April 1, 2020
Episode #7 – Web Analytics TV With Avinash Kaushik and Nick Mihailovski

Episode #7 – Web Analytics TV With Avinash Kaushik and Nick Mihailovski


Nash: Hi everyone. Nick: Hey. Nash: Welcome to the next episode of Web Analytics
with Nick and Nash. That’s the thing were going with. And first before we get started we’re thanking
our good friend Michael for sending us a lovely card; beautiful, gorgeous; and also sending
us two beautiful buttons, that say “I’m complicated.” They were so good, Michael, somebody took
them. So we sad; we we’re gonna pose – Nick: We we gonna wear ’em. Nash: with them today, but we don’t have them. Nick: We have the card, thank you. Nash: So all of you feel free to send us stuff.
Okay. [both laughing] Nick: Okay. Nash: Alright. Here we go. Let’s get going. We received see some questions for you, Nick.
So let, let’s get started. Nick: Sure. Nash: More technically oriented questions.
The first one’s from Matt in Ohio. It says, “When will Google implement a better method
of testing? Tried and wait 24 hours plus doesn’t work on a truncated timeline.” And so, actually
that’s a good question. Nick: It is. Yeah. Nash: And so what, what are the options there? Nick: Right. So at Google Analytics you have
a piece of JavaScript or, or some tracking code that will send data to analytics and
so what you can do instead of waiting for the reports to process the data, is capture
the data that’s being sent and insure that that’s valid. So using either Chrome Developer Tools, or
Firefox has a Firebug tool, you can actually see the request. It’s _ _utm.gif, which is
the one by one pixel image that we sent to analytics and a bunch of parameters. And we’ll link to a document that describes
how you can troubleshoot; that way you don’t have to wait the 24 hours for the data to
process; you can look at it right away and troubleshoot. And that’s really the way that
advanced users will troubleshoot installations. Nash: Exactly. And, and for most people actually,
if not for everyone, the data in Google Analytics is processed a much more faster so I think
publicly it has been stated that the data in Google Analytics will be about an hour
old. Just make sure you click on the little clock that brings – Nick: Uh-hum. Nash: all the day reports and splits it by
hours. So the data should be at most an hour old. I know that it’s actually turned faster
than that ->Nick: Right. Nash: but at least publicly the commitment
is in hours, but using Chromeback and other developer tools, fantastic. Nick: Yep. Nash: Genius idea. The next question also for you Nick, from
Michael in Barcelona. And it says, “Can you please offer ga.gs with gzip compression,
Google’s own page speed tool gives the alert “a ga.gs file is not gzip and it slows down
page load, although in a very, very, very, very, very, very small way. Thank you, Michael. So, Nick, what’s up with this zipping? Nick: Yeah, so everything we serve so ga.gs
is served compressed. gzipped, every time. There is a tool page speed that we ha, are
working with engineers to make sure that they’re reporting it properly, but you can be assured
that’s always gzipped and you can look at the headers and make sure it’s being sent
that way. Nash: So async rocks and if those of you on
the video not using async already, please get on the band wagon because it does two
wonderful things. One is it will collect more data, so your data will be more accurate,
plus the loading and back and forth happens really, really fast, so please go ahead. If you’re still using urgent.gs, jump directly
[laughs] to async. Nick: To async. Yeah, don’t use our urgent.gs. Nash: [laughs] Using urgent.gs also jump to async because
it will give you more accurate data. Here’s the next question from Ben in Salt
Lake City, Utah. “I would like to use site custom bar at a visitor level for first [inaudible]
attribution without losing last [inaudible]attributions. To do this successfully I would need to see
if a value exists so I don’t know, so I don’t know how to write it. How might this be accomplished?” I don’t even know what he’s asking for. Nick: Yeah, well he’s asking a lot of questions,
but you could actually for visitor level session variables or for custom variables at the visitor
level, we save them in a cookie and we have a method and a tracking code that you can
extract that value – Nash: Ah. Nick: You can save it in set intract. So we
already have it; we’ll send a link; you can start doing this; we’d love to hear how it
works for you. Nash: Exactly. Check out the blog post on
analytics.blogpost.com where this video is embedded. Because the links are not in YouTube,
the links are all on the blog posts. Nick: On the blog, yeah. Nash: And we’re gonna post it. Good, good. Okay. Nick: So it looks, time for a question for
you. This is from Pearl D. from The Netherlands.
“Can annotations be added to analytics automatic through the API?” Nash: [laughs] Nick: “Another way, for instance, a URL request?
When sending a newsletter for instance you can make the annotation by the system instead
of manually.” Nash: Oh, that very clever idea. So one of
the great things about the team that’s doing API with, that Nick leads and a bunch of other
people are there, has actually created an open forum where you can actually go in and
rate the features that you find important or the, the things that you want the API team
to prioritize higher. So in this blog post that goes with this video
we’re gonna add a link to that particular forum where you can go and vote and make your
voice heard. Get all your friends to vote for this feature and boom it’s gonna come
out. Nick: That’s right. And with the API we’re
really tryin’ to prioritize on what people want, so let us know how this is gonna help.
Vote and we’ll help to prioritize. Nash: Good. Nick: Great.>Nash: The next one’s for you as well, Nick.
And again from our friend, Michael in Barcelona. “Why does traffic from Google Images still
appear as a referrer instead of a search engine?” This is a good question. Nick: It is. I think um – Nash: And, and I think it’s, it’s that you
know very, I can see my top referrer reports or referring URL’s I can see Google Image. Nick: Um-hum. Nash: While some people would like to see
it under the, the search engines report. Nick: Right. So the, the issue is when you
search from Google Images and you click an image it actually creates a new page hosted
by Google and loads your page in an iframe. Nash: Ah. Nick: So it’s not like a standard search;
it’s slightly different and that’s why it looks like a referral. There’s a couple work arounds that you can
do. We’ll, we’ll link to one solution. The other way is definitely talk to a GAC or one
of our authorized consultants; they can help you out. Pretty much what you wanna do is if you could
identify that the page that’s being rendered for your site is coming, that has a parent
page from Google Images, you can overwrite the referral information; set a, a custom
search engine and you can add a new search engine and it’ll appear as organic. Nash: Oh. Nick: So there’s a couple of ways to do it.
It’s a little complicated for this video. We’ll send a link to how some people have
done it before. Nash: That’s very clever. But I could use
it across all search engines not just Google to do this whereever, you know, that kind
of behavior is happening. Nick: Right. Where you’re loading in an iframe. Nash: Oh. Wonderful, beautiful. Nick: Okay. Nash: I like that. For example at Twitter when many people post
links they, it opens this thing on the top and then the Website is rendered under that;
particular sort of this tool bar type thing – Nick: Okay. Nash: I could bust that and use – Nick: As long as it’s, yeah with the iframe,
yeah, definitely. So it’s pretty flexible. So here’s a question for you. Nash: Ah. Nick: This from Alfire from Montreal. “The traffic source report shows my site as
a referral of itself. Why is that?” Nash: [laughs] Good one. Hello, Canada. There are a number of reasons why your Website
would show up as a self referrer. But let me just cover two of the most important ones
that sort of address 90 percent of the time that we see this happening. And the very first one is that your entire
Website is not actually tagged. Nick: Um. Nash: Especially your T landing pages are
not tagged, whether you created them for search engine, email or whatever. If, if I come to
a landing page on your Website that does not have the Google Analytics code and I click
on a link to go deeper into your Website on, on which there is actually tracking code,
to that tracking code it will seem that this is the first time you entered the Website
and you entered the Website from this landing page that had no tracking code. Nick: Um. Nash: So that’s the first most common issue
and make sure you audit your Website frequently enough that all your pages are tagged. So
that’s number one. In fact go check the page which you are seeing
in self referral and sure enough that’s a great way to find out. Nick: Look at the entrance point. Nash: That’s it exactly. Nick: Exactly. Yeah, Nash: The second one that also is very, very
common is using redirects of some sort. Nick: Um. Nash: And, and what happens is you want, let’s
say you had three pages, so you go from page one to page two to page three, but suddenly
you got rid of page three and, and you’re actually sending people directly. And, and in this case you are using a redirect
and we use redirect on landing pages, we use redirect in side to side views, redirect in
many different ways and what happens then is that if, if the original referrer gets
lost – Nick: Um. Nash: If the original referrer gets lost it
looks like the person came from this particular page; from their redirected self in the middle.
So make sure they use permanent 301 redirects which actually pass the referrer; the original
referrer, not your Website URL that referred the person. If you have an, if you don’t have,
if you have a temporary redirect you will notice that your Website has self-referring
URLs. So, those are two of the most common cases
why self-referrers happen and, and if, just make sure that you don’t have those on your
Website. Nick: Sounds good. Nash: Here’s a question for you Nick from
BBC Guru, Los Angeles, California. And it says, “How do you break out the different
local Google Websites, such as Google.com, Google.com/es, Google.com.uk and things like
that?” Nick: Right. Nash: Um. Nick: So this whole, the question is can I
add a new search engine to Google from the existing ones that we define? And you can.
There’s a method called add organic. We’ll send a link to it and pretty much in the tracking
code on every page you say, “add organic”, you put the value Google.com/es, the query
parameter and it’ll be a referral as a search. Nash: Oh, perfect, perfect. And also sometimes
on Websites some of these ads ask me like I wanna treat business.com as a search engine
referrer or something. In which case, the same way. Nick. The same method, exactly. Really flexible. Nash: Excellent. Nick: Alright, so question for you here from
Mike FK from Fremont; popular commentary here. “The site overlay report is a great tool within
Google Analytics, but was wondering why the tool reports the same number of clicks by
two same links on a given page. Does this not defeat the purpose of visually analyzing
clicks?” Nash: Yes, yes. And it’s a great question
and, and as you all notice site overlay is one of my favorite reports. I think it’s,
it’s great at visually reporting regardless of the Web Analytics tool you use. Check the
out the site overlay report in the tools. And it’s common that, that a, it’s coming
up. It’s actually common in all Web analytics tools. It doesn’t actually matter, use Google
Analytics or not. What happens is on, on one page you have two different links; up here
you have a link going to nick.html. Nick: Um-Hum. Nash: And you have another link in the body
that’s also going to nick.html. Nick: I’m popular on this page. Nash: [laughing] And, and what happens is
what your Web Analytics Tools, Omniture, Google Analytics, whatever it is, is, it, what it
reports is somebody went from this page to this page. What it doesn’t know the way data
gets captured by JavaScript is that you clicked on this link or this link; it, it doesn’t
know. And when it doesn’t know it displays both of the, the data in both of those places. So there’s a very simple fix to this. In any
Web analytics tool you use, and what you do is you make them distinct. And so what happens
is in this particular link when I link to nick.html, I’m gonna add a query parameter
and then I’ll say sc=1 because that’s the link on the top. Nick: Um-hum. Nash: And the second one I’ll it, I’ll add
a parameter to it nick.html parameter sc=2 and now to your Web analytics tool these two
links have become distinct. Nick: Hum. Nash: They are not the same link even though
they point to the same page. So adding this kind of query parameter is one method by which
you can teach the search engine that these are two different links. It’s a very common
method that is being used to overcome the challenge of site overlay in any Web analytics. Nick: It sounds pretty simple. So something
people should definitely do. Nash: It’s a very simple fix.>Nick: Great. Great. Nash: Here’s Andy from San Diego, Nick. Nick: Okay. Nash: Question for you. A good one. It says,
“there are current, there is currently no profile filter to include exclude traffic
using custom variables – Nick: Um. Nash: underscore set custom more. Now that
set var has been deprecated, how I can I exclude internal and dynamic IP traffic?” Good one. Nick: Right. So we had something called set
var, a user defined variable. We’ve deprecated that from the tracking side in replace of
these more powerful set custom var custom variables. We have a Webinar coming up. You should watch
it on that. Nash: Oh, yes. Nick: On custom variables as well, so some
promotion there. So in the meantime, while we’re waiting for
profile filters to, to support the new format, you can use advance segments. Nash: Exactly. Nick: And then you can just segment it out,
slice and dice, and it should work for you. Nash: Yes, exactly. And actually if, when
you run into these kinds of problems, explore the features that are available in advance
segmentation because lots of dimensions and metrics are available that allow you to do
things that you might otherwise have thought are not possible. So we strongly encourage you; we both love
advance segmentation; without it life is not complete. [both laughing] Nick: Great. So a question for you, Avinash.
And this comes from Melicene from Baltimore. “Site overlayer was working for us until
– Nash: [laughing]] Nick: November 2009. We haven’t added anything
new to our site in form of JavaScript and clearing browser caches. There’s currently
no overlay at all. How can we get this to work again?” Nash: [laughing] I know they actually working;
I was just looking at my own Website yesterday so I do know that it’s working. There’s probably
something unique about your Website that is causing this issue. Typically when you run
into problems like this we have four different methods in which you can get help. You can use the Analytics Help Center, there
are, there are articles that help you diagnose issues. You can use our Analytics. Web Analytics
Forum and there you can post a question and we have people from Google as well as other
people from outside who are actually gonna help you. There is our code site that helps
you understand some of the nuances of these things; where you can get help as well. And
finally, of course, perhaps the best route for these kinds of complicated problems, is
that you get a GAC. We have a, we have a army of fantastic, wonderful
GACs that charge very reasonable prices compared to really all of the web analytics providers.
And they will be able to help you diagnose these issues. So you have four issues, some sort of self-help,
some external help or user help. Please use one of those methods to find help for yourself. And we’re gonna add all of those links to
this video so that you can solve these problems that come up to you. Nick: Great. Nash: Okay. So the next question is for you,
Nick. Nick: Great. Nash: Another delightful question from Andy
in Toronto, Ontario. Hello, Andy. Nick: Hey. Nick: “Why do some pages show up under not
set for content by title even though they have page titles and appear in other areas
in analytics correctly? This is an interesting question, right there. Nick: It is. So fundamentally there is a property
in the browser called document.title and what we do is use that value when we send each
request to analytics. But whatever that value is we’ll report on it. And it’s strange because,
you know, with URLs, every URL is unique, but for titles you can have the same title
across multiple pages. Nash: That’s right. Nick: So it’s not necessarily a one-to-one
mapping – Nash: Ah, ah. Nick: as you you would think. Nash: I see, I see, I see. I see. Nick: And in some cases certain browsers,
certain circumstances we might not capture this value from the browser. And that’s what’s
typically happening. So what I would recommend is going back into the pages, tryin’ to see,
tryin’ to maybe look at what were the browsers, what are the operating systems; try to isolate
the, the challenge; think you know what was causing that to, to better diagnose. Nash: So you mentioned before like sometimes
in mobile phones when it is being captured it might not be coming out optimally or things
like that. Nick: Right, right. Nash: But let me, let me also take this opportunity
to pimp async code. Because, because sometimes if, if the page doesn’t finish loading completely
and I already kind of move away — Nick: Um-hum. Nash: Google Analytics or literally Omniture
or whatever doesn’t have time to capture all the data, Nick: Um-hum. Nash: but if, if, if you Andy switch to the
async mode you know it’s kind of up there in the head and it works asynchronously. So
your page is going fine, but actually Google Analytics collect data more accurately. Nick: Right. Nash: There’s a feature that’s only available
Google Analytics and we strongly encourage you switch to the async code and see if that
licks this problem for you. Nick: That’s a great suggestion. Nash: Right? Nick: Yeah. Okay, question for your Avinash for Jens Vitz
from New York. “I’m learning Web Analytics now and finding
it daunting. Nash: [laughing] Nick: I have Avinash’s book – a big fan
– Nash: Yea. Nick: and they are helpful.” That’s good.
“So what is the best way to approach learning web analytics without getting spun around
and around in numbers and reports?” Nash: [laughing] Nick: It’s like where to you begin? Nash: I know. It’s, it’s, it’s a good question
Jen and, and, and it depends. [laughing] Okay. But, but let me tell, let me tell you a, a
fantastic way that I use to focus my attention when I start web analytics things; and, and
the, the trick is money. [laughs] And by that, here’s what I mean. The two places
I will start my web analytics journey regardless of what side I’m working with, large or small
or otherwise, is the first thing that I focus on figuring out where is it that my company’s
currently spending money. So let’s say were running Nick and Nash, Incorporated. Nick: Um-hum. Nash: I’m gonna say, “Okay, Nick you know
you’re, you are our V.P. of Marketing; where are you spending money? Nick: Um-hum. Nash: And he’s gonna tell me “Well we’re doing
some page search; we’re doing some email marketing and we’re doing an affiliate.” Okay that’s great. That’s exactly where I’m
gonna start. I’m not gonna start with referrers; I’m not gonna start with organic search; I’m
not gonna start with any of those wonderful places, by the way. I’m gonna say, “Where are we spending money
today and let me try to understand if I can reduce the amount of money we have to spend
to acquire traffic.” Nick: Um-hum. Nash: It’s a great way to focus because any
improvement that you do, any insight you find from analytics will help you reduce the acquisition
cost. In this case my V.P. of Marketing is gonna be happy. Nick: Um-hum. Nash: The second area I use to focus is, how
much money am I making and where, and why not. Right? All related to the outcomes from
your Website. And we got another question outcomes that we’re gonna answer in a few
seconds, but I focus on outcomes and say, “Okay, how much money am I making? Are people
dropping off in the funnel? Why is it my conversion is not big?” So, so those two sets of questions: where
am I spending money and how much money am I making and why am I not making more money?
Really dramatically help you focus your attention initially. And make sure in both cases you’re
working on things that add to your bottom line. So that’s, that’s the first filter. If you wanna learn the other, other recommendation
I have is let me, let me make two other recommendations. One is we’ve got our analytics forum that,
that people are always answering questions and, and pointing directions and all that
stuff. But the other thing is Nick mentioned a few
minutes ago that we’re gonna do a Webinar.>Nick: Um-hum. Nash: And let me take a moment to promote
Webinars by all vendors, right? Google Omniture does great Webinars. Coremetrics
I think from time to time, not, not as much, Webtrans not as much. But at least Google Analytics and Omniture
do Webinars quite frequently. Look up Webinars that are done by these companies, because
these Webinars will have 20 percent of pimping and telling you how great the vendor is; we
do that too. Omniture does it, everybody does it. But actually many of, most of the content
80 percent is actually very useful and a great way for you to learn because we’ll bring the
best practices; we’ll bring and Omniture Nick: Um-hum. Nash: great best practices to you and that’s
another great way to learn. So, so that’s, that’s kind of what I would do. Nick: Great. Great answer. So here’s another question for you Avinash.
This is one from Leonard in Boca Raton. Nash: Wow, Florida. Nick: Yeah. “When I send a larger date range
and see fewer pages or key words in my report, why is this true for some accounts while others
roughly the same amount of transactions totaled correctly?” Nash: Yes, Leonard, one of the things that
happens is in, at, in Google Analytics, we want to give the answer as fast as possible. Nick: Um-hum. Nash: And so what happens is if you think
really large date ranges, three years or whatever, or, or depending on the amount of number of
visits and page views that you have on your Website, what happens in automatically trigger
sampling of your data and what sampling does is it basically says, “Don’t wait until tomorrow
morning to get your answer.” Nick: Right. Nash: Or for three years, like you might with
some tools. Nick: Um-hum. Nash: We’ll give you an answer very, very
quick; I will statistically sample the data and give you the best possible answer; will
tell you the error margins and everything. But when sampling happens on your data you
can imagine that each time Google Analytics intelligently processing in data and trying
to find you the best possible answer. And, and when you run the same report a couple
different times, or or, on number, on different number of days, there might be slight variations
because the number of rows that get picked to be sampled – Nick: Right. Nash: is actually a different one or slightly
different and that really causes this issue to happen. So if this is particularly egregious
normally doesn’t, either a couple two, three percent here and there, but if it is particularly
egregious in your case definitely reduce your time ranges – Nick: Um-hum. Nash: and, and use shorter time ranges and
that way of course a) you’ll be sampling less data or most likely sampling won’t be executed
– Nick: Right. Nash: and you’ll get the answer that you’re
looking for. So those are two methods that you could use
to overcome this problem. Nick: Great answer. Nash: Here’s another one for you, Nick, this
one. Nick: Okay. Nash: How about our favorite motion chart?
We, we did a music video of this. We should link to that. Nick: We’ll link to the music video – Nash: [laughs] We have a music video. Nick: Really, really great. Nash: [laughing] “What would be a good use of motion chart
bar graph? I can easily see the use for the motion chart bubble graph, bubble charts
– Nick: Um-hum. Nash: but I have a hard time finding use for
the bar graph. That, that’s a new feature we had introduced Nick: Right. Nash: a little while back. Nick: Right. Right. It’s actually pretty nice
so motion charts, one of the biggest values that you can use with this is if you have
a lot of different data, a lot different metrics, you can plot four metrics at a time using
colors, size, and x and y axis. And you can play that over time so through the normal
reports over time you’d have to actually select each one of them to see if there is a spike.
But by putting them altogether you can quickly see if something’s moving around a lot. Nash: You can time travel. Nick: Time travel through your data. Nash: It’s amazing. Nick: It’s really, if you haven’t seen it
you should definitely like play with it ’cause it’s really interesting. But sometimes when you have all these different
bubbles going around – Nash: Yeah, yeah. Nick: it becomes really noisy. Nash: Ah. Nick: And so one way to reduce the noise
– Nash: It’s cluttered. Nick: Right, is it’s on a two dimensional
graph. So one way is to keep one of the axis, we constrain the x axis and so only the y
axis moves up and down. So if you wanted to – Nash: I see. Nick: make it easier to analyze you go into
the bar chart and just see what’s going up and down; it might make the data a little
bit more clear. Nash: So another question for you, Nick. “For ecommerce Website that’s sells say more
than 20 items, what is the best practice for using goals. Use them, them for products or
categories? Would love to hear your thoughts.” It’s from Meir in Israel. And he also asked another question which was
“as practice with using goal refinement – any best practice.” So Nick I know you’ve spent a lot of time
in this area. Nick: Yeah. No, setting up goals is really
important because you can get different funnel paths, you can tie everything to a conversion.
For ecommerce transaction sites, definitely on the final receipt page don’t only just
make it as a transaction, but also make it as a conversion call. We just recently released two engagement type
calls, one of them is Time on Site. So what you might consider if somebody spends more
than 20 minutes, more than 5 minutes even – Nash: Yes, yes. Nick: there might be engaged customers who
are really going through the shopping experience; they’re not window shopping; they’re going
in the store looking at products. [Noise in background; door opening; someone
talking inaudibly] The other one to take a look at is pages per
visit. Nash: Just a minute. [voice inaudible in background; door closing] Nick: the other one to take a look at is,
for people who transact is to create a segment and look at what their average page is per
visit. Most likely it will be much higher than people who didn’t transact. What you can then do is set up a goal for
that high level of page per visit and then see how many people who came in could of actually
convert it; and most likely to actually purchase. Nash: That’s great. So and then we’ll link
to some articles that will be helpful. I also want to say we got feature requests
from Andy in San Diego, Paul in the UK and from Beiling in Neufchatel, Switzerland. And so we’ll pass those on to the product
team guys. Hope you had fun with this video and please
go to the door at the moderator site and submit your questions. And we look forward to doing Episode Number
8. Nick: Yep. Nash: Thanks so much.

Leave a Reply

Your email address will not be published. Required fields are marked *