A (2:41)

There's one thing I'd add in there. Very often people are using social media platforms to help share content and push it out there. And most of the social media platforms your audience will by and large be consuming on a mobile device. And when people are doing that, you know, they'll see something that piques their interest, they'll click on it, they'll have a read of whatever it was you promised and then they'll leave. Because actually it was just a quick flyby and grab type, type session. And so if you're, you know, if that's a major way of you reaching your audience and pushing stuff out there, you will, you will see that. Also see similar thing very often with links in emails as well. Works in the same way. So, you know, you can't just assume, as people used to, that bounce rate is a really bad, bad thing for certain pages or if certain content's been shared in a particular, particular way and has been successful in that you will naturally see quite a high bounce rate.

B (3:32)

That's it. And I think when I was reading an article that you highlighted for me, Kieran, the other day on Unbounce. So Unbounce is a kind of landing page website and tool and we're going to talk to them again very soon. But very long article saying what's the value of content marketing? Because what happened is they'd looked at the conversion rates of their content marketing and seen they were really low. And actually what the conclusions were really to some extent. I mean, it's a big, big in depth article. But it basically came to conclusion that yes, a lot of people come in, they leave, they don't convert, but then they come back via another channel later on and they end up converting. So it is about that branding piece to some extent. So do not read too much into bounce rates. You got to look at it in context. Now you might then say that leads me onto my second one which is time on site. So you might go, okay, they've bounced, but I want to see how long they're on the website. If they're on the website for a long period of time, that's fine. There is a fundamental problem with how analytics calculates time on site.

A (4:21)

There is. It's flawed, Dan, massively flawed. Tell us how.

B (4:24)

So what happens is that analytics only really sends any data back to from your webpage to analytics when the page loads, generally speaking. So if I go onto a page, the analytics code loads and it sends a message back to analytics saying someone's arrived on this page. Then when I go to another page, the analytics code loads and it says someone's on this page and it knows it's the same person because my cookie's there now. The only way it knows how long I've been on page A, the first page I get to is by what time I arrive on page B. Therefore, for example, if there is no page C, it has no idea how long I'm on page B and it doesn't count it. So for example, if you're a bounce visit, so say you've got a 60% bounce rate on a page, that means 60% of people are entering XD on the same page. There is no page B. That would mean 60% of the people, you are not getting any time calculation for how long they're on that page. So the reality is in most cases time on page is actually underestimating pretty fabulously how long people are actually on your pages because the last page you're on is very often the page run for the longest, not always, but very often. So there is this kind of built in limitation. And the reality is that that might sound like a bad thing, but it's better than you leaving a tab open and it just keep on counting up the whole time the tabs open and so on. So it's a slight problem. And it means that you can't calculate how long someone was on a page when they've bounced.

A (5:46)

If somebody's on a particular page for longer than half an hour as well, it does 10 times out and assumes it's a new session. So you've got all these, these little caveats with analytics rather than free ca about it though what I would say is, look, it's wrong, but it's fairly consistently wrong because it does measure things in that way across the board. So you know, if you're comparing like for like, which is one of the great, great joys of analytics, you can compare one month against another, you're going to have, you know, consistent inconsistencies. That's a good thing, right?

B (6:16)

Yeah. I mean, it's the same with email open rates, they're not actually very accurate because the fact that if you open something for a second and then delete it, it still counts as an opened if it's loaded, this little pixel that calculates it. And equally if you don't have images loading when you load your email, never calculate as that will never count as an open as well. But open rates, it's apples to apples. You're just comparing like for like. So it's inconsistent by roughly the same amount each time. So we don't worry about it too much. So it's like measuring the least wrong thing, as you call.

A (6:42)

It's a lovely image I've got there of the number of people that open up the email. And you think that's good, they opened it, but actually how many of them are screaming, for crying out loud, I've asked you not to send this to me. You know, where's the unsubscribe link?

B (6:54)

So on from bounce rate and time on site to one that I think is. It looks brilliant. And I teach lots of people analytics and they open this report, they go, ah, this is exactly what I needed. And it's the user flow. And user flow is a kind of visualization in Google Analytics that shows you how people travel from one page to the next.

A (7:09)

And so just describe it to us.

B (7:11)

So it normally starts with countries. So you've got this many visits from India, this many from the uk, this many from Germany, and then it shows you little kind of where they went next in groups. And then it says this many people got to the homepage, this many people got this particular piece of content, where did they go next? And you can keep highlighting through it. The reality is it's a lot less useful than it first seems. I mean, you might identify as things. Well, most people from the homepage go here, here or here. But actually you could do that in your popular content report to some extent. I mean, there's things around that it's not bad, but I've never really found that anything that actionable in it.

A (7:43)

Well, because there's too much data in it.

B (7:45)

Exactly.

A (7:45)

And actually you can't see it all in one go. You have to keep on sliding and zooming out. I do quite like to look at it from a landing page perspective because that gives you a rough idea of the route through the. And certainly I've used that in the past to highlight some user issues. It's also quite useful if you separate it out and just look at the mobile visitors segment or I think this is it segment, because then you can start. But you do need to get a bit more granular on what you look at. If you look at it in aggregate, it's kind of.

B (8:17)

It's a bit overwhelming and it's a bit of a kind of spider's web when you get into it. So I think it's one of those reports, there's a lot in it, but it's not one that you can just use off the shelf. And it takes a bit of knowledge in terms of segments and things like that to look at it as well. Now I'm going to be a bit controversial. The next one. I'm going to say that search console reports are massively misleading. So basically you might be aware if you connect Google Search Console up to your analytics, they then give you a load more data about what keywords people have been searching for in what countries and what devices. The thing is, it's lovely but a.

A (8:49)

Sense of but coming.

B (8:50)

Yeah. What happens is it says you had 50,000 visits from Google or you appeared in 50,000 searches in Google, but we're not going to tell you 22,000 of them, for example.

A (9:00)

Yeah.

B (9:01)

And what's happening is there's so many phrases you're not being told that I think it's slightly dangerous in that it gives you some data, but the data that's missing, we don't know what's missing.

A (9:12)

I've always assumed, rightly or wrongly, I'm about to find out.

B (9:16)

I'm about the look on your face. I reckon wrongly, because this looks like an assumption.

A (9:20)

I've always just assumed that they were very long tail search terms and that.

B (9:25)

I think there'll be a load in.

A (9:26)

It aggregated that all together.

B (9:28)

Well, the reality is that it will show, generally speaking, in the search console report, phrases up to about three or four words. It doesn't seem to show many more than that. And there will be a load of phrases, but in reality I'm not sure what percentage it is, but there won't be that many phrases that are five, six, seven words or more. I mean, there will be some, but I don't know what percentage it would be. But we don't know. This is the key thing. We don't know what's missing from that report and therefore we are probably reading a lot of assumptions into it that it's only very long tail or it might actually Be the phrase that you're getting the most traffic from is missing from it. There's no real guarantees on it because Google don't tell you anything of thing. By the way, the reason they don't provide all of it, our understanding is that they don't want you to better reverse engineer back how Google works, which is perfectly kind of understandable.

A (10:10)

If you get too good at getting free traffic, you don't pay for traffic.

B (10:14)

Exactly. So there's a commercial objective there. But I think it's a good report to look at. But I would also think beyond what it's telling you because you might miss some real kind of nuggets of keywords and phrases and things like that in there as well. So I would always use an external tool like SE Ranking to actually say that's telling you what you are getting traffic for what you are appearing for in search results. It's not telling you what you're not getting rankings for. So what you really need, if there are certain phrases that are important to you, you really need to make sure you're monitoring those within the search engines themselves rather than just looking at your traffic levels.

A (10:44)

The problem with tools like SE Ranking is actually they rely very heavily on Google Search Console data in a lot of cases.

B (10:52)

But I mean there is the key thing of actually look, I want to rank for this phrase and I'm not ranking anywhere for it, I'm not getting traffic from it. The reality is that as I start going up, I'm not really going to be clear on that until I've got an external report. So again, it's not a bad report. There is some stuff in there, but you need to actually supplement it.

A (11:09)

Now before we move on to the next one. Just wanted a little digression. Very excitingly, Search Console have released a new version which check your emails. They seem to have been sending emails out over the last six weeks or so to various different people with accounts. They invite you to try out the new version, which I think the most exciting thing about that is Search Console. One of the big problems, problems I've always had with it is it only has the last 90 days worth of data and outside of that you don't see anything. And that's where if you've got a search engine monitoring platform you can store a lot more data. But actually what the new version of Search Console does is it actually goes back 18 months, which I think is great. So you can begin to start like year on year comparisons and the like. One of the things that Google do Say when they send you that little email is not all the reports in the old.

B (11:58)

Yeah, don't panic because it panicked us.

A (12:00)

Yeah, it did. Me too. Then not all the reports that are there. You can still get to the old version and you need to use them both together because actually when you first start looking at reports, it's horrible. But if you haven't checked it out, do you do need to get that little email inviting you to look at it first? But it should be, watch out for it if you haven't had it already.

B (12:18)

Right. Last one I will talk about for now, but I would like to look at the goal reports. So conversion reports under goals. And when you look at your goals, it'll tell you how many goals you get, but it will tell you the goal source. And this is probably the single most dangerous report in Google Analytics. So you've set our goals up and it's telling us, you know, are people doing the things we want them to do? Great, okay. And then it will tell you where that goal came from. So I would go. And it goes, oh, you, most of your goals have come from organic search. And I'm like, that's brilliant. I am going to focus more and more efforts on organic search because it's working really well for us. But it's last click attribution and it's not that clear really. You know, if you're not used to looking at this, it doesn't really clarify for you why it's nominating that channel. And it's basically saying this, this click that converted came from this channel. And it's probably not something you want to look at. You want to look at the multi channel funnels because multichannel funnels will tell you actually the journey that ended in this conversion. What were all the steps that were taken along the line there as well? So I think it's one of those things that it's not clear when you start using analytics, the whole last click attribution piece. And actually I don't think it's that useful because we know from our website that a lot of people will come from an email, they might come from social media, then they do a Google search, then eventually come in directly because they know the website, their browser is remembering it. And if you look at this, it tells us that 60, 70% of our conversions are coming directly, which is just, it is true in the last click, but it doesn't tell you the whole journey. So it's just again one of those reports that's a little bit misleading. And unless you understand it fully, you're going to come to completely the wrong conclusion.

A (13:50)

I agree. And one of the big lessons for me with analytics is how much the various mediums within analytics overlap. You know, there's no one customer that is only an email visitor or they're only even down to devices actually, or I only visit on a mobile device. I do both. And you have to bear that in mind and read all of the reports with that context in mind. Because actually, you know, I will do an organic search and find a cool company and then maybe later on I will click click on a pay per click ad and buy. I think traditionally we're all pretty lazy, especially when we get on our mobiles. You know, we've got quite heavy thumbs. They tend to settle on the first thing that crops up, which hey, guess what, in most cases that'll be a PPC ad. So you need to understand these contexts and apply that when you're reading these reports to really inform what your digital strategy is going to be.

B (14:43)

That's it. And the conclusion I would draw from this is that you can't look at one piece of data. So what you really need is, is a kind of dashboarding approach. There will be key reports you look at every month. But the important thing is iteration. It is that you should look at your analytics, you should draw a hypothesis from it and you should say, I think this is telling me this. I think this landing page needs improving. I think I could get more traffic from Facebook, I think our traffic from Twitter is great, but it's not convert, whatever it may be. And then you say, right, I'm going to make a change off the back of that. I'm going to tweet more, I'm going to tweet less, I'm going to do better tweets, I'm going to change the landing page, whatever it may be. You make that change and then you go and look at the analytics again. And that needs to be the process. There is a analytics being used for iteration approach and there needs to be a dashboard that you look at or some core reports that you look at that you just get into this process. Because otherwise what happens if you're not careful is you start off with this kind of concept of marketing mix, which we've spoken about in marketing for years, which is how much should I be spending on each of my different channels, how much my effort should be split? What's the perfect marketing mix for an automotive company or a financial services company? And there's no answer. I can kind of give you best practice. Well, we know that this works and that works, but then you start off with a mix of different channels and then you need to start iterating, squeeze more and more out of those channels and too often people will say things like, well, we were doing Twitter and it was working. And the thing is there are so many variables at play. So the quality of the tweets, the timing of the tweets, the frequency of the tweets, the number of Twitter followers, they're all variables you can change and then you change each of those and you see what impact they have. But bear in mind, your changes to Twitter might be directly impacting your search traffic because actually people are reading stuff on Twitter, they really like it, and they go and search you in Google. So you need some way of identifying the connections between your channels. And that's why dashboards are pretty important with all this. So the idea of a dashboard is that you work out what are the key measures for each of your key channels. So for email, search, Facebook, whatever it may be, and then how many conversions you're getting and then are those channels contributing towards those conversions and you can get that contribution percentage from that multi channel funnels report in Google Analytics we'll put the show notes, there's a link, we've got a description of the dashboard that we've built for a couple of different things, so we'll put that into the show notes anyway. But just think about the fact you need to look at a set of reports. Really where analytics then comes in useful is you've got your kind of high level dashboard data and then you jump into analytics and try and analyze it a bit more and you might segment out some data, you might look at particular campaigns, you're having a dig and you go, I think it's telling me this, but very often the clarity that we'd kind of expect from having all this data isn't really there. And the more I've used analytics, more we've got into analytics and looked at it for hundreds of different companies. The reality is that it's not quite as straightforward as it first seems. Just because the user journey is not as linear as we would expect it to be in any way, you're jumping around. People make weird decisions, they do things in weird ways. Think how complicated our day by day lives are, multiple devices, multiple channels. So therefore you can't look at the kind of individual journey, you have to kind of aggregate it together and start making a hypothesis from the whole thing as well. But the other Thing what I'm saying, you can't do it an individual level. You've got to be careful because analytics doesn't tell you about intent. It doesn't tell you why people are doing things. And this is something that Kieran's raised quite a lot. And the fact is we need to try and understand why people are doing things to understand what it is that's turning them on or off about our particular content.

A (18:00)

It fascinates me. And I've shared this or not before. It was an amazing piece of art that I saw in London, which was a giant orange canvas, actually. And do you know, for years I've despised modern art and just think it's rubbish, doesn't mean anything, it's just meaningless. And that for me was the epitome of bad modern art. But when I actually took the heathen. I know, right? But when I actually took the trouble to read the little typed blurb next to this enormous 20 foot orange canvas, what was really interesting to me was the artist was really interested in the effect that the orange color he'd painted had on the space between the user and the painting. And I thought long and hard about this because it challenged me.

B (18:44)

This is very profound. All of a sudden.

A (18:46)

No, no. So I'm going to share this with you because I've been on a bit of a journey with this particular painting. And yeah, it's interesting because I think as digital marketers, we have details of our customers activities on our website, but we are missing that space between the customer and the click. So actually in this analogy, the click is literally the canvas that we can see. But actually what we need to be analyzing is that space between the customer and the click. And just because I clicked doesn't necessarily follow through that, you know, why I clicked or what was going on in my head when I clicked or where I was when I clicked. And this is, for me, this is some of the big unanswered questions that when you're aggregating data together, we need to sort of make some attempt to fill in.

B (19:34)

Yeah. And I think we've used tools like hotjar that have filled in some of the gap. I mean, two techniques really. One is speaking to people, is asking them questions.

A (19:40)

Yeah.

B (19:41)

I also think you've got to be careful of not listening to one person too much because someone will tell you, oh, I don't like this. And you go, right, they didn't like it, let's change it. Oh, they didn't like that, let's change that. And you end up Just being a kind of running in circles, but also hotjar, you could see people using the website and there's one bit of our website which is the qualifications bit that people getting really confused what they're actually getting. Which is why we've then gone on to say, right, we will offer full proper CIM chart and Egypt marketing qualifications to make it a lot clearer about what we're offering. So the reality is that actually you could start to see that space a little bit more. There was like, well, I want to get qualified, but I don't really understand what on earth you're offering. And I can't really get if this, this is going to make me kind of recognized in my industry. And you could almost see that in the mouse movements where people just kind of going around and around and around in circles in this one bit kind of a bit confused. So I think you can look at analytics, you can draw hypotheses, you need to then do things, change it, but don't make it your only source of data. You want to probably speak to customers, people using a website, tools like hotjar, screen recording, AB testing, all those kind of things are going to be really important as well.

A (20:37)

And user testing and user journey and, you know, things like what users do, which is very good. You know, you get recorded sessions of targeted people you've asked to analyze your website and they actually speak their thoughts. All these qualitative sources of data are hugely useful for filling in that gap between the customer and the click. And I would say, yes, it's a fair bit of effort to go out and get these things, but my goodness, can you supercharge and multiply the results you get back if you actually find the key things that are getting in the way and improve that user journey? It's quite phenomenal. I'd say it's possibly one of the most valuable things you can bring to the party as a digital marketer. So ignore it at your peril.

B (21:22)

Exactly. So good luck with all your analytics efforts and we'll speak to you again on the Digital Marketing Podcast.

A (21:35)

Thanks for listening to the Digital Marketing Podcast brought to you by Target Internet. If you're investing in your digital marketing skills, take a look at our free benchmark skills test and look at the wealth of online learning we provide. To help marketers get up to speed and stay up to date, just visit targetinternet.combenchmark.