GitLab Analyze:Product Analytics UX meetings, 30 Aug 2023

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: 2023-08-30 Product Analytics UX/PM Sync

Description

Today we talked through:

1. Tracking for Dashboard Designer and how it'd be a great dogfooding opportunity for funnels! - https://gitlab.com/groups/gitlab-org/-/epics/10939
2. What the UX for the `user_id_type` should be, on or off by default and why. - https://gitlab.com/gitlab-org/gitlab/-/issues/421328
3. Priority for scoping the Usage Quota work is below current milestone. It's a nice to have in 16.4. - https://gitlab.com/gitlab-org/gitlab/-/issues/419407#note_1499123178

A

The product analytics PM, ux, weekly sync for August 30th, almost said July, it's August, 30th 2023.. We have a couple of things: the standing agenda for the ux team front, end team and I wanted to review any questions about upcoming implementation, specifically dashboard designer. It's not an issue or an effort that is in the current milestone.

A

Rob I, don't know if you've got anything assigned to you for this one. It looks like yawn and a winner taking most of it or have the issue design to them so far, but are you gonna? Are you looking to pick anything up here? Is there anything I can answer or make more clear um in what you've.

B

Read no as far as I know, the I mean the tracking for dashboard designer is in the backlog. um It's not been allocated in one yet, but that's more of a back end orientated task.

B

um The removable panels. um Ellen's picked up he's just working his way through it, and then we've already had that conversation last week regarding the dashboard designer NBC in the direction we're going to take initially um and I, know yarn's all over that.

B

um So I don't see any other issues, I'm, not sure if any there's other than rolling out the feature flag, I'm, not sure if there's any other related issues, at least for this epic is there.

A

um Alan picked up the issue yesterday, which I should put into this epic or filtering out not product Analytics, visualizations,.

B

Oh you mean like Dora and Etc, that sort of stuff yeah.

A

uh I'm gonna add that into um since we did assign it I'm gonna. Stick it into that milestone.

B

That should be fairly straightforward, though that's um it's not a UI change. That's an engineering, filtering change, yeah.

A

And that would be an interesting iteration when we get to dashboards for other areas of the product um where you want to pull in data from say, error, tracking and product analytics or body stream dashboard. How do we start to make a menu of sorts of those without blowing it up into just this huge list that no one can grok and you're Paralyzed by choice.

B

A

That'd be a fun uh fun task for the New York person when they get here.

B

Yeah I mean the most obvious choice is to break it down by feature or group, but then you do that you're sort of siloing. That's not necessarily what those charts mean. That's just where they're situated yeah. So that's going to be an interesting one. Yeah.

A

Specifically is group data, whereas all the product analytics ones are project data today, so there's a natural split there, but then, within that we got to start to figure out. What's your job to be done? Who are you what's the right data to present um yeah and we can go down that path later, but yeah, I, guess I.

A

Think starting by splitting group and project is a good good way to split it, because otherwise you're going to add a chart that you're never going to get data for potentially, yes, yeah cool um I want to go back and talk uh tracking for dashboard designer real quick. You mentioned that's a back-end task, um The Proposal, which is very light that I wrote It's a snow plot tracking. Is that something that we can add from the front end.

B

A

B

Yes, yeah we can, we can um it's just normally the tracking we end up doing. Is you know, service ping, snow, plow orientated from the backend point of view, um thinking about it rereading but I I didn't read the issue um reading it now, saving could be done on the back end through the graphql API, but I mean at the same time, on success.

B

We can easily trigger a snowpile tracking event on the front end too and when it comes to visiting the dashboard designer uh most likely, that would end up having to be the front end because that's uh router as a view router, which means the back end's, not even involved.

B

So, yes, that is a problem I'm actually going to update the label to make that clear, so I'll add the front end label to that.

A

Oh and I added a extra bullet um for a stretch of users who edit the dashboard, not knowing. If that's going to be the same sort of view, components there, we can differentiate between I've created a dashboard versus saved an existing dashboard.

B

Okay, so users, who save an existing dashboard uh yep? That's that's the exact same process, essentially cool, so we can do both possible.

A

B

Of priorities, what would be the uh how soon or what? How soon do we want that done? I, guess that's what I'm trying to say in my convoluted I was trying to think in terms of Mark in terms of our group Milestones with beta Etc yeah.

A

um All right, we're I'm thinking about dashboard designer functionality he's separate from beta, so we.

B

Can release the beta.

A

And then roll out the dashboard designer, um so when we get to it great, we do have uh service, pin tracking today on New dashboards created. So we have some data already, but we'd love to see a little more real-time, the snowplow data, um how many people are going to and finishing up low I think this would be a great one for our first funnel tracking as well as we start to document that um how do people get through the flow of the dashboard designer? Where do they fall out?.

B

Is this okay is? Is this a case where we could actually use our own product to track our own product? It.

A

Is yeah analytics instrumentation is getting close to having the Mr merge for um at least huge views tracking, uh if you don't have cookie consent, so we won't get unique users um on getlab.com, so that'll be fun. Exciting, probably knock over the cluster repeatedly for a week and a half and then figure out how.

B

A

A

Awesome um so then reviewing upcoming design issues, speaking of uh Anonymous users or user ID type, that's the one that stuck out and I think it's assigned to a milestone that I wanted to see. If there were any questions there, um Ian and I had talked about in a previous call, making those filters. Eventually the tokenized filters like we have in other big views like jobs and Mrs, and things like that. But for now just adding a filter or a check box to exclude Anonymous I think gets us where we need to be.

B

Yeah I agree, um I did add a comment um just before the meeting started, whether we could consider where we could consider using the toggle rather than a chat box um just because it seems to fit the design parameters better in this scenario, because there's a binary that is always defaulted to one value um and has the uh added benefit that we auto update the dashboard as soon as you click it we're not doing any kind of saving or anything like that, so visually from a user experience perspective I feel like it makes it easier to pass and slightly larger.

B

So it's not so hidden um wording would be a bit tricky or I. Guess it's no different than a checkbox. Is it never mind wherever Lorena decides I'm acting with I think do.

A

We yeah there's been some discussion down in a comment below um about the verbiage or the wording and I think yeah I updated. um Let me I'm gonna update the proposal, while we're talking I'm going to share my screen while I do it so there's something to look at in the recording.

B

Although they're not pretty faces right.

A

Now speak for yourself, um toggle text exclude Anonymous users and followed by that um model ux um on true wait. How would that be so toggle I guess it would be yes.

B

Well, uh the if the terminal or the terminology is a checkbox that says exclude Anonymous users, then by default it'd be awkward in it because we're including all users and then, if you want to narrow it down or is that the or is the idea to have it. The other way.

A

Well, the the default for the SDK.

B

A

You're gonna get or that you have cookie consent boss, so all of these users would be anonymous.

B

A

So if this is toggle yes to default, you're, yes, excluding so you're going to show zero users and zero page views.

B

Yeah, so that would be no by default, wouldn't it yeah.

A

I know by default so that you see you're going to see what uh design saw one user, 100, 000 page views, then.

B

Yeah I mean yeah, you wouldn't be able to drill down into the users themselves, but you could see their paid shoes and their other date. Yeah.

A

That gets you around the I start. Instrumenting I, don't collect I, don't do anything with the SDK options. I don't collect a cookie. I have a bunch.

B

A

That's going to the cluster, we don't see any of it in my charts. So instead of troubleshooting you're going to see one user lots and lots of pages potentially.

B

If we want to have the toggle uh defaulted to on um from a point from a user from a ux perspective visually, does it make more sense to have that toggle on, or does it make more sense to have that toggle off you see if we want it on by default, we could just change the text to include Anonymous users instead and have a positive message rather than an exclusionary one.

B

This is where I want an actual ux designer to tell me: what's the under the.

A

Nbc the exclude, since we already got Lorena to give us a yes on it.

B

A

Have it off by default right or no by default, so don't exclude Anonymous users, because the probably the default behavior for somebody's instrumenting, at least with the browser SDK. First, that way they get data and we have a help. Our tooltip there yep about why they're, seeing what they're seeing I think that's the best case for someone who's new to product Analytics or our version of it and.

B

It follows the idea of I guess: Progressive disclosure, your the user, is actively turning this off yeah to change. What they're saying okay I can um create a quick mock-up of this as well. It I mean it's just a pajamas component, so I can just Chuck it on the page. So you get an idea of what it's going to look like if we want before we give the green light to move it into uh planning breakdown, I'm.

A

Sorry and then I think um Dennis hadn't asked uh for someone to create an issue for this Cube scope of work. For this, which I don't see a related I see a related issue around the feature iteration, but nothing for.

B

um Yeah I'll do that at a second time, I'll. Take this one excellent! Thank you, sir.

A

Cool and then the last one was asking to get an estimate on the usage quota. um I know uh Dennis picked everybody earlier this week as well. Thank you for putting it on your to-do list. Rob.

B

It's been on my to-do list for well over a week now, I'll get to the with fill parameters have taken a far longer than I was expecting or wanting it to so. um But yes, it's my next higher priority item. Normally, um when I meet with Dennis every week, we go through my items and just make sure they're all aligned um he's not feeling too great today, so we'll be doing that tomorrow.

B

Instead, so you can take it easy um I'll see if that adjuster could like me to Pivot away from what I'm doing now to focus on this. But if not, it is the next item that I'm going to be hitting before the end of the week.

A

Cool, so my preference would be finished up what you're working on for the existing well done. The commitments we've made future scoping work can always be second fiddle to that per.

B

Second, after that, and if anyone's watching this and wants to take this on and I haven't started or even if I have feel free to excellent short toes and all.

A

That I'll link to what we're talking about in the notes below the video too awesome well we're at the end of the agenda. um Thank you. Rob.

B

That's all right, I, don't think we've yet we've we're quite light on design stuff. For now, for now, let's see how we go.

A

Awesome all right well have a great rest of your day and you cheers ciao.