Krista Seiden, Google - Measurement for Growth

October 23, 2018

Measurement for Growth

By Krista Seiden of Google

Slide 1:

Thanks for having me so much here at the Analytics Summit. I hope you guys have had the opportunity to listen in on some of the other sessions, and I’m going to talk to you today about measurement for growth. So, let’s dig right in.

Slide 2:

So, you just heard a lot about me. I’m the analytics advocate as well as a product manager for Google Analytics, but I’ve actually been in the analytics and optimization space for about a decade with a lot of that time on the practitioners side at companies such as Google, The Apollo Group, and Adobe. I’m the co-chair for the Digital Analytics Association San Francisco chapter, and as mentioned, I’m a huge advocate for the women in analytics movement.


If you want to get a hold of me, you can find me on Twitter @kristaseiden and on the web at


Slide 3:

So, today, we’re going to talk about a few things. We’re going to start with a baseline of what is growth. Then we’re going to dig into some growth strategies. And finally, the meat of the presentation is going to cover measurement for growth.

Slide 4:

So what is growth? Growth hacking is a process of rapid experimentation across marketing channels and product development to identify the most efficient ways to grow a business. Growth hackers are marketers, engineers and product managers that specifically focus on building and engaging the user base of a business.

Slide 5:


You hear a lot of buzzwords when it comes to growth. You hear growth, AB testing, SEO, product, engineering, development, a lot of different things.

Slide 6:


Essentially, a growth marketer is wonder woman.


Slide 7:


So, a little while back, we invited a local growth marketer, growth hacker, to our digital analytics association San Francisco chapter event to give us a talk about what growth hacking and growth marketing is all about. Specifically, how it’s different than analytics or optimization or a lot of the things that us, as analytics professionals, are used to.


So, he started talking about some of the similarities and differences, and then when we pushed him, he kind of got into this funnel.


Slide 8:


He started to draw this five part funnel and essentially, what he said, was this.


Slide 9:


He said, let’s talk about a growth hacker versus a digital marketer. Now, traditionally, digital marketers are going to be the people that look at the very top of the funnel. They’re going to focus on aquicistion, so getting people to a website. And activation. Getting people to start that first time experience. Whereas, a growth hacker, is actually interested in those two as well as further through the funnel to retention, revenue, and referral.


Now, a few of us in the audience, myself included, are listening to this and ya know, we took a little bit of offense maybe. I’ve been on the practitioners side for most of my career and in roles where I led analytics and optimization. I was focused on every single one of these things. Not just acquisition and activation. So I think that a digital marketer can and does do many of these things. I also hate the term growth hacking, so for the rest of this presentation, I’m going to call it growth marketer, growth marketing, and growth.


Slide 10:


So, growth does a have a few things that do set it apart as a discipline within digital marketing. One of the things is product market fit. So product market fit is the idea that you have to be in a good market with a good product that satisfies that market. As a consumer, you can think about it in terms of, if that product didn’t exist anymore, how would you feel? If you would be heartbroken, it’s likely that product has product market fit. But if you just move on to the next app or business that fills the same need, perhaps it doesn't.


The good example of this is Uber in China. Now, Uber doesn’t actually have product market fit in every market that they go into. In China specifically, they struggled to really get that business off the ground. And in August of 2016, Uber actually pulled out of China and seeded their business to their competitor Didi. Now, I’ll never forget this date because I actually happened to be in Shanghai when all of this happened, and all of a sudden, I couldn't use my Uber app anymore. But I also couldn’t use Didi because I needed a chinese credit card and I did not speak mandarin so I couldn’t understand the app. So, for me, it was unfortunate, but for the market overall, there was a clear signal that there was not product market fit there.


Slide 11:


Another thing that sets growth apart is the way that growth teams are organized. Growth combines elements of many different roles, and all of these roles come together in what is known as a Growth team. So, for example, there might be somebody focused on analytics, experimentation, product design, engineering, product marketing, content marketing, or SEO. All of which are going to roll into this growth team.


Slide 12:


So now let’s move into some growth strategies.


Slide 13:


Now a typical growth framework might look something like this. Might look like the funnel that we looked at earlier where you go from awareness to acquisition all the way down to reactivation or referral. Typically, growth is come out of the startup scene. You might have started a growth team with one growth lead or maybe a small growth team. And as that company scales, maybe each business area or product unit or business unit would have their own growth team focused on that particular area.


Slide 14:


At Google, we’re thinking about this a little bit differently rather than a five part funnel. We’re looking at it in three parts. We call it the AAA growth marketing framework. This starts with our strategic drivers. This is acquisition, activation and adoption. Next, we have our desired outcomes that are related to each of these drivers. Perceived value, realized value, and ongoing value. Each of these is going to have some primary tactics that you use to drive these desired outcomes such as paid media, SEO, email, help center, in-product, all of those kind of tactics depending on which driver we’re looking at. And for each tactic, we’re going to have some key measure to be able to measure our success. So, we’re going to look at things like impressions or new accounts, conversion rates, success rate to sign up, or adoption and retention. But all of these things are going to lead to one strategic goal. In the case of a lot of our businesses, this is monthly active users.


Slide 15:


For the Google Analytics 360 team in particular, we’re looking at a centralized growth product team. So we have our email, SEO, media and ads folks all working with our growth marketing team who is partnering really closely with our content and education team to build the content for these programs. And each of these people is partnered really closely with our engineering side. So we have our growth engineers, our data scientists, our growth PM, all of which are coming together to drag growth for our business.


Slide 16:


Now that we’ve covered some of the early parts of this presentation, I want to dive in to the meat. Which is going to be measurement for growth.


Slide 17:


So, measurement starts out with this core concept in growth of the north star metrics. The north star metric is the single metric that best captures the core value that your product delivers to customers. Optimizing your efforts to grow this metric is key to driving sustainable growth across your full customer base. This is a quote that comes directly from Sean Ellis. He’s kind of credited as being the founder, the father, of growth marketing.


Slide 18:


So, let’s look at some examples of what a north star metric might be. Facebook is famously known for having daily active users as their north star metric. Airbnb is nights booked. Uber is weekly rides. It’s important to realize that the north star metric should comprise value from both the user and the company. As the user, I’m getting value for every night that I’m staying in an Airbnb as well as Airbnb getting value in terms of the revenue they’re collecting.


Slide 19:


To put this in the context of Google Analytics, our north star metric is monthly active users. For Google Optimize, monthly active experiments. And for Google, they just use monthly active users. So across the board, we’re really looking at that monthly active metric for our business.


Slide 20:


That north star metric should be what every person on that growth team is striving towards. So you have all of these different roles as we mentioned earlier, and all of them are working towards the same goal which is the north star metric.


Slide 21:


So let’s look at some key metrics by life stage. If we were in that acquisition stage, a couple of things that are going to be important would be impressions and of those impressions, how many we’re converting, what that signup rate is, or our cost per acquisition.


Next, we have activation where we’re going to look at inactive users and retention rate. And activated users and cost for activated user. And finally, adoption and adoption rate. But let’s go ahead and dive into each of these a bit more and look at some of the tactics that we can use across each of these different areas in the AAA framework.


Slide 22:


So we start with acquisition.


Slide 23:


We’re going to just mention this, but we’re going to look at impressions and how we can turn those impressions into conversions, new users, and be able to measure the cost per new user.


Slide 24:


So first, we need to capture that impression and turn it into a click. We have a lot of different acquisition strategies across digital marketing. Some of the channels you might be focused on are Facebook, Twitter, Instagram, you might have a blog where you’re pushing out a lot of different blog content to drive users and drive eyeballs on your content. Display advertising is a big one. Organic and SEO traffic. And finally, referral and affiliate traffic.


Slide 25:


And once we have that impression driven to a click, we need to make sure that we are capturing that as an acquisition. So a few things that are important here. The first is campaign tracking. Now, you see a screenshot on my screen of campaign tracking in Google Analytics. Here we’re looking in a source medium as well as a secondary dimension of content. Now in this case, this is from my personal blog, and I’ve actually tagged every social post that I’ve put out leading to a blog post with very specific campaign parameters. I tagged the social network, the medium of social, and the content cloud is being filled by the actual title of the blog post that I’m driving my traffic to. And what that results in is actually the ability to do deep analysis on the types of posts that work best on which channels.


For example, when I blog about google tag manager, I actually tend to get more traffic from Google plus than from Twitter. Digging deeper, I found out, that’s because there’s a very active google tag manager community on google plus. Whereas on Twitter, I get much more traffic for pretty much anything else that I might post.


Other things that are really important for capturing that acquisition are page tracking, event tracking, form tracking, cross-domain tracking and referral exclusions, and having the right data structures in place.


Slide 26:


So, thinking about data structures, we need to make sure that we have the right data to collect or that our analytics platforms are giving the ability to do the analysis that we need by having the right data already available for us. That’s where the data layer can come in. To collect custom information.


Slide 27:


So, let’s look at an example of this.


Slide 28:


I was at home in San Francisco, and I was craving some Italian food delivery. So I googled “postmates italian food sf.” Postmates is a delivery food company. And that first one, “Tommaso’s restaurant in San Francisco” sounds pretty good. I’d go ahead and click on that, and I land on their page on postmates. And that lasagna looks pretty good, the food all looks great. I decide I’m going to order from here.


Now, if I flip my perspective and I pretend for a second that I’m the analysts at postmates, and my boss asks me what are the top five types of food in our top five cities because we want to build some special marketing campaigns to target those. Well, I look at that url, “”. It is not giving me a lot of information. It does mention San Francisco, but it doesn’t do it in any kind of directory structure. So it’s not that I can actually look at my pages report in Google Analytics and drill down by city or by type of food because it doesn’t even mention the type of food that is here. When I look at some of the other pages, San Francisco isn’t always in the name, and so that type of analysis even filtering on city names is not going to do me much good. But that is where the data layer comes in.


Slide 29:


So here you can see, I’m going to add “city”:”san francisco” and “pageSubCategory”:”italian” to my data layer. And I can capture those pieces of information from the data layer and send them to Google Analytics as a custom dimension. I can then use that custom dimension to slice and dice my data in Google Analytics and do the analysis my boss is asking for so that I can come back to her and say, “Okay, here are our top five cities, here are the top five types of foods that are most popular in these cities, and here is some creative that we might want to test out.”


Slide 30:


Another thing that’s really important here is using event tracking for form field interaction. And really understanding how people are moving through your forms. I like to use onChange events for this to measure each form field interaction. This is great because of the work on single page forms, multi-page signup flows, checkout flows, and many more types of forms of flows. You can pinpoint exactly where somebody might actually drop out of your flow and look for errors or look for ways to optimize that.


Slide 31:


So for example, here we are looking at the event tracking for the Google Analytics marketing page. I’m looking at the top level, event category. And if I click on one of these, for example, my number eight, contact form. I’m going to get to the second level of the event hierarchy, not the event action.


Slide 32:


In this case I’m using the action to notate which product people were interested in when they clicked on that contact form. So if I click into one of these, Analytics 360 Suite.


Slide 33:


I’m actually going to get to the final level of the event tracking hierarchy, which in this case is the label. Now, I see some really interesting things here. First off, I need to note that normally, event tables in Google Analytics, are sorted by the first column after the primary dimension, which in this case would’ve been total events. But I resorted this by unique events. You can tell by the downward facing arrow in the top right part of that column, and when I do that, I see that this is actually sorted as a funnel or as a flow of my form. This is actually the order that people are going through my form. First name, last name, job level, email, all the way down.


Slide 34:


There’s one outlier. That’s region. That’s because that’s an optional form field. But I think that there’s something else that’s really interesting here and that catches my eye. That’s line number nine. The submit field. When I look at this, I see that there are significantly more total events for submit than there are unique events. And so, to troubleshoot this, I actually go to my website. I look at this contact form. I think there’s, ya know, not a lot more than you can do that is going to be better to troubleshoot this than actually get that form in front of you and put yourself in the user’s shoes to figure out what’s going on.


And so, I’m looking at this form, and what I’m realizing is that the fields that are required actually have very small little asterisk next to them that indicate that their required. So it’s not very clear to the user that they must fill in all of these other fields. And so I think what’s happening is that people are filling in some fields, they’re hitting submit, and they’re getting an error because they may have missed a field or two that they didn’t realize was actually required before.


So I made this suggestion back to my design and product teams and we went ahead and fixed this on the form. And this is looking much better now.


Slide 35:


The next part of the AAA growth framework is activation.


Slide 36:


Here we’re going to look at things like inactive users and our retention rate and if users do go inactive, we’re going to send them right back through our acquisition phase with infractions. But what we’re hoping to see is our activated users, our success rate, and our cost per activated user.


Slide 37:


So one thing that is often interesting and useful to businesses is something called network effects. So this is important if you have some kind of a social network or really any type of product or network that gets better as more user are using it. This is called the network effect, and you need to achieve network growth to make this platform very useful. So good examples would be LinkedIn, Facebook, Twitter, Slack.


Facebook is very famous for doing a bunch of experiments for network effects. What they did, what they found after a lot of this experimentation, was that when you add seven friends in ten days, your news feed is going to be more useful to you and the product is going to be more sticky. They figured this out after doing a lot of experimentation with number of friends required and the number of days that you had to do this. So they experimented with their messaging, their signup flows, in-product message. You can do all of these things to encourage users to increase their network and their network effects.


Slide 38:


Now other things that businesses are focused on, especially if their not a network effect type of business, would be time to first interaction. So examples of this. Uber is definitely focused on time to first ride. How long after you sign up is it going to take you to take your first ride. Deliveroo or postmates or any food delivery platform is going to be that time to first order. Pretty much any type of ecommerce platform is going to be focused on the time to first purchase.


An example that we’re looking at on the screen is google AdWords or what’s now called Google Ads. Signing up is important here, and that’s the first step, but activation is creating and running that first campaign.


Slide 39:


When they put this in the context of Google Analytics, acquisition happens when a user signs up for an accounts.


Slide 40:


But activation happens when the user actually installs the tracking code and they start using the product. So then the question that we need to ask ourselves as digital marketers or as Google marketers, is how can we make this time period shorter. What are the experiments and the product tweaks that we can make that might encourage people to activate faster? And that is really the essence of growth marketing is this rapid experimentation to drive people through our growth funnel.


Slide 41:


So a hypothetical example of this. Hypothetical because this is not something that we currently do, but it would be cool. Using different experiments to help make that implementation process easier. So here we are looking at a screenshot of the tracking code for Google Analytics. This is what you would see if you haven’t installed it or even if you have installed it, and you’re looking for your site code. But you have to take the site code and install it on your page. It might actually be difficult for some users who don’t have access to development resources easily or are not using tag management platform.


But wouldn’t it be cool if we had a button there that just said “add google analytics code.” By pushing that button, we’re going to automatically inject this code onto the page for you if you have the proper permissions, and you don’t have to go through the process of tracking down the proper development resources and making sure that that code gets on your page. That would definitely make the process of implementation easier and get users using the product faster or activating faster. I think that would be really cool.


Slide 42:


So one thing that we are doing is experimenting with activation cues in product. So when somebody is in their product and they have that code on their page, but maybe they haven’t started using the interface or they’re not using it as much as we would like. We’re doing something that we call guided flows.


These are these little blue pop-up boxes. So for example, we’re looking at one on the screen about custom segments. So if you’ve never built or used a custom segment before, you might see this blue pop-up. If you click on it, it’s going to open up the segment folder. If you keep following the instructions on the blue boxes, it’s going to walk you all the way through creating a brand new custom segment. So it takes you from never having used it before all the way to having created a segment and apply it to your data. So great way to actually move people through the process of activating on individual features.


Slide 43:


So activation programs in general are all about optimizing for the base value realized. So in our case, we are developing our strategy and our creative for these or engineering our implementation triggers. So if we see that somebody hasn’t used a certain function, we’ll make sure we have a trigger in there to trigger a guided flow or something like that. Then we’re going to run those. We’re going to pull the data, understand what’s going on, and launch and iterate. That’s going to be backed up by our educational content about these features.


And our goal is really to drive as many new users as possible to realize that base value for the product while making it as easy as possible. And throughout all of this, the metrics that we’re going to be focused on here or the things that we really want to measure are going to be our activations complete, our success rate, our average time to value in days or weeks, whatever makes sense for the business, as well as our cost per activated user.


Slide 44:


So the final part of our AAA growth framework is adoption.


Slide 45:


With adoption, we’re really focused on that overall adoption rate. But again, if users fall out, we’re going to send them back through this whole process with activation. Acquisition, activation, finally adoption.


Slide 46:


So one of the things that we are doing for adoption is an email nurture campaign. So on day zero when somebody signs up for analytics, we are going to trigger a signup email. It’s going to say something like, welcome to google analytics, you want to make sure that you can sign in, get started. Here's some educational information. Here’s a link to download our app. Really gives people a lot of the materials that they need to get started using our product.


And then on day 30 or really every month after that, we are going to send what we call a monthly performance report. And this is going to give you information about your top analytics account, and what is going on in the past month. Kind of the snapshot of that information about your users, conversion rates, for goals, or ecommerce, and a bunch of information that you might find interesting by email. But it’s also going to have queues to click through into the product so that we’re really trying to drive users back into the product.


Our overall goal here is to encourage repeat usage, reduce customer churn, and provide product value to users. And the key metrics we’re looking at here, what we’re measuring our success on, is open rate to those emails, click through rates from our in-email queues into our product, return and re-engagement rates for our product, and overall, our north star metric of our monthly retention rate.


Slide 47:


And then what I think is really important for looking at adoption overall is full funnel analysis. Understanding and acting on the full funnel of the user experience from first interaction all the way to a closed lead and re-engagement is a really critical part of this whole growth marketing business.


So one of the things that I think is really cool here is the new Salesforce integration with Google Analytics 360. In this case, you can actually send lead status changes as events to Google Analytics 360, and here, you can see in the screenshot, we are actually looking at a business that goes from online to offline with that sales team driving those leads through this lead status change in Salesforce. So people are online, they submit a lead form, that lead goes offline, and we actually see by this lead status change event being sent back to Google Analytics, how people are progressing on the whole funnel, online and offline.


And not only that, if you click on those downward red arrows, it’s going to actually give you the ability to create a segment and remarket to users who are dropping off. Even if they are dropping off in your offline flow, you can then remarket to them online. I think this is a really cool way to be able to re-engage people even from offline drop off online with the data that you have about them. Their full funnel journey.


Slide 48:


Next, we have retention and lifetime value. So we want to measure our user lifecycle by looking at lifetime value by user-ID, time to repeat purchase, the frequency or recency of purchase. And things that we can do here is experiment with offers, nurture campaigns, or push notifications to decrease the time to repeat purchase and drive increased retention and lifetime value.


I think a unique example of this is some of the tacts that does. So you may have seen the subscribe and save box on some product offerings on amazon. This is generally for things that you would be using on a regular basis or probably should be using on a regular basis like toothpaste and deodorant and toilet paper and those kind of things. And it’s actually going to pull up in search results, the lower price to subscribe and save. And then you’re going to see that you are subscribing to purchase this one every two months or three months or six months. You can choose to then make a one time purchase, but it’s going to be a higher price. That actually gets at the heartstrings of consumers where they don’t want to pay more if they have to, and they know that they’re going to use this product again. So I think this is a really good tactic that they’re using here to drive repeat purchase.


Slide 49:


Finally, we have targeting for re-engagement. And here, we’re looking at this in the context of Firebase and Google Analytics for Firebase. And using that data you’re collecting for your app analytics. And what we’re doing here is we’re targeting audiences that contain at least one of something like purchasers, lifetime value is greater than zero, or high scoring users, with an offer for a free character in their game as a way to incentivise high value users to come back.


So we’re measuring things like message views and interactions, repeat visits, and offer acceptance or purchase.


Slide 50:


So, then, after all of this, you’re probably saying to yourself, “Aren’t we all growth marketers?” Maybe. But maybe not.


Slide 51:


I think the biggest difference between a growth minded organization and a more traditional product and marketing driven organization is the willingness to take risks. But it’s also north star metrics and full funnel analysis and team organization and an experimentation mindset and the idea that velocity is greater than accuracy and the idea that risk equals reward.


Slide 52:


So in summary, growth marketing may or may not equal digital marketing. It depends on the organization. Growth teams are fast acting due to team structure, culture of rapid experimentation and risk taking. The AAA growth framework is acquisition, activation, and adoption. And please don’t call it growth hacking.


Slide 53:


Thank you very much. I really appreciate you listening to this presentation, and enjoy the rest of the analytics summit.

Previous Video
Till Buettner, Deutsche Post DHL Group - The Perfect Dashboard Rocks
Till Buettner, Deutsche Post DHL Group - The Perfect Dashboard Rocks

Create advanced dashboards to give more value to your organization and separate yourself from the competition.

Next Video
Judah Phillips, Vizadata - Machine Learning for Marketers: Predicting Tag Compliance
Judah Phillips, Vizadata - Machine Learning for Marketers: Predicting Tag Compliance

Learn about supervised machine learning in the context of tag auditing and monitoring.