Peter O'Neill - Google Analytics Crash Course: 30 of the Best Tricks and Hacks

November 23, 2016

Slide 1:

Good day. My name is Peter. I feel like we’ve had a long day of everyone listening in here, only two sessions left to go. I’ve only got 30 minutes, so for a quick introduction first of all—

Slide 2:

As you’re probably picking up from this accent, I am an Australian living in London. I’ve been here twelve and a half years now. I’ve worked in digital analytics for over 10 years. I founded L3 Analytics back in 2010. We’re a London-based digital analytics consultancy, a GA, GTM certified partner. Currently in the process to merge with AEP Convert to form LeapThree, which is a name you’ll see associated with me a lot more in the future. Which means we’ll be supporting Goggle Analytics and Adobe Analytics services from set-up to CRO, personalization, and data strategy, working with clients of all sizes across all sectors. I’m also the founder of MeasureCamp, a digital analytics un-conference. We’ve had 25 of them around Europe and in Australia and Hong Kong. We’re hopefully getting across to the US next year. Also the co-founder of MeasureBowling.

Slide 3:

Moving on, a quick explanation of this session here. I was originally going to give 10 Google Analytics tips, tricks, hacks but decided that wasn’t enough for me. I wanted to go for 30, unfortunately, I can’t get all 30 in in 30 minutes. Now I’m targeting 25 tips in what’s left of my 30 minutes. Some are very basic, some are advanced, but they cover business principles, Google Analytics set-up, reporting, and analysis. So basically, if you start to learn something and you find it’s boring or not relevant to you or something you already know, give it a minute something comes along next. For now, let’s get straight into it.

Slide 4:

The first few tips are before you actually open Google Analytics itself. Number one: before you start anything at all, remind yourself, the purpose of digital analytics.

Slide 5:

The purpose here is to provide the intelligence, the insights, the information that inform the business actions leading to an improvement in performance for online organizations. These business actions are things that people in the company are doing every day anyway. They’re taking actions, they’re making decisions. We’re trying to make those actions and decisions smarter, and because they’re smarter, they’ll lead to improvement in performance. Performance improves, and basically the end goal of all of that is to make the company more money. We as analysts, we don’t actually do that much ourselves, we just provide the information and the insights, it’s the rest of the company that actually does the work, that actually take these actions so they drive the performance forward. Keeping that in mind always, makes analytic much more effective. Because of that, you shouldn’t be trying to create a Google Analytics solution for yourselves, for the analysts.

Slide 6:

You’re trying to create a solution that works for the non-analysts, the people that are making these decisions and taking the actions. That’s what matters. You don’t want a situation where you can’t get the really valuable information. If you go to this customer report here, click through there, apply this segment, add a secondary dimension, and wow, there it is. If he’s someone who doesn’t understand numbers, possibly doesn’t even like numbers, to go in, go to send a report, and say, “Here’s the answers to questions they need.” That’s when Google Analytics becomes really effective, really valuable. Don’t include any abbreviations and don’t make it hard to find the information. People shouldn’t have to think to be able to use Google Analytics.

Slide 7:

Before that, to get started, you’re looking to invest time upfront to understand what information these people need that no one else in the company, who aren’t the web analytics team, what information that they need, to make these decisions, to take these actions. So go out there, talk to people, talk to those end users, understand their decisions, understand their actions. And then paper and pen and write down a plan for the information you need to capture in Google Analytics. From that, you’ll get a pretty long list. You’re not trying capture everything though. You can capture so much information in Google Analytics if you really try to.

Slide 8:

It causes problems. If you’re trying to capture too much information, everything extra you capture, is an extra piece of work for the developers. That’s just not realistic in most cases. There’s only so much time and resources available to you. And if you could actually capture all the information, it’s overwhelming. These people have got too much information they’re looking at. They wouldn’t know where to look, where to start, where to go with it, and they usually end up ignoring everything.

So what a much smarter approach is: to track the most useful information first. I aim for that core level. The macro conversion actions. The critical information. At the utmost, going to that silver levels are where we capture those micro conversion actions and directly useable information. Basically, that first stage of implementation, you should be able to connect every single piece of information you want to capture to an action that someone in your business can take. If you can’t do that linking, if you can’t drive an action forward, don’t bother. It’s not useful information at this point in time. It’s nice to know, let’s save it for the future.

Slide 9:

And going from there, you actually—it’s a weird word to use within Google Analytics, or in any analytics at all—it’s creativity. But it’s one of the biggest barriers to getting true value out of Google Analytics. You need to think of the data to capture before it can be tracked. So what information is most useful? While you’re being told that, “Here’s the action being taken, the decisions they made,” and you’re looking for information to inform that, you need to think, “What is really, really valuable here?” Because you can capture so much. You can capture the weather while someone’s on the website. You can capture how many different product sizes or product SKUs are in stock for each product.

For people who have a subscription to the website, you could be capturing how long it is and when it expires. So you can understand a clear performance of people who were in their first week of subscription versus last week of subscription, and how they behaved differently on the websites. People who are looking to book some holidays, you look to see if the holiday booking they’re looking for, during the school holiday period or a different time, because again, that behavior is different. This information here is where the value really comes into it. Out of the bubble, out of the box. So you can be creative and think through what information is most useful to those end users and you capture that.

Slide 10:

So going on, here’s my useful things to catch within Google Analytics

Slide 11:

First of all, I always capture the page URL and the page referrer, content grouping or hit-based custom dimensions. Simple things, but a key use for it, is to identify cause of 404 error pages. Create a custom report very easily for displays. Here’s a list of URLs that generate the 404 errors. And you click through into the referrers to that page. Looking around and here is the list of referrers that lead to 404 error pages, clicking through into the URL of these 404 error pages, either way, that should identify a list of sources and of broken links on these pages, on both your own and on third-party sites. Find these broken links, fix them, improve performance.

Slide 12:

Next one: always record the number of search results being returned from internal site search. Use a hit-based customer dimension for this. The key is, for this one, to identify search terms with zero search results. And basically, quite simply, fix it so people can find what they’re looking for. They could be misspelling the terms or putting in a different sequence of words, looking for things that actually already exist, so make sure that they can get the right page for that. Potentially as well, you’ll be able to identify a new business opportunity. I’ve seen this time and time and time again in different companies; people are on their website, looking for something they can’t find right now, but it’s a product that could be solved there.

Slide 13:

Form validation errors; they’re things that are blocking people from completing forms and getting in the process. Record the form name, field name, and error message in hit scoped custom dimensions. Then create a custom report to identify the cause of form errors. An example can be seen in the screenshots here where I listed out the forms, the number of errors per form. Clicking through on that, there’s different fields driving those form errors. Clicking through on the field name shows the error messages being displayed. That tell you, straightaway, really great insights into why it is having issues, or in this case here, where the developer needs to adjust the settings to allow for telephone numbers to be entered with a space it it because that’s common sense.

Slide 14:

Campaign tracking, something everyone always talks about. It’s really boring, something that’s not happening enough in most companies.

Slide 15:

First of all, there’s five Google Analytics campaign parameters to be used. Simple tip there—use all five. Ignore the variable names here. It’s medium, source, campaign, content search term—I don’t really care. To me, it’s just five fields: campaign terms, if everything goes well, it’s used for paid keywords. I don’t care, it can be used to capture anything at all. It doesn’t have to be a paid search keyword, it can be used for any channel. To really capture the granularity campaigns, you look at performance at a high level and at a low level. An example of this for email; the type of email was promotional or operational. I want to capture the name of the email category or the name of the distribution list it was sent to, the name of the email itself, the name of the email campaign—or possibly the date the email was sent—and a link identifier. So a lot more information being captured there, a lot more granularity available to me to really understand and monitor performance, to make changes, to make the business more money.

And as a bonus tip, if you actually need more than five parameters, that’s fine. Using custom dimensions changes the way the session is being scoped. Capture all the additional information in a single campaign parameter using a separator and using view filters, fill it out and populate these additional session scope custom dimensions within your campaign parameters until you’ve got, not just five, but six, seven, ten different campaign parameters to be used in reporting and custom reports.

Slide 16:

While many companies still aren’t using campaign tracking, it’s very frustrating. It’s really crazy when you’re paying for a service such as emails, display ads, affiliates, anything like that because in any case here, the vendor you’re working through, they already built it into their systems. They can flip a switch in the background to automatically add Google Analytics campaign tracking to all your links in all your campaigns. So my tip is very simple—tell them to do this for you. If for some reason they can’t do it or if there’s a fee for it, well it’s ticking a box in the background, they shouldn’t have to charge a fee for this and their solution isn’t that advanced, or basically, they’re trying to hide your data from you. That’s not good. Recommendation there: switch to their competitor who can do this for you.

Slide 17:

Don’t forget please—about your email display ads or your paid campaigns there—don’t forget about campaign tracking for social media. So many companies don’t worry about it, but they say, “It’s fine, it appears as a referrer in Google Analytics.” And while, yes it does, but only if the visitor clicks through from the website and there’s no redirect in the process. If the visitors click through from an app, from the Facebook app, Twitter app, or anything else, there is not referring website. It’s an app, and therefor appears in Google Analytics as direct traffic or as last direct click there on the previous traffic source. What that means is, you don’t get full credit for the work you’re doing around social media. If you want that credit for the work you’re doing, you’ve got to know how to actually improve it and make it better for it as well. Invest that 30 seconds per link under the campaign tracking.

Slide 18:

Moving on, Google Analytics configuration—mostly around some goals and setting it up.

Slide 19:

Simple one, first of all, hopefully you create goals for your macro conversion actions and your micro conversion actions. Don’t forget as well, you can also create goals for negative website experiences, so creating a goal for viewing 404 error pages, creating a goal for for validation errors. Then set a target on this: if goals completion rate—basically the conversion rate for these goals—is above X percent, drop everything else you’re doing and take some action. Set your own limits for this, but it should definitely be below five percent—I mean two percent, three percent. If fiver percent of your visitors are seeing an error page when they come to your website, you’ve got some major problems. Forget all the other work you’re doing and fix that problem first of all.

Slide 20:

One of the frustrating aspects of Google Analytics is the fact that it’s all based on last and direct click. And it’s frustrating to me that you don’t know how many sessions you actually get from different channels; from organic search, paid search, social media, everything else. So This is a trick to get that information quite easily. What you do is create a goal for All Sessions. It’s a detestation goal and it “begins with /”. That matches all landing pages or pages, therefore the number of goal completions pretty much matches sessions. Then you go into the assisted conversions report within multi-channel funnels, you change that conversion type so it’s just that All Sessions goal only. And what happens then is that Last Click metric now displays sessions. This report shows you then by channel—or medium, or anything else there—the number of true sessions per channel. Compare that against you channels report and look for the difference between last click and last non-direct click.

Slide 21:

Funnels. There’s apparently a bit of backlash right now happening against funnels, but to me, they’re still really useful. Whenever you create that big sort of goal on the website, you ought to include a funnel in that as well, stages of the process. One of the problems though, is that they can’t be segmented. For me, I actually want to look at segmented funnels. The trick to do this is: rather than creating just one goal for the end point and create a funnel leading up to it, create a goal for every step in the process. What you get from that is a horizontal funnel. Look at those screen on the bottom of the page here, I’ve pulled some custom reports with dimensions: channel groupings, device categories, countries. I can see for each one: how many sessions, what portion of the session is going to see an ecommerce visit, what portion would go to product, what portion create a cart, commence checkout, place an order. I can follow performance across that dimension. That’s really powerful at identifying problems, bugs in your website, and potential opportunities. But I want more.

Slide 22:

So better still, use the new feature in GA: Calculated Metrics to calculate the completion rates for each funnel stage. Based on the goals you’ve already got there, create calculated metrics based on goal Y over goal X, goal two above goal one completions, goal two completions above goal three completions, goal four completions above goal three completions, and so on. Then you create custom reports with these calculated metrics instead and actually show progress through the funnel at each stage. The comparison then becomes so much more powerful. You can see in this report here: desktop, mobile, tablet for an ecommerce website. Sessions, how many get to the ecommerce part, how many from ecommerce to product, product to basket, basket to view basket, view basket to commence checkout, checkout to placing an order. What this shows you is that while performance may be exactly the same, comparing desktop and mobile for every stage, it’s different in one stage, that product to basket stage. That’s the problem area where only one stays in the process. On that blog post, we have more information about how this approach, how to do a lot for yourself.

Slide 23:

From all that, what’s more useful information you can capture within Google Analytics?

Slide 24:

One thing I always do—I love to rename the pages in Google Analytics. Pages names have to have three features associated with them. One is: they must be unique. One page name per page on the website. Two: there has to be a hierarchy in place. Pages must be sort of grouped in a very logical manner so you can apply filters and look at deeper and deeper pages within a category. And thirdly: they must be very intuitive. Anyone in the business, from CEO down to the brand new intern, should be able to see a page name name and go, “That must be this page on the website,” without having to think about it. The page names you get by default in Google Analytics are based on URLs, and often, they don’t meet these rules. They are created default to the CMS or designed for SEO purposes, which is fine, but not great for analysis. What you do is you rename all your pages in the tags in Google Analytics. Example being there: the auto name crafter being used—or Google Tag Manager is even easier. You want to make these pages as useful and user friendly as possible. The best example of that: what’s the homepage called? Forward slash homepage. Everyone knows what that means.

Slide 25:

While you’re doing this, group the pages as well. Capture the page type as the content grouping in Google Analytics, as per the example on the right hand side there. There are different uses for this bit or information, it’s great. It’s especially useful for seeing website entry points, so looking to see how they actually got into the website. I mean, if you can’t tell me what point people enter your website on the homepage versus any other page, you don’t know how to do your business. It’s also great for understanding website navigation at an aggregate level. People that click from product list A through any product page. I don’t care which one it is, I just want to know they’re going from this product on this page to any of the product pages from that. Again, to see the flow through your website it’s really powerful just grouping pages.

Slide 26:

Talking about sort of grouping, best type of segmentation possible is to look to look at different types of visitors. Simple way to do that is just split them up into prospects and known customers. Capture this information in a session scope custom dimension and use it to understand performance, how behavior differs between your prospects and your customers. Say your conversion rate is five percent for the entire website, that’s pretty good, sit back, pump money into our marketing and take a look at whatever comes through to you. But if you add this segment to it, you might find that your customers love you and convert 40 percent, but your prospect only convert at 0.5 percent. You’ve actually got some major issues with your website. People who are new to the website, aren’t used to it, don’t know how to use it. You need to get in, fix the website, fix that conversion, so you make that money to bet better money for the marketing. You’ve got to get that segment applied, this is complicated to set up. The approach we use uses two different cookies.

The full instructions for this are in the blog post listed there. Prospects and customers is just a starting point. Even a lot smarter still, imaging segmenting your customers into loyal customers or discount shoppers or possibly different marketing personas, readily available information.

Slide 27:

Next one. I mentioned before product availability. Here’s the approach for that. People can’t buy what’s not available to them. So what I do is record the percentage of product variations that are in stock on a product page. So product A there, there’s 14 sizes for that product, only one’s available, therefore it’s recorded at 7 percent availability. Product B, 4 out of the 14, 20 percent availability. Product C is at 25 percent. D is 100 percent. What this shows you if you’re looking at your product performance—I hope you all laugh looking at an ecommerce website—one of the key reasons why certain products underperform and could do better, is the availability. That could start telling you how much you should be investing in your inventory to get product availability up to drive back that product performance.

Slide 28:

One frustrating aspect of enhanced ecommerce is it doesn’t actually expose the basket value or any pricing data, prior to purchase. These things need to be captured, but it’s not displayed anywhere, not available at all. So what you can do is create a new product scoped custom dimension and record this: the value of all the products as they’re added to the basket, then you record the products removed from the basket as a negative value, and also record the products in a transaction, also as a negative value. This custom metrics then gives you the value of abandoned baskets on the website. Share that with our senior management and see their reaction at that point in time. Caveat for this metric: the value doesn’t quite calculate the purchases correctly across sessions, but it’s still a very good indicator of how much money you’re leaving on the table.

Slide 29:

Useful reports you could be using.

Slide 30:

My favorite one in Google Analytics is the entry point versus traffic source report. For this one here, go to your landing page report, change the primary dimension to the page type content grouping you set up earlier—hopefully—click to change it to a pivot report, change the pivot by medium. I prefer channel, but it’s not available in this sort of report, I believe it actually is in the custom reports. You get your sessions as your quality metric by default, add a second metric, a quality metrics such as bounce rate or conversion rate, and what you’ve suddenly got is how they got into the website, what the entry point is or which source they’re coming from, with quaintly and quality as your metrics. If you’re looking at this and you can’t find any insights, you’re really not looking hard enough.

Slide 31:

Going on further—performance diagnostic report. Every row here is a different metric, every column is a different segment. This is a big A3 Excel file printed out with small sort of details. But you’ve got every row, a different metrics: traffic metrics, engagement metrics, your funnel, your sales metrics, macro conversion actions, micro conversion actions. Your segments, your columns: old visits, new, returning, prospect customer, device type, location, browser, traffic source, entry point, everything’s there. So get it all out there, define internal benchmarks for what can possibly be achieved for each different metric across segments. And look for any bugs in your website. Look for any big opportunities of how you can make more money. An online version of this is being developed. Currently it’s in beta. Look up profitgrid.io.

Slide 32:

Tactical reports as well, go from the performance diagnostic into tactical. This one here is one for ecommerce and the key metric there is the product page success rate. For product A and B, they’re the product you know about already. They get the most views, sell the most units, make the most money. That’s great, but I want to know what could have made the most money, not what did, but what could have made the most money. Looking at product C, it’s got a product page success rate of 6.3 percent. Other products are being added to basket at a much higher rate, around 20 percent, this one’s definitely underperforming. It could be because of that poor availability, a bad photo, bad ad copy, bad reviews, price too high. These are the things that which the merchandiser on your team should be fixing every day anyway, that’s their job. So make it easier for them, make it smarter, give them the report which tells them where to look, which products need to be fixed.

Product D, similar issue—20 percent checkout completion rate. It should be 50 percent, so something’s wrong here like shipping costs or time for delivery. Product E is reversed, look at that 47 percent, it’s in the sweet spot, price is right, it’s got stock, great ad copy, great ad photo, it’s all sort of looking good right now. If you needed more, put that product on the home page, social media, newsletter, and watch your sales go up straightaway based on the data. Make smarter decisions.

Slide 33:

From publishers and content reports, we’ll look at a tactical report there as well. Looking at your content, your articles, basic difference between read, shared, commented on. That second article is popular, but not very effective. It’s not very engaging. People are reading it, sharing, it, commenting on it at a very high rate compared to those internal benchmarks. Whereas article number 11, way down on the list there, is getting viewed a lot less, but has a high read rate, massive share rate, massive comment rate. That’s the article to promote based on the data. Put that on the homepage, social media, newsletters, and you can guarantee you’ll get more engaged readers for long-term success.

Slide 34:

One last one to go.

Slide 35:

If we ever have the argument Adobe Analytics versus Google Analytics, one big point the Adobe fans always bring out is hit based segmentation. And it’s a really powerful feature, I admit that. In GA, you’ve got user based, session based, you don’t have hit based, but we actually do, hidden away. What hit based segmentation does is scenarios like, when the category is visit, I want to allow sessions in which someone watched a certain video, and you try using events category: video interaction, event is another interaction, label is a video name. So if you create the segment A there, it returns all sessions where the visitor saw 75 percent of a video and had an interaction with product A video.

That’s not good enough because they could have watched 75 percent of a different video and pressed play on product A video. It doesn’t really match our need. So the trick here is to switch across and look at sequence segments, then within a single step, put the factors in there. This now returns all videos where the visitor saw 75 percent of a video and had an interaction with product A video in a single GA hit. So basically, this returns all sessions where the visitor saw 75 percent of product A video. That’s the one you want. Hit based segmentation is possible in GA.

Slide 36:

That’s 25 tips, tricks, hacks, within 30 minutes. Thank you very much. There are my details, you can find me in numerous ways. Thanks for listening and looking forward to seeing your questions.

Previous Flipbook
How HPE Cleaned Up Their Analytics Across Business Units
How HPE Cleaned Up Their Analytics Across Business Units

Learn how HPE used ObservePoint's WebAssurance to improve their analytics across their business. Fill out t...

Next Video
David DeVisser - Stop Analyzing Devices—Start Understanding the People
David DeVisser - Stop Analyzing Devices—Start Understanding the People

Devices don't buy things—people do.