How and what to do with your adwords campaigns in Google Analytics

Last time I blogged about how you set up your campaigns in HBX and Google Analytics.  Whilst I have given a bit of a description on what to do with your campaigns in HBX before (more on that one later), I thought this time I'd write about Google Analytics.  Specifically, having joined up your adwords and analytics accounts, what should you do with the outcome of it.  Hopefully out of all of this, you should be able to go and find out that critical information to calculate your Return on Investment.  That way you can work out how much more money you can spend before it stops making you more money than you are spending.



How Do I find out how many visits have come from my Adwords?

There it is in the left hand menu, we can see how many visits our campaign has had, along with a few other key metrics.  I always think that this maybe shouldn't be the screen that you get presented with, but you should see the 'goal conversion' tab.  More of that later though.



The advantage of linking your adwords campaigns to your analytics account is that you can view more metrics here than you could by doing each of them individually.  More to the point you have the ability to drill down into your campaigns and look at your adgroups individually.  If you have set up your adgroups correctly, you can then focus on each of these individually to optimise your keywords.  Having picked your ad group you can delve into the keywords themselves that are driving traffic:

The beauty of this is that you can take each of these keywords and try and optimise individually.  You know that you can do exact matches, broad matches and negative matches.  Start out with your broad matches so that you have a long tail of keywords.  You can then give the keywords that are performing well more money by splitting them out into a new group and giving them an exact match.



Why don't my visits match my clicks from Adwords?



This is a popular question.  These two systems may never match.  Fortunately I have given this explanation enough times that I can just copy and paste out of an email:

• Anyone who double clicks on a link will be counted twice by Google Adwords


• Anyone who clicks on a link and can’t connect to the page gets counted by Adwords but not by GA


• Anyone who clicks on a link, connects to a page and the GA tag fails gets counted by Adwords and not by GA


• Anyone who clicks on a link, connects to a page and then clicks on a link on that page before the GA tag has loaded gets counted as a click in Adwords and not by GA


• Anyone who clicks on a link, connects to a page, loads the GA tag, but then gets filtered out of GA because they’ve filtered out internal traffic from their GA account will get counted by Adwords and not by GA


• Anyone who clicks on a link, connects to a page, loads the GA tag, but then gets filtered out of GA because GA thinks that they are not a real person will get counted by Adwords and not by GA (although in theory they'll get filtered out of Adwords this doesn't necessarily happen)


• Anyone who clicks on a link, connects to a page, loads the GA tag, but blocks any cookies that GA attempts to give them will get counted by Adwords and not by GA

The funny thing is, with Google Analytics, you can tell.  You can find out what the difference is.  Click on the tab that says 'clicks' and you get your answer:

Again, you can delve down into this.  I like the way that it tells you per adgroup, per keyword.  You can delve down per keyword to find out which is the most competitive (ie you have a low click through rate - impressions divided by clicks and you have a high cost per click).  You should also be able to work out which is the least competitive, which should help you generate the most traffic from those.



How do I work out ROI?



The most important of the tabs is the one that says 'Goal conversion'.  This is the one where you work out if the money you are spending is making you more money than you are spending.  The first trick to this is to work out what you want your users to do.  On an eCommerce website, this is fairly obvious - you look at your page that says 'Thank you for buying my product'.  You set that up as a goal in Google Analytics.  How do you do that?  Well follow these simple steps:

• Log into your analytics account


• Click 'edit' on the settings column on your website


• Under 'Conversion Goals and Funnel' click on the edit to the right of G1


• Type in your conversion page ('thanks for buying') into the box 'Goal URL'


• Type in your goal name (eg 'Buying product A') into the 'Goal Name' box


• If you know how much it gives you in terms of value, type that in too in the 'Goal Value'


• Click the 'Save Changes' button at the bottom of the page


• Repeat for all conversion points


• View your reports

Now you have your goal set up, each of your reports now shows you all your goals too.  This should allow you for your adwords campaigns to do all the things you were doing before, but with real conversion points associated with them.



This gives you the real option of looking at each of your campaigns now in terms of conversion rate as well as clicks, visits, CTR, CPC, etc.  If you can see one of your groups isn't converting at all well, lets look at the keywords in that group and find out which ones aren't working, so that we can put a negative match in google Adwords.  We can take the ones that are converting well, look at the click through rate and cost per click to work out if there is any more scope to increase the volume without compromising the conversion.



More importantly, the search functionality will allow us to group our broad match keywords together and get a nice sum at the top.  We don't want to take all of those single click and single visit terms on their own, but being able to search our broad match will give us more knowledge of what is going on with it and whether it is converting.

Setting Up Campaigns in HBX and Google Analytics

Again, since I did a presentation on it last week, I thought I might as well blog about it as well. Last week I was talking to our conferences team about how they should use campaigns in HBX (although I'll explain about Google Analytics here as well) to work out their Return on investment (ROI) of their email campaigns. I thought I'd go through how you should do it in HBX and how you do it in Google Analytics.

The campaigns I am going to be talking about are when you direct someone to your site via some sort of campaign. In HBX, you can of course create campaigns based on many things, not least by putting them in the tags on particular pages (and also you can create a campaign in the same way that you create a conversion point - more on how to create conversion points later).

First off in HBX you have to work out how you are going to do it. We've decided that we want to do it dynamically (ie not having to set up the campaigns in the user interface). There are many ways of doing this as well. You have two parameters in your HBX tag under campaigns, which we've altered slightly:

• hbx.dcmp = "_P::cp" and
• hbx.hra = "_P::attr"

Why have we done this? Mainly so that when we put our campaign codes on our destination urls, we don't have to worry about duplicating query strings created by other codes.

HBX campaigns work by using query strings on the end of urls. If you have url rewriting systems (which strip out query strings), then you'll need to instruct the HBX tag to pick the campaign code up out of a session cookie (which in turn would have picked it from the url before the rewriting). In theory though, you'll still need to put the query string on the end of a destination url.

The first thing to check is that you've got your destination url properly. If you are using vanity urls and redirects, you'll need to redirect to the full campaign code version or make sure that your redirect keeps the campaign code (or stores it in a session cookie).

The next is to see if it has a query string on the end. Query strings on the ends of urls are used for many things (eg printer friendly pages, etc). A query string on the end of a url starts with a question mark (?). If there isn't a question mark already, you'll need to add one. If there is one already, then each set of attributes is split with an ampersand (& - anyone know why it is called an ampersand?). Then you attach your campaign code on the end as such (this is for a paid search campaign):

http://www.fwi.co.uk/Home/Default.aspx?cp=KNC-AGFW-adwd-cows

How is it broken down:

• URL (http://www.fwi.co.uk/Home/Default.aspx)
  
• Start query string (?)
  
• HBX campaign starts (cp=)
  
• HBX campaign type - this is a paid search campaign (KNC)
• Industry and site - Agriculture and Farmers weekly, as we have many websites in the same market (AGFW)
• Who the campaign is with - Google adwords, Yahoo!, MS adcenter, etc (adwd)
• Campaign name (cows)

For paid search campaigns, Analytics systems will tend to assign all referrals as organic unless you specifically tell them so. You do this in HBX by assigning the start of the campaign code as being KNC or KNL (for lead generating campaigns).

Here we have an newsletter campaign (although this principle works for all email campaigns) and I always like email campaigns to work slightly differently:

http://www.fwi.co.uk/Home/Default.aspx?cp=NLC-AGFWweekly-20080624&attr=topleft

• URL (http://www.fwi.co.uk/Home/Default.aspx)
  
• Start query string (?)
  
• HBX campaign starts (cp=)
  
• HBX campaign type - this is a newsletter campaign (NLC)
• Industry and site - Agriculture and Farmers weekly, as we have many websites in the same market (AGFW)
• The actual newsletter name itself - this one is just called 'weekly'
• The date that it was sent on - this is important if you send the same newsletter frequently (20080624)
  
• Start of a new parameter on the query string (&)
• The secondary attribute in HBX (attr=)
• Any split that we may want from the links in the newsletter (topleft)
  

And that is it. HBX has a number of campaign types, but I always find these are the most common (KNC for paid search, EMC for email and NLC for newsletter). Here are the others that you can use:

• AFC Affiliate
• BAC Banner Ad
• CAC Classified Ad
• DMC Direct Mail
• ILC - Internal Link
• LEC Link Exchange
• OTC Other
• PAC Print Ad
• RAC Rich Media Ad
• SPC Sponsorship

Don't forget that HBX is case sensitive!!!

In Google Analytics it works slightly different. There is a really good explanation by Justin Cutroni which I am going to do a bit of copying and pasting from.

The principle remains the same. You put the campaign on the end of your destination url and Google picks up the campaign code from there. The campaign parameters that you set in HBX are slightly different. In Google every thing is free form, there are no standards to follow. However that shouldn't stop you setting up a naming convention, otherwise you won't know what the hell is going on.

Firstly you don't need to do anything with your Google Adword campaigns. You can just link them in your user interface in adwords. Firstly, give your Google Analytics account admin rights to your adwords account. This is really easy:

1. Login to Google Analytics
2. Click on 'edit' next to the site that you are linking underneath the 'Settings' column
   
3. In the bottom box (Users with access to Profile) click add user (in the top right hand corner of said box)
4. Under 'email address' give the adwords account ID, also fill in the name of the person
5. Access Type should be set to 'Account administrator'
6. Click Finish

Now you just need to link the two accounts together and this is how you do it:

1. Log in to your AdWords account at https://adwords.google.com.
2. Click the Analytics tab.
3. If you already have an Analytics account, click I already have a Google Analytics account.
4. From the Existing Google Analytics Account drop-down list, select the name of the Analytics account you'd like to link to.
   
5. Keep the checkboxes selected on this page
6. Click Link Account.

Now your adwords and your analytics accounts are together, all your campaign details - visits, clicks, etc - will appear in your analytics account.

You don't have to do any tagging for Google Adword campaigns in Google Analytics.

For your other accounts you will have to, so make sure you set them up as Justin says. Put them in the query string as you did with HBX. You can choose how you set them up. Alex Cohen on his Digital Alex site has a great recommendations on how to set these up, so I won't copy him too much. Here are his suggestions on the campaign medium parameter (utm_campaign):

• Affiliate - affiliate
• Comparison Shopping Engine - cse
• Email - email
• Local Search - local
• Offline media that drives online click-through - offline
• Paid Inclusion Bulk - paid-inclusion
• Pay Per Click Search - cpc

That last one is vitally important for your other search campaigns (MSN, Yahoo!, etc) because you want them all grouped together in Google Analytics with Google Adwords.

What you end up with is a url that looks similar to our HBX one, but with different parameters:

www.fwi.co.uk/Home/Default.aspx?utm_medium=email&utm_source=AGFW&utm_campaign=weekly-20080624&utm_content=topleft

Next time we'll go through how to pick up results from these.

Engagement is Conversation

Well I've just come back from the RBI eMarketing conference, where I did a couple of speeches. The second one was on A/B testing, where I did a couple of examples and got people to vote (thanks to all those people that I nicked slides from). However, I thought I stick this post up here about the first one that I did with James Kelway on Engagement building and metrics (before he has a chance). What I'll do is briefly cover his bit on building engagement (and link to his post when he writes it) and concentrate mainly on my bit of measuring.

I've talked about Engagement before on this blog (and so has Kate when she described how people are cats and not dogs), so I won't go into too much depth. I thought what I'd actually do is mention what the metrics that we are using to describe engagement are:

• Non homepage entry visits
• Bounce rate and visit duration
• Conversions and micro/step conversions
• Brand search term strength
• Frequency
• Emotional responses

Thanks to Dave Chaffey actually for this, because I used much of his slides and melted them down into smaller versions.

Ok, so our first measure is all about non-homepage entry visits. I tried to stir it up a bit here, by claiming that "the home page is dead". I only mean metaphorically, but it gives you the right impression. For many of our sites the home page doesn't get more than 50% of the site entrants. For some of them it is 15%. We need to focus on the other areas that visitors are arriving at the site and optimise them for landing pages. Specifically I always think the best way of doing this is to group your content into either content groups (or if you are using Google Analytics into your url structure) and then surveying each group. You can then work from this to work out the areas you need to be concentrating on most.


I won't wither on too much about bounce rates and visit durations, because I've posted about bounce rates and visit duration before. I always think a better way of looking at bounce rate is to look at the visits viewing more than one page. They are the important ones and we want to increase them. And then we can focus on these people and work out how we get them to stay for a longer period of time (ie how many visits last for more than one minute).



So having got them into the site and managed to get them to stay for more than one page, our ideal situation is not them wandering around aimlessly for hours on the site. We have key things that we want them to do on the site - the conversions. How many of them go on to convert? For each of our marketing campaigns how many of them go on to convert? These are the areas that we need to focus on. Having got that information, we then need to go back a step and wonder what we can do with it. This is where our step or micro conversions come into play. We need to look at each of the individual steps along the way to the conversion that the user has to take and work out why they are dropping out at each stage. We need to do this not only as a whole site, but also for each of our individual marketing campaigns.

Now we come onto the most important thing for our Marketers. Online, the user experience is the brand experience. If your users are having a good user experience online, then they'll remember it and they'll come back because they'll remember the brand. How will they get back there? They'll search for you in a search engine. It's vitally important that you measure how much traffic your brand related traffic is generating and measure if it increases or decreases. There may be other factors that will effect this including television and radio advertising.




I haven't ever posted before on frequency. It is something that I quite often put out there to the Business though. We want people to come back more frequently (in Google Analytics they call it Visitor Loyalty) and hence one of our key measures should be how frequently our visitors visit the site (how frequently do we want them to visit?). I always think a good way of looking at this is the volume of visitors that visit more frequently than once in your set time period. We also have a recency metric, whereby we can see how long ago visitors visited the site. Any that visited over a certain time period ago are no longer going to be engaged with your brand and should be treated as new visitors. Up to a certain point though, you want to keep them as 'active' as they'll remember you if you market to them.

Finally, there are the emotional responses. These are what the user thinks. The best way of finding out is either to ask them using email surveys, exit surveys, etc. Or you can go to your expert groups (on facebook, on your blogs, etc) and ask them to let you know. This can't be underestimated. Whilst the above measures will show you what the users do, the emotional response will tell you why they are doing it.

Let me end on the bit that James looked at. These are the methods for building engagement (as measured above):

1. Content – build for focus and deliver on what is promised
2. Calls to action – give clear choices and ensure relevance.
3. Context – build trust and credibility by using visual hierarchy

Remember that measuring engagement is a waste of time if you aren't going to do anything about it. Get that focused content out there, get the context around it (scent!!!) - where people are coming from and where they'll want to go to - and then give the clear calls to action on the page.

And so this finally came down to our point. What is engagement? We couldn't find a definition anywhere - so we made one up.

Engagement is conversation.

It is about your users conversing with the website. Whether that be by reading it, posting comments, looking at another page, buying the product, signing up to the email newsletter. You can't easily converse back mind you...