-
Notifications
You must be signed in to change notification settings - Fork 1.4k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Epic - Nail Seamless Analytics Experience #6513
Comments
Love that you put this together @mariusandra! Happy to help with the IA refactor if needed. Can you clarify Goal 2? Not really following. |
Can we include #6172 in our new chart options. |
This is an ambitious undertaking. The insights UI is actively changing in other sprints as we consider taking this on. I think we should still go for it, I'm just calling it out as a complex problem. Please include me as you are able to - I have a lot of in-progress unification work that hasn't been shared too broadly, yet. You can see those at https://www.figma.com/file/9yWtngNb1AIuf6KmXaEPJA/?node-id=827%3A137854 |
@clarkus While slight UX changes can definitely be part of this (and you will be looped in, 100%), the bullet point was more to address the fluidity of navigating between the scenes. Not necessarily where the buttons to go between the scenes are, but to get rid of the loading indicators and make the experience fluid. Not "the insight zooms in from a 3d cascading dashboard and then subtly tilt-shifts in a futuristic minority report UI way" fluid, but a snappy "everything is there instantly" fluid. However, streamlining the UI and design between the pages is going to be a closely second undertaking, somewhat of an extension of the IA task. @paolodamico we could definitely use input in metrifying this goal, and I'm also fine with skipping it. Basically the idea is that currently there are so many things that don't make sense in the overall flow between the pages, and we want to improve many of them. Things like "why the F does it load two times?", "why the F does it say I need to send events to PostHog when the chart just hasn't loaded yet?", "why the F is there a huge 'view' button, but I can't click on the insight title", etc. Browse the pages for 5min and count the "wtf/min". |
That makes sense @mariusandra! In addition, @clarkus to make sure we achieve our goals, and we don't present a completely new product in a single release, I think we should only do the basic Information Architecture improvement standalone, and then work out improvements. Thoughts? Re metrics. The kind of improvements we're looking at will be hard to quantify with anything other than usability tests, but a percentage change will be very hard to measure (what would our baseline be?). How about, "have a maximum of 3 WTFs in a single insight/dashboard session" (we can ofc run several sessions and average). I think this will really help push our quality forward. |
@paolodamico do you think we should test some of the IA changes before committing? There's still a lot of in-progress feedback there. I think we can get the design locked down by mid-week. I think there are some good ideas in here, but it is a lot of change. Once we get this locked down, we should try to work on a communication plan for users. We will also need to look at our support documentation to ensure it still aligns to the latest IA changes. |
Re @clarkus. I think the best approach here would be to build the design we have (almost) finalized. I'm worried it'll be hard to get actionable feedback just from Figma (as it's mostly a visual change and you feel the real impact when using it).
Also pinging @Twixes here for input on the approach above, as it'll definitely have some technical implications keeping both versions. |
@leggetter brought up a need today in figma about making dashboards easier to find. We solved this in the saved insights work, but haven't really done much to bring dashboards up to the same level of "easy to find". As part of this work, can we tack a simple keyword search on the dashboard listing page? It can literally be the same component used to find insights. |
Cool I can put together a few quick prompts that encourage users to opt into the new design, but still makes it easy to opt out if there is an issue. Will continue on that in the main figma for the IA work. 👍 |
Things identified for "saved insights" to spark joy. Card view saved insights
List view saved insights
General
New Insight
Insight table on Insights
View mode
Dashboards
|
@mariusandra worth considering #6668 as part of this effort? Tooltips can be pretty critical to validating the insight is done and ready to be saved. |
Yes, that's exactly the type of "annoying, usually not important enough though rather quick to fix" usability issue that we should fix as part of this initiative. Other tasks such as:
.... may or may not also be part of the next release, depends on how far we'll get in the next 2 days, how much muscle we'll have for the next sprint, and what will still need to be done. The list above is just taken from searching "insight" in issues, not a definite list of priorities. Please keep the issues coming that you think make sense to be here. I'll aggregate a bunch more tomorrow. That's why I also posted the "saved insights" list here not under its own issue --> so we have one huge list of things to do in one place. Goal 1 of this epic is almost done, goal 2 to decrease the "wtf/minute" count when using insights is still in progress and will continue next sprint under the proposal goal for next sprint of "Get the UX to a high quality state with new Information Architecture ready for Product Hunt Launch". The more we document the scope of the problem this week, the better we'll be able to solve it next week. |
Turbo mode todo:
|
Another big improvement I think we should do here, when refreshing an insight if you're on "edit" mode you should stay in edit mode, it's very confusing to get taken to the view mode if you just refreshed (probably also because something went wrong). |
Fixed another bunch of small things with saved insights. The big issues to tackle tomorrow are:
Postponed:
|
Also, we might need to prioritise fixing our slow-loading tables as well: #5361 This is visible when you use turbo mode. Whenever you open a page that has a huge resizable table (events, feature flags, etc), the page takes a lot longer to load, compared to other pages with huge non-resizable tables. There's something really inefficient happening with these tables and it should be at least looked at. The video above is with a M1, and the problem is probably much worse for folks with less single-core performance. |
I'd vote for solving tables later. They're a critical component in the product and there isn't a ton of variation across instances. I am solidifying some component designs for these now. It might make for a great follow up issue to make tables awesome everywhere™. |
Related #6771 |
This issue has 2012 words at 18 comments. Issues this long are hard to read or contribute to, and tend to take very long to reach a conclusion. Instead, why not:
Is this issue intended to be sprawling? Consider adding label |
@mariusandra I found an issue with turbo mode. In some cases, depending on how you get to your target location in the product, you'll experience the top offset not being reset for the new context. An easy way to reproduce - go to the insights list page and scroll down to find a funnel in some form. Click on that and once the insight loads you'll see that you're already half way down the page looking at correlation features instead of the header of that insight. tl;dr - we should reset the top offset when navigating across features. |
Later / whoever:
Even later:
Insight table:
|
We've done a lot of work on this front. Still remaining tasks:
What else is out there? What's outside the scope of this issue/epic? What falls under the IA epic? @Twixes @clarkus @paolodamico
|
There's a few UX-y things that we could still do, not high priority but to be considered:
|
|
Alright, as agreed, did a full rundown and here's what I found. We may want to split some issues into their own scope: High priority issues
Low priority issues
|
One more potentially quick fix for this effort #7269 |
Another idea from watching recordings today. We currently sort the event/property list alphabetically but most of the stuff we show above the fold is not helpful (e.g. almost no one filters by Anon Distinct ID), instead we could sort based on relevancy (i.e. filters used most often) |
Amazing list @paolodamico . These all sound like important things to work on. A few more issues that were just reported that fall under this umbrella:
Edit: also |
A simple change would be to modify the labels for favorites to specify "Your favorites" or "Add to your favorites" to better communicate the scope of favorites. @mariusandra |
Split out the search bug into its own issue, #7368. Will update the board so we can use it to track this epic instead, as the myriad of comments have become to complex to handle here. |
Alright, this issue was getting pretty unwieldy, so I created issues for outstanding things, added them to our board in a special column and we can now close this. |
Expanding on this sprint planning comment
Nail Seamless Analytics Experience (Working title)
Owner: @Twixes
Why?
Why? Insights and dashboards are very sluggish and inconsistent to use. The app doesn't spark joy.
Goals
Goal 1: "It should be possible to go from a dashboard to an insight and back in 0.1 seconds"
Goal 2: TBD, working draft: "80% decrease in WTF/minute during usability surveys."
Steps
P1, fix before other stuff. Persons modal loads forever on trends @mariusandra
Finally release saved insights @mariusandra
Implement the Information Architecture flow and refresh the look @Twixes
Combine the UX of insights, saved insights and dashboards into a coherent whole @Twixes and @mariusandra
Chart UX improvements:
The text was updated successfully, but these errors were encountered: