see all the data (too much left-to-right scrolling). I imagine the horizontal split would only work on larger screens if the search pane was the only pane open. If search is opened in a second pane, then I imagine a vertical split providing a better UX.

Two areas for results is not an ideal solution, since there will be two scrolls, no way to use keyboard and other usability problems. We will not go there.

3 Likes

Fair and reasonable. But still… think outside the box :grin:

About requires fields, I’d suggest to target a wider approach where each fields support 2 formulas: one to drive show/hide state and the other to drive required/optional state
That would deeply enhance flexibility, in particular to complex scenarios.
Thanks a lot for these monthly updates!

1 Like

Can you expand this idea with some use cases? For required field there will be just a checkbox in settings “Required”

Oh we always love ideas!

1 Like

Yeah. Share item in menu is not good enough. I found workaround where I create smart sections - Tasks, Projects, Clients and Guest see only items that are shared with them in these sections. Problem is that they see all the smart sections even when empty for them.

1 Like

This is exactly what we will fix, improve and simplify

2 Likes

Yes: in many cases fields can be dependent on values of other fields, both in terms of being required or not and meaningful or not. For example: an entity can be a task that can be repetitive or not, a field “is recurring?” drive the appearance of period of recurrency, that is not meaningful if the task is not repeatitive so it must be hidden, while it is meanigful and required if the task is repeatitive.

1 Like