GitLab Working at GitLab, 1 Sep 2020

Previous Meeting

Next Meeting

⏯

youtube image

►

From YouTube: Practical Tips for Handbook Editing_Leveraging Google Analytics_2020.06.12

Description

Shane Rice (Manager, Inbound Marketing) shares tips and tricks for analyzing site traffic in the GitLab Handbook. Find other helpful tips at https://about.gitlab.com/handbook/practical-handbook-edits/.

A

B

So I can only play like 15 seconds of this, or else we'll get like a strike.

C

A

One of our school videos got flagged because it had, I think, the awards dinner, because it had some stuff in there like.

D

Background music or people coming on stage or.

A

It was one yeah, some like some pipe hype up music to celebrate and stuff right, and it was it was it got rejected by youtube because of that. Oh.

D

Jeez so youtube algorithms.

A

Yeah, that's pretty it's impressive that it catches that right, but it recognizes the sound all right thanks for everybody joining today, we're excited to have shane rice as our guest speaker, so shane a little bit of background. This uh we've formed this group and have been meeting regularly to talk about just some practical tips for using the handbook, and so we've got even a practical tips, handbook page that monica helped us create and uh emily's been guiding us through and some other guest speakers of hey how to do this. How to do that?

A

One of the topics that's come up actually a couple of times now in this series is: how can we, uh you know, gather some insights into traffic and visits to our handbook, page and understanding.

A

um You know how many clicks we're getting that kind of thing, so we'd love to uh yeah reached out to you to just give us an overview. I know you've got access to some of that information and you take us from here I'd love to share and then we'll ask ask questions after that.

B

Yeah, great thanks david. um So what I'll do is I'll go ahead and share my screen and just walk through some of what we have in the google analytics application directly and then we'll talk about some tools we have outside of this. We can, you know, get some of this data into a tool called data studio, some places where it can be more publicly available so that you don't have to have um an access request in to be in the google analytics tool to find this and.

A

Shane started: can you tell us yeah who you are what group you're with? What do you do in your day? Job? Where are you located yeah.

B

I'm gonna definitely do that so yeah, so um I'm shane rice. I work in marketing, I'm the manager for inbound marketing and um I had to think about that because it changed this week, but it's not bad. I like it and um so one of the things some of the things I do are um you know help focus on the strategy for organic discovery, so make sure people can find us when they search for uh related terms. uh You know git lab ci.

B

Those kind of things um then measure our success in that, and so we use google analytics to do that on the marketing website, as well as docs um forum customers and those all have. The google analytics cookie firing when someone visits and then also um conversion testing, making sure that we're optimizing for conversion. um You know converting people to hopefully leads um as they visit our site, so um I could have stopped sharing my screen when I did that. So sorry,.

C

B

So diving into google analytics. If we go here, uh we can look at the behavior for specific pages in the tool here. So what I can do is since the handbook you know it lives in a very specific folder. We can just you know, just look for anything that has handbook slash handbook, slash in the url, and so this is just for this week. Let's expand it out and look at say may is a month.

B

All right, so we can see that for the handbook we had, you know 301 000, page views, um 223 of those were unique. um I can tell you this spike right here. I made a note of this in the tool here we had a temporary spike in page views for about four hours. One day when there was a problem with some redirects on uh the about.getlab.com site. I was constantly refreshing the page and so page views went up during that.

B

That's kind of a rare thing, though um so anyway um time on page really useful for the handbook, because it tells you if somebody's engaging with the content, are they hitting that page and you know reading through? Are they hitting that page and bouncing? um You know? We see it um to give you an idea um on the about site. It's about four minutes uh is how an average about how long somebody stays on a page. um Docs is more like eight or nine.

B

So docs is a lot more like you know, in-depth coverage like you know, going into how to do something, and so people are like really staying on the page, going through all the steps um so that the time on page is a little higher there.

B

I would expect the handbook overall would probably be a little bit higher than um about on average, um because you know it's used so much internally and people are going there to reference, but it could also be that because people are going there to reference, they go find what they want and they're bouncing back out. So some interesting things we could dive into there so entrances. This is.

B

This is where you know we find people entering the site, so they're entering the site and going to a handbook page first, um so I personally the way I search the handbook is, I use google, so I do a site, search where I say, site colon, about.getlab.com handbook and then space whatever term. I need to find, because I find that helps me find what I want a lot faster.

B

So I enter the site a lot of times in the handbook.

B

um Bounce rate, so how often does somebody hit this page as their first page and then immediately leave the site, so they can stay on the page for a minute whatever, but if they don't go to another page, they bounce, so that's what's counted in bounce rate and then percentage exit.

B

How many people leave the site from one of these pages and then so page value? It's so what I do with page value is I set any event that we want to track to the value of one dollar and then anytime. Somebody takes that event, like you know, doesn't have an event on a page. Then we get a percentage of how often people are interacting with that event on the page, so this is kind of you know very high level.

B

What you can see in google analytics, um rather than just you know, continue to guess at what y'all want to know. Are there pages you want to look at? Are there specific things you'd like to know if we have in google analytics what questions can I answer for you.

A

Yeah thanks shane uh one one would be to start in the sales handbook, so handbook, slash, sales, okay and, and this what you're, showing here like is, is only was this only shane has access to this or how would like our team get access to this information when we think in the future about something that we want to take a look at.

B

Yeah, that's great, so um no! This is not something only I have access to. This is widely available on the team. um You know if you need access to google analytics, um you know you can put in your access request. um I find that one of the things that's challenging for with google analytics is, if you're not using it constantly, it's really easy to pull up a report one time and then not be able to find it in the future.

B

You know you've got that number and you know what the number is, but you don't know how to get back to that number, and so um I like data studio for that reason. So, um while I'm sharing my screen I'll share data studio- and this is something everybody if you have a gitlab.com email address- you have access to all of these um reports that I'm I've got listed here. um So let's look at self-serve um metrics here, so I can go to a page and I can say handbook slide show that's not working, of course.

B

Well, when it works, you can go here and you can type in handbook sales, and it will tell you, um you know, only give you the urls that have that included. So you can deselect all these urls and you can say, slash handbook, and then you can select all these.

B

And now, instead of looking at you know the stat the sites across the uh the domain you're, looking at just these pages for this time period, um obviously there's something I need to do to fix this. These are kind of mvc, so until somebody says hey shane, this is broken. I don't know it's broken. um So if you find something broken, please let me know um I'll add a link to this in the um agenda as well.

B

B

C

And so david, what would be some use cases for when we might want to access this analytics page.

A

You know one from like a communications perspective with monica's she's looking at uh field flash. I think mailchimp already has some that baked in here, but if we just wanted to see just traffic in general to understand of things that she's promoting and communicating out, I think just in general.

A

If we want to understand of all the pages underneath the field enablement piece, you know how much traffic or each of those getting and if we find some, for example, just not getting any traffic like we can say all right well, should we kill this all together? Is there a better way to organize it?

A

um This is just a couple but yeah open that up to the team.

D

Yeah I'll, um if I can I'll piggyback off that david, I think what you said is spot. On the other example, I'd give of what you outlined is is wow. This page is getting a ton of traffic. Let's actually spend a little bit more time, making sure that people are getting the best experience here. So let's say you're talking, you have a page about pen sales and why clicky pens are better than screw pens like maybe the traffic on it's so great.

D

You want to add a video and maybe create more materials to support it and kind of judge it up, based on the fact that it's being used well.

C

E

Thank you, and- and I would say too, as we think about I know like melinda on james's team- is- is kind of spearheading like reworking the sales handbook, but as we think about that as well, we can. We can start looking at okay, like what are the. What are the pages like?

E

What's the content like getting the most traffic um to david's point, and then we can basically like look at how those pieces are laid out like what's the what's, the format of those like what's working well um and then go ahead and like use that as like kind of a baseline, so reiterating didn't really say anything new there, but just kind of thinking out loud and reiterating what david and emily said.

D

I actually have a question for shane, so um let's say someone lands on the slash sales page. I know you talked about entrance points, but is there a way to figure out what was the page that brought you to that page specifically,.

B

Yes, so google analytics has a limited ability to tell you like you know, one step back. Basically, where were they at before they visited this page? um It doesn't give you a full funnel.

B

There are some like behavior flow visualizations in google, analytics that a lot of marketers like, um but I don't find them to be incredibly useful because the data there is often sampled and not really, um it gives you a good visualization, but it doesn't really give you a sufficient depth to understand where people are coming from so um sharing my screen again, I will just show what this looks like for the page, we're on now that what we're looking at so we can add a secondary dimension to say previous page path,.

B

And so yeah, so you can see here, like you know most the highest, you know um previous page is an entrance for the sales main sales page in the handbook I mean you can see. You know some something's happening here where the page is getting reloaded and that's why we have the previous page path matches the current page um that you know the people are on the business. Ops data team kpi index hit sales from there.

B

um You start off in the handbook index page command of message, field operation, and then you know so we can. Even you know, enrich this further. So you know hey the entrance data is not really something I care about. I want to see pages that people are on, so we can actually exclude that we can say.

B

Oh actually, I don't have it's already in the previous yep cool. So if it matches entrance, we don't want to see it and then apply that.

B

B

B

There we go so now um you know it's the text. Wrapping is weird, um not much. I can do about that, but you can see here. This is a list of you know all the pages where somebody entered or hit the next page, the page we're actually looking at in google analytics.

B

So this is a little a little messy because we're looking at all the pages that are handbook sales, so we can click in and we can go to say just the sales index page and we can look at previous path just for that page. So let's do that same thing again on previous page path and then we will hit advanced, add dimensions. Previous page exclude.

B

I know it's like riveting content watching me talk to myself. While I do this, so I appreciate y'all hanging out and you know you're all laughing on you I want I want. I want to be on the record that everybody's laughing on me.

A

Well, you provided the uh awesome intro music before so. You know you already set the bar high.

B

Cool, well I'm glad I I set the bar high. That's that's I'm going to the weekend on that high note. Yeah.

A

B

So yeah, so here you know this is this is every every time somebody hits? You know, handbook sales, that's not an entrance, here's the page they were on before they did that.

D

This is really cool to see. Some of the things that jump out to me are like the sal job. Family is number four there. So it's interesting that that's sending a lot of people there, the other the kpi index is number two.

D

My guess, if is that that's going to change now that the kpis aren't on the sales like the slash sales page anymore, it would be interesting to see if that drops over time, um and then that search seems to be kind of the number five way that people land there, which is um interesting anything jump out to you, david or monica or kelly, or michelle.

A

I'm also surprised about the the sal. You know job family link um commanding the message yeah near and dear to my our hearts. So.

C

A

Interesting yeah they're, going from there to back to the sales main page.

C

I was thinking the same thing, david.

D

I wonder if what do you think about? um I wonder if one of the actions you could take with this would be to look at that command of the message page and see what links are sending people back to sales, because you know we without you know, click tracking. We might not know what links are sending people back to the sales page, but it's probably worth going through that page, seeing what people might be going back to that sales page for and seeing if that's information that should be on that command of the message page.

A

D

Don't know just a thought.

A

Yeah I'm like going there now just to say I'm at the top of it yeah. You can navigate back.

D

A

E

I'm also trying to so so I'm grappling sorry, sorry, shane, um I'm kind of grappling with two things. So one thing um I um sorry I'm trying to figure out how to like.

E

So I think the sales hand like the sales home page, is super long and in talking to melinda and she's the dri on this, I'm you're just kind of getting it secondhand, but in initial conversations with her and thinking about how she was going to approach kind of refactoring that home page, we talked about the fact that, like it's just super long like there's a lot of information, there's probably there's definitely a need to break out that home page into sub pages um and have like a cleaner um homepage and almost like directory.

E

And so when you think about like that use case. What like what comes to mind for me, is like looking through. Okay, all the links on the home, page and figuring out like what are people engaging with the most. Is that like, where you would kind of start to figure out? I guess like what content on the homepage people are going to, or I'm not really articulating myself that well, but I'm trying to kind of think out loud while thinking it's really how this data can help us.

E

I guess prioritize what people are finding the most helpful on that page and go from there.

B

Well, uh I think, there's another marketing tool used that can be helpful with this and that's hot jar, which is basically what it does is it can record sessions. It can also do click tracking. So, like you know, heat mapping, a page to say, like hey, you know people are scrolling down. They are you know, um they're clicking on these links, those kind of things um you know like I can tell you, like you know um some of the things you do. That too is like you know.

B

Pages are valuable, but you kind of want to see what's secondary in value like you know, pages that maybe need to be highlighted in different places or something like that you can exclude. You know certain. You know parts of the page from your test, so, for example like when I test anything with nav like if I want to see you know like what's going on with the nav, I have to exclude sign in and register well, not really register, but definitely sign in because sign in is like always like 97 of the activity.

B

But if I exclude that, then the other things become way more visible. So um that's a tool that we could, you know, talk about finding a way to implement for handbook testing, and we haven't done that to this point, but that would definitely give you some of that. You know more page depth as far as you know, what's happening with certain links. How deep into the page are people going on average? um You know those kind of questions um you know we could.

B

We could set up link tracking in google analytics, um but that would be a really kind of heavy lift, not lift but like it would be a really heavy solution, because now we're creating events for every time. Somebody clicks a link um and that'll be a lot.

D

5000 handbook pages now.

E

No, that makes sense- and I was going to the my second question was going to be you know, is, is there value on the google analytics side? Is? Is that kind of time on page metric like a good place to to look out for some insights, or is that am I kind of hearing like hot jar might be a better place like holistically to to look at? You know, like you, said, behaviors on the pages and then like how people are are yeah, I'm I'm david, I'm gonna stop mid-sentence like we talked about earlier.

E

Stop my thoughts on that.

B

So I would say that time on page is a great metrics to kind of group pages together. You can say like okay, so based on what we're seeing here like this is above average, so these are pages that are outperforming. What are we doing here? That we can? You know, do in other places, um I would say hot jar is a lot more for, like you know you, you have specific questions about a specific page, um whereas, like you know, google analytics is really more like groups of pages.

B

What can we identify from these different groups of pages and how we can use this data to help us improve those other pages, because I will tell you, like you know, remember how we were saying. The um the handbook in general. The average time on page is like a little over four minutes like for the sales page. It's like 5 15, which I just noticed. While we were, you know having this conversation, so you know on average, people are spending more time in the sales part of the handbook.

B

There are other places, but even there you can see. Like you know, um process version check right. The number two page people are there: six and a half minutes. um We can, you know, uh sort by this, so we can see you know longest pages and we can. You know refine that. But you know like territories right like people are spending 12 minutes 12 and a half minutes on the territories page, like you know, like.

A

After it's interesting.

B

Yeah, on average, let's add a 2 000, uh what 2191 unique page views right so, like you know, that's um you know a lot of time on that page now. People need to be leaving the tab open right.

A

B

One person can leave the tab open that raises the average, but that's really high for that number of page views um and that's really the the biggest outlier I'm seeing right now. um So that's interesting.

E

Yeah and even if they're leaving the top open, I mean that still tells us like it's a resource page right. So even you know it still means something. Even if it's skewed a little bit, I'm probably guilty of leaving. I have like five handbook page tabs open as we speak that have probably been open for, like three weeks.

A

Is there can you um subscribe to a report that just gets generated on some kind of regular cadence.

B

So google analytics has a customization area where you can create dashboards and reports there and then have those sent out um on a you know on a regular basis. It can be weekly daily. Whatever you need, um so it's possible yeah, it's definitely there. I don't use it a lot um like I said I feel like the reporting. The customizations in google analytics are just not as robust as you can get in data studio.

B

Data studio is a lot more flexible and what can you you can report and and how it can be visualized and those kind of things, so data studio is my preference um but yeah you can hold on I've. Never looked at that. Can you okay? So, with data studio, you can schedule email delivery with that as well. So that's an option with that tool which would be good.

A

Yeah, I know we're almost the time I mean for me. I think there there are some questions about just what's available those kind of things, I'm always in the opinion like. Let's not look at data just for data's sake. um The biggest thing is like if you've got something where uh there's some kind of actionable insight. That or a hypothesis that you have that's gonna inform some other action.

A

Maybe there is like coming up with hey, I'm really wondering about user behavior data. What is what can google analytics or data studio or whatever tell us for there so really be targeted with the insights that we're trying to gain but um yeah? I I discourage us from just trying to go. Look at all the data, that's there, because that you probably spend all day.

B

Yeah, you definitely want to start with a hypothesis, like you know, people that are spending more time on average are there for this reason, and based on that, we see this trend and based on that trend, we want to analyze these pages. um You know, um I agree with you. I don't I don't like to just say like hey. This is great that this page is doing this like, let's you know, do everything like we do on this page, like that's, not really um the best way to use what you learned from this data.

B

I would, I would say anyway, yeah.

A

Thanks so much shane, I think we're at time, but any kind of final questions or comments. Any questions, questions from the team we get.

D

Outro, music.

B

Do you have a request.

D

B

um Hold on, I gotta find something now something awesome. uh Let's see wayne.

A

Some johnny cash.

D

We're going to get a weekend babe.

C

Yeah, that was.

C

D

Happy family, okay,.