GitLab Growth, 12 May 2020

Previous Meeting Next Meeting

⏯

youtube image

►

From YouTube: GitLab.com Signup Form Conversion w/ Sam Awezec & Danielle Morrill

Description

A conversation between Sam Awezec, Senior PM for Growth: Conversion + Danielle Morrill, Senior Director of Growth Marketing

A

Hi, this is Danielle I'm here with Sam and I'm learning more about our new user trial. Funnel Sam is on the product growth team and he's got some beautiful dashboards. Did you a man so.

B

This dashboard shows us for the gitlab comm trials, the conversion rates of people that view that trial form and submit it and then also get into it in a minute, but also the source, the the locations that people are coming from, the way that we're doing that is using custom, URL parameters. So we can tell where people are coming, what CTAs people are coming from in the product and then we can talk about this later on, but we can also add that to the website as well, so we can pull it all.

B

The way through to revenue is the ideal scenario here. So what we're looking at here is views to the first trial form to submitting the second one. It's not gonna add a link up here. It's not but the two four two page form for signing up for a SAS trial. The first one is the whatever it is 8 questions that you have to submit, and then the second step is: where do you want your trial?

B

It is it on your personal namespace, or is it on a group, so this is going from the first trial to submitting that second form, so the all the way to completion- and we can see that the conversion rate is ticking up nicely. So we've iterated on this forum a few times in the last quarter, or so so.

B

We've made it easier so that the group URL doesn't error out if the URL path has already taken and then the big thing that we did is for we check to see if you're, an existing user or not, and we change the Skip trial logic. So if you're a free user there. Sorry, if you're, not a user of gitlab, the CTA at the very bottom says skip trial, continue with free account and we've switched this so that if you're already a user of gitlab, it just says go back to my account.

B

So we removed that that that big call-to-action to say, like oh, it's, free, free, free and what we've done is we've drastically reduced that skip trial rate for the sash trials. By doing that, so this is stuff that we're constantly looking at and then we can see on the second form here, specifically the submission rates actually really high. So we don't lose actually that many people on this page, which is a good thing.

B

A

Rate with all these steps from muta conversion.

B

Right around 25% right now.

A

So this is from I'm confused. This is view through from first to second trial, but, like oh, is that completion rate all the way through yeah.

B

That's completion rate all the way through okay,.

A

Cool, so that's really helpful um yeah. Obviously it's like it doesn't feed back to our Google Analytics view, but that might not really matter um at least not right. Now, yes,.

B

Let me uh I'll scroll down and I'll show you, so this is just to like run through them. This is viewing the first trial form to skipping the trial, which is that at CTA that where we change the logic for existing users and now we're got it down to whatever this is 17%, and then this is for people that view that second form, which is really like. Where do you want your trial and the completion rates right in the ABS, which is pretty good?

B

We can get this up probably higher, but 80% is pretty good and then the view to submit first trial. This is where we're seeing it go up nicely and that's from the change on the skip trial for.

A

B

Users and then.

A

What's this one, this.

B

Is showing us the conversion rate of if somebody selects a group trial or a personal trial.

A

B

So one of the next experiments that we have lined up on the growth team is, and this again for the people listening in the future we were just talking about. Why is this form live in product and not on a marketing form?

B

One of the benefits of this being a product form is when this question gets asked now, one of the experiments we're gonna quickly run now that we have access to this data is we're gonna, actually say: okay, if you say that you're using this form or starting this trial for your company, we're not even gonna, show you the personal option right and then we're gonna have a second variation, where we don't even show the personal option to anyone, and that what that'll tell us for that cohort is what percent of those users were likely acts looking for a group trial because we'll see the uptick in conversion rate to paid for that group, because there's likely a cohort of people that are just selecting an individual trial, because we don't do anything to tell them what a group is and the value that a group.

B

You really need a group to see team value and get.

A

Okay, so this is kind of where the marketing question goes, which is obviously group is a much higher leverage conversion rate, so I would love to find a way to send more people to the places that they need to go to become essentially a look-alike for someone who converts as a group user, like there's, definitely pages of gitlab, Docs and pages of gitlab content that are more geared towards those people and like we could have more aggressive calls to action so I'm curious.

A

Do you have a sense of like a path that led that person to be to make that choice? We.

B

We don't right now really that caught the the verbiage on the page off the top. My head is really. It tries to spell out that you should pick a group if you're setting this up for your team, so we try and call it out, but it's still on the user, but do.

A

B

A

Guess I'm trying to say where were they before they came here? Yeah.

B

Referring yeah, referring is I'll get to that in a second, it's a little bit tough, because we don't we didn't. We could tell a little bit about the referring URL, but there's a lot of like pass-through URL. So a lot of times, they're, referring URL, will just be the previous page in the trial process, and we can't see you know five URLs, but before because we're not tracking user ID.

A

We have a cookie I.

B

You have a cookie on Google Analytics. We we can't track users on periscope once they become a user. We cannot track them at an individual level.

A

Okay, so that would be like the counterpoint which I'm not trying to make with any convictions. I really don't.

B

A

Make would be if you had. If it was a non product form, then you would be able to track deeper into the form and and answer that question. Yes,.

B

But I'll get to get to a point. I'll show you something in a second here and how we can bridge that gap.

A

B

A

What you're saying not and I'm not making a case for that yeah.

B

Yeah totally now this shows us the conversion rate by cohort. So this is each week that people sign up for trials, excuse me and then how many days it takes for them to convert. So we can see the vast majority of our conversions are coming right at the end of the trial. So it's that twenty the thirty day and then anything this long tail. This isn't saying that their trial was extended. This long, it was just that some people upgraded this late after their trial right now.

B

Some of these, these are probably without digging into the data. These are probably more like, like sales assisted conversations, somebody does a trial and then they have a longer sales cycle working with a colleague in the sales department.

B

Now what we can see for trials with our custom, URL parameter tracking is, we can add, basically trial source to any URL, and then we can see where they're coming from for particular CTS. So what I can show you here? Is this top one?

B

It's telling us for specifically for trial skips where they're coming from so this isn't the best. It tells us they're coming from the previous page. But here let me show you so for dot-com. We just need to add some JavaScript to the page, to the trial source page or that trial, picker page, where you pick SAS or self-hosted, and then we'll be able to pass through any kind of UTM parameters, but at least a custom.

B

One and I can show you how it's structured, but just to show you in app I can see all the CTA's that people are coming from to start in app.com trials, hmm and what we can actually see here is the vast majority are coming from this one. That's called the gold call-out, which is just this kind of banner, that is to start a trial.

A

B

No, it's actually the one it's on the parent of the its. What is it the pair of the project, the project overview page and then specifically, you can see within this next chart here- shows us of these gold call-out clicks. The 66% are coming from this specific URL of get lab. Comm backslash. You only reach this specific URL when you first sign up for get lab. That is you the very first page you land on, and there is a trial banner on that page. So what we're gonna actually test in our backlog?

B

I, don't know, then the next few milestones is. This clearly shows that the vast majority are trials are actually coming from brand new users and signed up for free and then selected a trial, so we're actually going to add the start. The trial signup process in as an experiment into the free user signup as an option so will be required, but we'll ask them during for user signup.

B

Do you want to sign up for a trial, because this data supports that the vast majority of people are at least of our trials are interested even from the free user side?.

A

Okay, that's really useful I mean it supports the premise we're operating on, which is that if we simply grow new user traffic to get lab comm like oh, we are able to hit most of our other numbers. Yep, there's a pretty predictable path there. It's so like this seems to say that's also they the power lies there as well for you, when you look at this yeah.

B

And then what this? This is actually what I was referencing before, but but for tracking sources from about Duquette lab. The vast majority are just coming from that the free trial picker page, but we can add some JavaScript to this page and then you could add UTM. You could pass the UTM parameter into the like that.

B

The button the sass button just that that href and then we would pick it up and we would have it in all of these charts and then we could give it to you all the way to revenue the exact dollars in revenue, yeah.

A

How do we get that I want I want to be with update all of our CTAs. To give tracking I mean you want to understand the ROI on different activities.

B

That's exactly what I want yeah.

A

I, don't really care how we get there to be totally honest with you. It's much more. Just I, just I like to believe that all the activities were doing or adding up to something. Yeah.

B

So all it's looking at is we have two fields: GLM source and then a value, and then GL and eve off the top and check the other report, but I believe it's GLM content and those that that's what's just powering these. These reports.

B

So essentially, I can I can help write an issue for you, I, don't think it's worth just talking over, but I sent I did this at my last company.

B

The way that we can pass this value is we have JavaScript on your trial, picker page that essentially looks at your the URL, which contains your your Google, Analytics tracking and ideally, some sort of CTA reference, and all we do is copy that Google Analytics over to our GL n source and add it to the HR f2 for the SAS trial button and when the person clicks that they go into the SAS trial, signup flow. And then this report starts distributing specifically, which CTAs they're coming from based off of your tracking awesome.

A

Yeah, if you want or an issue, that's awesome, I think what we really should do is I'm happy to drive this is we should make a handbook page that within the marketing team that says, are you building a CTA? Here's like the checklist of stuff. You need to do and this happen.

B

A

I mean we already want, is for other reasons, because we're having some consistency, challenges with conversion, and this is just another really awesome thing that everyone should be doing. Also. We want to feed you data to like this is a way for.

B

A

Yeah, thank you that was super helpful yeah.

B

So that the idea is ideally I want to get to a point where I can say here all the CTA's that drive the most value to trials and to direct revenue for get them, and then we can work together and iterate on which ones are performing the best and what are common traits in those and maybe the ones that aren't performing the best. Do we get rid of them or do we try and work together on cycles on how we can improve them if they're not converting?

B

Well, so I want to be able to get to a point where we can actually have a dollar value for these CTAs so that we can predict future revenue by saying, like, oh, you know, Danielle's team did a great job. They increase traffic to this CTA by X percent and within 90 days, based off our historical data. That's gonna generate the business-y amount of ICB funny.

A

You mention that because I'm literally kind of tasked with building a similar view from a budget perspective, so I know that we have probably a struggle which is we don't have a lot historical, so we're gonna need to get the trip going, but obviously that's that just increases my sense of urgency like hey. What do I need to do to support you and getting these implemented, and we do have people and our don't do that audit and go and make those changes.

A

I just would love to get like a very basic issue, spec from you of like here's, how we would like this to be done consistently and maybe an example.

B

A

B

And I think the probably the best way to go about it too is, and we can work on this together and I'm, not a Google Analytics expert by any means. But if we know the the source traffic to that trial picker page and we can just go top down and pick the biggest sources and just work our way down and get the vast majority of traffic probably pretty quickly and then work through all the outliers over time. Yeah.

A

And we have Shane rice on. Our team is definitely our Google Analytics nerd and he won't be thrilled to work. Yes,.

B

It's me yeah. He always helps me anytime. I need help he's my.

A

B

A

I will I'm gonna post this on unfiltered if it's cool with you and I'm gonna, ask him to just watch this video because I think, based on what he and I talked about this morning, he'll probably be like sweet, okay, cool and I. Don't need to solve this problem and he'll get pumped about getting these CTAs wired up. So this is a win for everybody. Yeah.

B

No, this is gonna, be awesome. Well,.

A

I'm gonna stop the recording, really quick.