Upcoming LunaMetrics Seminars
Washington DC, Sep 22-26 Boston, Oct 6-10 Chicago, Oct 20-24 Seattle, Nov 3-7

Google Website Optimizer is Dead. Long live Google Analytics Content Experiments

Until a few days ago we’ve had to keep this one under wraps, but with the official announcement on the Google Analytics Blog Friday we can finally talk about the all new Google Analytics Content Experiments. This is the all new, tied directly into your analytics, testing software to replace Google Website Optimizer. Google Website Optimizer will slowly be decomissioned over this year, and replaced fully by these new Content Experiments. So if your’e starting any A/B testing anytime soon, time to do it in here rather than in GWO.

Let’s do a quick run through on how to use and set up these new tests. It’s really simple, and A/B testing is something pretty much everyone is capable of doing, and should give a try.

The first screen below here is where you start. It’s under Content in the left navigation, and from there you’ll see “Experiments”. This brings you to the first page. The first step is to simply enter in the URL of the page you’d like to A/B test. Enter the original URL, so if it’s your index.php home page then put in www.mydomain.com/index.php for instance (Where obviously you’re replacing mydomain with your actual website domain.)

Google Analytics Content Experiments Screen 1

I’d like to test my “What are Content Experiments?” test page. I have a picture of a cute puppy, and a green button, but I don’t know how effective that is. How many more people could I get to click the button if it were say red, or orange. Or if I had a picture of a cat instead of a dog, or I dunno…. A manatee.

Google Analytics Content Experiments Screen 2

So I enter in my url, and a little screenshot shows up. Interestingly it doesn’t load the image of the dog for the preview. Still, it has it, it’s the right page, so I click “start experimenting”

Google Analytics Content Experiments Screen 3

The next page I get to is to start choosing my variations. I name my experiment “Button Color Experiment” just for my own way of identifying it. No user sees that. Then I check the original page.  Time to add a B to the A/B. I have created a second page, with a kitten image, and a red button. I add that URL to the variation 1 area and it loads up the preview. I could just go on from here, but I’m going to test a third option as well. You can test up to 6 variations at a time.

Google Analytics Content Experiments Screen 4

Now I have my three variations I want to test loaded in. green button with my original page, red button on my first variation, and an orange button on my second variation. Time to move to the next step.

Google Analytics Content Experiments Screen 5

Here’s where some people might get hung up. Right now you need to use a goal already in Google Analytics for your site. If you want to read up on goals we have plenty of blog articles on them. Right now you can do a URL Goal, or an Event. If you don’t currently have a goal set up for the page you want them to go to for a URL Goal for instance, you can click “add new goal” and go set it up right in the flow.

Google Analytics Content Experiments Screen 6

Clicking add a new goal, takes you right into your admin goal area. If you’re not an admin on your account remember you won’t be able to do this at all. So here you can create a new goal. I’m going to put it in set 2 to keep it separate from my other Goal Page.
Google Analytics Content Experiments Screen 7

So I name it “Button Goal” and make it a URL Destination goal, and then in the Goal Details I put in the URL of the goal page that I want them to reach upon clicking that button. I gave it a goal value of 1. Why not. If you click save and return here, it’s pretty slick… It’ll take you back into your Experiment, so you don’t have to navigate back there or click the back button awkwardly.

Google Analytics Content Experiments Screen 8

Once we’re back in the Experiment we can then select that new goal from the drop down list. We can also choose how many visitors to include. IF we get alot of visitors we could make that a much smaller number, and keep the testing to just a small segment of our user base. This is useful when you have enough visitors that not only you can get a good sample from a small percentage, but it keeps a massive amount of your user base from seeing these changes you’re testing, particularly when you don’t KNOW whether they’re good or bad. I’m leaving mine at 100% because I don’t really have any traffic to my lonely test page… You can also enter in notes, and then when you’re ready, click next.

Google Analytics Content Experiments Screen 9

Now another stumbling block. You have to add some code to the first page of the test, the main page people go to. If you don’t have access to your website this is a problem, but you can email the instructions to people. However if you can modify it all you need to do is open up the HTML for the page you’re testing, and add this specific code right after the <head> tag.

Google Analytics Content Experiments Screen 10

It’d look something like this…

Google Analytics Content Experiments Screen 11

Once the code is in the page and on the server, click next to validate the pages. If the code is installed right and it can find all your pages you’ll get lots of green checkmarks and an “Excellent!” button. Click that.

Google Analytics Content Experiments Screen 12

You’ll then be able to preview your experiment pages just to confirm it’s working correctly.

Google Analytics Content Experiments Screen 13

Google Analytics Content Experiments Screen 14

Google Analytics Content Experiments Screen 15

And from there you can confirm to run the experiment and you’re off to the races!

Google Analytics Content Experiments Screen 16

You’ll then be able to click on the experiment and view the page. Of course once you start the test there won’t be any data in it. It can take a bit for the data to populate, and it can take even longer for Google Analytics Content Experiments to determine a true winner. Days, weeks, even months depending on how much traffic you have.

Google Analytics Content Experiments Screen 17

Also a note, you’ll be able to see the test is running by looking at the URL’s of those pages. There will be a new utm_expid parameter attached to the pages.

Google Analytics Content Experiments Screen 18

Once data starts to accumulate though you’ll start seeing the differences. You’ll be able to compare them to your analytics data, see them over time, and see the different conversion rates. And when one is determined statistically to be better than the rest, Google will let you know, and tell you which one is the winner!

Google Analytics Content Experiments 20

And then when a winner has been determined, it’s time to start a new A/B test with your winner and all new variations!

Conclusions

On the whole I”m pretty excited to have Content Experiments tied into Google Analytics. There are a number of benefits to the new system. There’s only one code snippet you need to include on the page rather than multiple pages of code. It really simplifies that aspect when you need to add new testing. You can also now use advanced segments to segment your results too.

There’s some improved statistical models too. Test results don’t even show up for 2 weeks or more, and all tests expire after 3 months, assuming you can’t get a statistically significant winner. If you have alot of traffic that’ll undoubtedly be true, but it’ll make it harder to do longer tests on lower traffic sites.

All in all though I think it’s great. If I had one wish it’d be to add Multivariate testing as well as just A/B testing. You can do MVT and pretend through an A/B test but it’s much more awkward.

So what are you waiting for? Go start testing something!

Sayf Sharif

About Sayf Sharif

Sayf Sharif is a Web Analyst, and expert in Usability and UX, who has worked with businesses large and small to maximize their online presence since the beginning of the Web, winning numerous awards along the way. Sayf has studied human tool use from the stone age (he went to graduate school for Archaeology) to the information age (he started programing on his father’s TRS-80), and is always interested in what goals people wish to accomplish using their tools, and how successful that experience was.

http://www.lunametrics.com/blog/2012/06/04/google-website-optimizer-dead-long-live-google-analytics-content-experiments/

70 Responses to “Google Website Optimizer is Dead. Long live Google Analytics Content Experiments”

hyderali says:

But where do I’ve to add that code? Which page? The original, 1st variation page or 2nd variation page?

Thanks for the post.

In the ‘old’ GWO tool an A/B test would break your referrals for 50% of all your visits. That was caused by the javascript redirect on the A page. That was the main reason to use the multivariate option to do A/B tests. Is that problem fixed?

Sayf Sharif Sayf Sharif says:

hyderali, sorry if that wasn’t clear. You just copy and paste that one chunk of code into the original page after the tag. Just there. Nowhere else. You don’t need to put it on the variation pages, or your goal page. Just that initial page you are testing.

So if you’re testing your home page, you just put it on ONLY Your home page, but NOT on the variation page you are comparing, or the goal page that you want people to reach.

Or if you’re testing an internal page, like a shopping cart, then you just put it on ONLY that shopping cart page, and not the variation A or variation B or C pages that you’re comparing it with. Those you don’t put any code onto, you simply identify them in the content experiment UI.

This can be a problem for people with sites that are built in ways where the head area of the document gets delivered the same on every page, but generally on most sites you can put in a little code to only deliver it on certain sites, particularly if you’re in a PHP site, or in wordpress.

Sayf Sharif Sayf Sharif says:

Andre, EXCELLENT question. I loathed that aspect of the old GWO, and could never truly figure out what caused that. It wasn’t consistent in that on some sites the referrals would carry over, and on others they wouldn’t, and you had to jump through hoops to make it work.

On my test server, the referral information DID pass correctly to the redirected page, and to the goal page. As I said however, it wasn’t consistent in my experience across servers, so it’s quite possible that same issue remains, but I have yet been able to reproduce it.

If you run a test on your server and find out that the referral issue remains, please come back here and post another comment and let us know! My hope is that this problem has been eliminated, but until we go a few months with nobody complaining about it, I’ll hold my breath (with hope).

Neil says:

Hi, should this be available to everyone already? I don’t have the experiments option showing up under ‘Content’ (‘In-Page Analytics’ is the last option I have). I was wondering if it’s being rolled out slowly or whether I have to enable something? Thanks.

Neil says:

Sorry, ignore the above – I overlooked it first time round but it does actually mention a gradual roll-out on Google’s blog.

Bronwyn says:

I’m excited about this.

I’m assuming these experiments could also be used in conjuction with Adwords to test landing pages, is that correct?

Thanks

Sayf Sharif Sayf Sharif says:

Neil, they do mention a roll out, but across several accounts I have yet to see one that doesn’t have it. I’m not sure of the precise roll out timing, so it’s possible there are people that won’t see it yet, but my feeling is that Google is moving pretty quickly on this one to get it live, as they want people in there immediately, as Google Website Optimizer goes away in only a few months. I’ll be surprised if there are people who don’t have access to it at the end of this week.

Sayf Sharif Sayf Sharif says:

Bronwyn, From their blog post: “With full integration in Google Analytics, we’ll be able to grow and evolve website experimentation tools within our broader measurement platform. Initially, you’ll be able to utilize important features like optimized goal conversions, easier tagging, and advanced segmentation in reports. We’re also working hard to release page metrics, additional goal conversion options and experiment suggestions.”

Which to me reads that they’ll be adding plenty of stuff in the future. Right now looking in custom reports and advanced segments I don’t see any related metrics or dimensions to create additional new ones based on current content experiments. It would be a great feature to be able to segment your currently running experiments while looking through say your Advertising metrics in GA. Unless I’m not seeing it, they haven’t implemented it yet.

If someone DOES see how to segment based on currently running experiments, please post and let us know!

Tim says:

Is there a mechanism to ensure that a user gets the same version of the page through their entire visit? For example, people can see, leave, and return to a Cart page several times during a site visit. If that is the page I am testing, is it possible their view will switch from original_page to test_page and back?

Sayf Sharif Sayf Sharif says:

Tim, once a user hits the test page, that’s the version that user will see. The variation they’re delivered gets placed in a cookie. The only way they would see a new version of the page is if they deleted their cookies entirely.

Im not really going to miss GWO :) I have been looking forward to this release and will be doing some experiments very soon. However i still see a lot of advantages of using a more sofisticated tool lige i.e. Visual Website Optimizer.

Marketeering says:

Is anybody not seeing the Experiments option in the navigation bar of their Analytics account? I’ve had a GWO account for years but it seems as though they haven’t activated Experiments on my Analytics account yet.

Sayf Sharif Sayf Sharif says:

Sebastian, I’ll be honest in that I agree with you. While GWO was very useful, too often we would run into some sort of issue, whether it would be destroying the referral information for 50% of the visitors to the site, or something along those lines.

Content Experiments definitely aren’t the most sophisticated solution out there, and if you need to do more than A-B and basic variation testing, at least right now, there are other ways to go. However for just constantly throwing up variations and running tests tied directly into your analytics, I think it can’t be beat for ease of use, and cost. Though as I said, I’d love it if they incorporated in some more MVT aspects rather than just A/B.

Sayf Sharif Sayf Sharif says:

Marketeering, sounds like you’re one of the unlucky ones that still isn’t seeing it. keep checking under Content in the left navigation. I’m surprised that it’s not fully rolled out now, a week later, but if you don’t see it yet, you’ll probably see it soon. Another thing you can try is to log out and log back in. Sometimes the new features won’t show up without logging out and back in.

Pez says:

Hi. It still isn’t showing up in my Analytics either. Waiting patiently.

I’m just designing a slightly edited version of my homepage ready for when ‘Experiments’ appears… But I have a quick question;

Will there be any Google duplication issues with having more than one page on the server with similar content on?

Should I use ‘noindex,nofollow’ on the variation pages or would that mean the test wouldn’t work?

Thank you.

derek says:

The article has the line “Test results don’t even show up for 2 weeks or more.”

Does this mean there is no data for 2-weeks or more?

I’ve setup some experiments that are 24 hours old. GA has validated everything with green check marks, but is still reporting 0 visits.

I definitely do have visits and I’m not filtering or excluding results via a custom segment. My pages also have the utm url parameters appended.

I’m wondering if I have to wait to see visits and data or whether there is something wrong with my tests.

Sayf Sharif Sayf Sharif says:

Pez, Supposedly no, it’s not going to be considered a duplicate page. Be sure not to list your variation pages in your sitemaps however, as they could be picked up by search engines other than Google.

Also just to be on the safe side I would use a canonical tag…

Sayf Sharif Sayf Sharif says:

Derek, They say 2 weeks, but I also saw some numbers earlier. But according to them you won’t see the results of the testing at the same time your other data comes in. Essentially its still separate information than your standard pageviews, so it won’t load up at the same time.

If GA says that the experiment is running, then this is what we’re talking about with the 2 weeks. You won’t see hits in there for a bit, even if the pages are getting hit. You need to wait to see the data up to 2 weeks.

derek says:

Thank you, Sayf.

I hope it’s on their roadmap to eliminate the delay. I see no reason for a delay when there wasn’t one in GWO.

Sayf Sharif Sayf Sharif says:

They list the delay along with talking about their new algorithm, so I wonder if the algorithm itself depends on a certain amount, or timespan of data.

Robbin says:

Sayf, I want to weigh in on the duplicate content issue. When I was at the GWO summit in Mountain View in 2007 (yes, a million internet years ago), Avinash Kaushik talked about how hard he had worked with Matt Cutts to ensure that tests were not seen as duplicate content. So I know that there is a strong desire to ensure that Google Search plays nicely. That said, it is new *and* Google is still not the only search engine. At least, last time I checked.

Dennis van der Heijden says:

Hi Sayf, if you miss multivariate testing or live results we launched Convert Experiments 2 days ago. GA integration and revenue tracking, hope you like it. http://www.convert.com

Pez says:

Thanks Sayf.

Good point about the other search engines Robbin. I take it that’s why you advise a canonical, Sayf?

Experiments STILL isn’t showing in my Analytics! :( I’m in the UK, could that be a reason do you think?

Sayf Sharif Sayf Sharif says:

Robbin, I’m sure it’s a concern on their parts and they do all they can, but I just think of the warning WordPress gives you when you set the privacy settings within it’s admin.

“Note: Neither of these options blocks access to your site — it is up to search engines to honor your request.”

It’s more likely that Google itself will honor whatever test parameters are set up on a variation page, and I’d like to think that the two groups would communicate, however Google Analytics and the developers who write Content Experiments are completely separate from the search engine folks. How much they communicate, or claim to communicate, I don’t know. Even if they’re completely on the same page however, the other search engines as I, and you, note, are definitely not. Given that you can go directly to the variation page in your browser without being a part of the content experiment means it WILL get spidered at some point.

Using rel=canonical tags on your page and variations will in no way harm your SEO, and can only help protect you even in Google, but definitely in other search engines.

Sayf Sharif Sayf Sharif says:

Pez,

I’m not sure if there is any location type roll out. Be sure to log completely out of analytics and google, and try logging back in. If it’s still not there, I can only recommend checking daily and being sure to log out and back in every day. If you don’t have it yet, it’ll surely be soon.

Heath Copps says:

I still don’t have Experiments. Must be last on the list. Of course, I don’t think I can get to the Head on individual pages of the site, so it might not matter right now.

In the documentation, Google definitely recommends using rel=canonical: http://support.google.com/analytics/bin/answer.py?hl=en&answer=2613318&topic=1745207&ctx=topic

Sayf Sharif Sayf Sharif says:

That support link isn’t working for me, but I’m glad to hear that they recommend using the rel=canonical.

As far as the rollout, it definitely is taking longer than I expected, though supposedly over half of users have it at this point. Keep checking!

Ander Jáuregui says:

Excelent review!!

Question,
What if my original URL have parameters, and these parameters depends of the previows steps… analytics will know that parameters and execute the test with al the possible convinatiods depending the previows steps?
Example of URL
portal-blabla.com/blabla/HotelResult.aspx?codiconc=01&fechaini=30/06/2012&fechafin=05/07/2012&numerohabitaciones=1&edadpersona0_0=30&edadpersona0_1=30&adultohab0=2&ninohab0=0&bebehab0=0&numeropersonas0=2&css=

So, analytics will know that my URL is portal-blabla.com/blabla/HotelResult.aspx and automaticaly will make the test with al the convinations?

Thx!

Sampatirao says:

Hi Sayf Sharif,

Thanks for the demo and its really helped me to give demo to my business.

I am having question on the variation page to display…How the variations pages will appear to the visitors on random basis? Can you limit these variation like 20%, 45% and 35% to appear to the users.

And also can we add multiple goals to the experiment?

There is an option on number visits to experiment. On what basis this will apply, say I selected 50% of my visits. On what basis this 50% will be considered?

Regarsd
Sam

Sayf Sharif Sayf Sharif says:

Ander, if I understand you correctly you should be fine. GA will pass your own parameters onto the variation pages so anything you pass to the original page will also pass to the variation page as well.

Sayf Sharif Sayf Sharif says:

Sampatirao, yes you can limit how many people see the test. You can have it include 100% of your visitors, 75%, 50%, 25%, 10%, 5%, or 1%. So for instance if you chose 5%, 95% of your visitors would see your normal page, and only 5% would be part of the test batch. As far as how often specific variations appear, we don’t have control over that, it’s part of the statistical algorithm that GA is using.

You cannot add multiple goals to the experiment, you can choose only a single goal.

Tamra Hamblin says:

I’m really frustrated. Our company’s GA accounts still don’t have this option. I really want to get started with the testing but without a tracking mechanism there is no sense in it.
Wish they could roll faster so even the small sites can have the information right away too. Obviously patience is not one of my virtues… :)

Sayf Sharif Sayf Sharif says:

Tamra, I feel your pain. They’re really cool, and with GWO being taken away soon, there’s almost no reason to start a test there, but without GA Content Experiments available you’re in a bind.

Be sure to try completely logging out of your google accounts, closing your browser, and then re logging into Analytics. It’s been known to happen that new features won’t show up in your GA until you log out and back in.

But barring that, you just are most likely an unlucky one, and when you’re impatient like you (and I), it’s frustrating.

Given the timeline for depreciation of GWO I’d be shocked if you don’t have them by the end of the month, but honestly the rollout is taking much longer than I anticipated originally, so take that with a grain of salt.

What you CAN do if you absolutely can’t wait for Content Experiments is to fudge your own A/B testing. Unfortunately for most people this means javascript coding, setting cookies of your own on the user browser, custom variables in GA, and then tracking in that regard.

If you’re interested in how to do A/B testing without using GA Content Experiments let me know, however probably by the time you implement something like that, you’ll see the Experiments option in your left hand navigation of GA.

It’d be Murphy’s Law at it’s finest…

Sampatirao says:

Thanks for your reply. Another query, is there any that we can add multiple experiment code to one page? because I want to test multiple Goals. As you said we can test only one Goal at time.

Regards
Sam

Sayf Sharif Sayf Sharif says:

As far as I understand how the code works, no you can’t add multiple experiments to a single page, other than by creating a few different variations.

However you can probably test multiple goals by creating an advanced segment. Create an advanced segment with the page dimension value containing your variation page for the variation you want to see. With that segment active you can see how visits that included that variation page affected any other data, including any and all your goals.

Would that help?

Tim says:

Hello Sayf,

I know we can not run multiple concurrent tests on the same page. Can we have two (or more) separate pages being tested concurrently?
Thank you

Sayf Sharif Sayf Sharif says:

You can run up to I believe 8 or 10 experiments concurrently, as well as of course up to 5 variations per experiment.

The question is should you?

Keep in mind the more variations and experiments you run, the harder it is to attribute success to a specific variation. I’m not saying DON’T do it, i’m just saying consider whether you can test via simply variations, or whether you need an actual secondary test.

An example of a good secondary test would be a button color on a shopping cart page (after a primary test of completely different product pages).

A complicated one would be a completely different category landing page, with a secondary test of a completely different product page, with a tertiary test of a completely different shopping cart page.

Now add 4 variations to each of those, and suddeently you have users seeing Category Page B, Product Page C, Shopping Cart Page A, and then they convert…But did one page of that work better? WAs Page B better, but Page C worse? You’d need a TON of traffic to ferret that out with a statistical signficance inside the 3 month window.

So sure… Go do more than one experiment, but just be careful with it.

And I reserve the right to completely change my mind and say “oh god no don’t do multiple experiments on the same site, it’s all broken!” because GA Content Experiments is still relatively new.

Rosenstand says:

I have on question: What about duplicate content issues? Build 2 or 3 versions of the main page with the same (or almost) content and put it online. Google will index the 2-3-4 versions and BAM! Or what?

Sayf Sharif Sayf Sharif says:

Rosenstand, Google claims that they’ll take that into consideration but they also recommend you use a rel=canonical tag on all your variation pages pointing to the main page. That way if they do get spidered by Google, or other search engines, your SEO won’t get diluted.

Rosenstand says:

Thanks Sayf. I appreciate the answer. I might even put “noindex” on the test pages just to be absolutely sure.

Sayf Sharif Sayf Sharif says:

I haven’t read anything about whether that would be bad to do, but I can’t see it’d be a problem. Google does warn about people using content experiments for seo “cloaking” but I don’t think this would trigger any issues. rel=canonical is what they recommend, and is probably enough, but “better safe than sorry” is generally pretty good advice, so I’d say go for it with the noindex as well.

Rebecca says:

Hi Sayf, I’ve used GWO for clients that don’t use Google Analytics and their conversion page is on a secure page. We’ve therefore used doGoal onClick tracking to track clicks on the ‘goal’ button and cross referenced the results against other reporting mechanisms (Omniture mainly). Now with GA Experiments, I understand that this won’t be possible nor will onClick tracking for more than one button…
Any ideas?

Ryan says:

Any chance you know wordpress plugins to add code just after the on a page at wordpress?

It will be awesome

Sayf Sharif Sayf Sharif says:

Ryan, why must you give me these sorts of ideas… Stay tuned to our blog. You may have inspired me.

Sayf Sharif Sayf Sharif says:

Rebecca,

You can have event goals, if that’s what you’re referring to. I’m a little hazy based on your description but here’s what I’d recommend.

You can set Goals in GA based on an Event, however you can specify the goal details. You can have it match any combination of Categories, Actions, Labels or Values, any of the inputs into a TrackEvent. So if you structured multiple buttons with different events, but combined their label for instance you could track them all as the same goal.

So for instance let’s say that you have four different conversion pages, with four different conversion buttons. You could track events on each of them that were say

Category: ConversionButton
Action: All different for each button
Label: All different for each button
Value: All different for each button

Then in the goal you just match the category “ConversionButton” and any event matching that category triggers the Goal completion.

If you’re already organizing things a certain way with categories, you could do it with actions or values, or even do a matching expression. Or you could set the goal value as a number… Say you want only events with a goal value of 443 to trigger a goal. You could do that.

That way you could set up a Content Experiment that could theoretically get tracked on an infinite number of buttons on the site because it’s just a matter of how you label the even tracked on the click.

Does that apply to your situation?

Mike says:

Thanks for the post.

Do you know if you can run one AB test across multiple pages? Specifically, I have about 20 different pages (unique content for SEO) that have the same call to action block. I would like to test a B version of the call to action block across all 20 pages (I need the traffic across all 20 pages to have a timely test).

Paul says:

I am a Westerner running a Chinese language website and we are trying to track experiment results for a process that runs from our main domain and ends up on our sub-domain. According to our Google Analytics page, the experiment code is validated and working. But, there is no data in our experiment results. Can you help?

Sayf Sharif Sayf Sharif says:

Mike, sorry for the delayed reply i’ve been on vacation.

You cannot by default run a single Google Analytics Content Experiment AB test across that many multiple pages. Particularly if the users will be hitting those 20 pages as initial landing pages, the GACE way of doing things goes out the window and your’e left to your own devices.

Now, if all the visitors were being funneled from a pervious page you could run a Content Experiment by having essentially two variant pages that are identical as landing pages, and then setting your own cookie on each of those landing pages, and delivering different call to actions on the following pages based on which initial landing page they hit…then measuring their event goal the same.

However given you mention their unique content and SEO I assume people are entering on those specific pages. Because of that you’d have to use a different AB testing software, or code it yourself. You could create a cookie for users randomly when they come to your site, and based on that cookie set a custom variable to denote which type of call to action they’re viewing. Then based on the cookie as well put the correct call to action there, as well as the event tracking on that call to action. Then you can compare the conversions based on the custom variable you set on the users within GA. It wouldn’t have the same statisitical algorithims applied, so you’d have to do your own math on that, but otherwise you’d be able to compare one versus the other at least.

Sayf Sharif Sayf Sharif says:

Paul, how long as it been running? It can take up to two weeks for data to start showing up in the content experiment area. If you set it up, and GA says it’s running and working and validated, but it’s within 2 weeks…. Give it time.

if you’re running it longer than 3-4 weeks and still see no data, I’d say you’re experiencing an early stage bug for Content Experiments, and should abandon that experiment, and start over.

Sudhir says:

Hi Sayf,

I have created two experiments – one with URL destination and other with Events goals. Surprisingly I dont see any tracking in first experiment with URL destination. I can see page views from content for the respective pages.

There is no issue with Event goal experiment.

Am wondering what might be the issue – is it any Google known bug?

Thanks
Sudhir

Sayf Sharif Sayf Sharif says:

We’re finding more and more that there are bugs with Content Experiments. My recommendation would be at this point to run one experiment at a time, make sure it’s working and your data is good, and then add in a second and check again after a couple weeks. If things start getting mesed up removed that second test.

As far as general issues with tracking URL Destinations there are none that would be indicated by this, but it could easily be a bug. I’d have to know more specifics about the page structure, the order users show up at the sight and where they hit the experiment pages, etc.

Or it could be working and just taking awhile to have data show up if it’s been less than 2 weeks. If it’s been less than 2-3 weeks let the data have time to show up. If it’s been more than that, i’d stop the test, and start a new one that’s basically the same and see what happens.

Paul says:

Hi Sayf, thank you for your reply. Wow. 2 weeks? That’s a long time … I’d gotten accustomed to the more “instant” type of results in Optimizer. I guess I’ll just have to wait … but seriously … two weeks feels like an eternity when you work for an organization that wants answers on a daily basis! Thanks again for your reply.

Sayf Sharif Sayf Sharif says:

I agree. I wish you could immediately see the results coming in, and I’m not exactly sure WHY they’re withholding the test data for that initial period. Something about the algorithm, but I don’t know.

Istvan says:

Hi. Great post. I am using the Google optimizer for a long and I was wondering if there are any chance to transfer existing running experiments into GA Experiment section? So I would keep all my result in the new section of GA. Thanks

Sayf Sharif Sayf Sharif says:

Istvan, unfortunately it appears that you won’t be able to transfer any existing and running experiments into Content Experiments. The best you can do is to recreate them in GACE and start them anew.

Eric Clark says:

Are you telling me that all my websiteoptimizer experiments are just gone now?!?!?!?!

Sayf Sharif Sayf Sharif says:

Uh. Unfortunately yeah. Per the official announcement on June 1st:

“The last day you’ll be able to access Google Website Optimizer, and any reports for current or past experiments, will be August 1, 2012″

http://analytics.blogspot.com/2012/06/helping-to-create-better-websites.html

I don’t know why they wouldn’t just have left the old reports and experiments up, even if they stopped tracking additional information, but anything there is now gone. You could try contacting Google about it, but you are probably out of luck.

Anyone else think that the 2 months from “hey we’re discontinuing this” to completely shutting off access to all data is pretty short? Particularly for content experiments were people would often leave things running for months?

Dave says:

Do I add the noindex tag to my variation pages to avoid duplicate content? And if my goal is a landing page on another site, is it true I do not have to add any code to that page? Thanks.

Sayf Sharif Sayf Sharif says:

No don’t add the noindex tag, you can add a rel=canonical tag

http://support.google.com/webmasters/bin/answer.py?hl=en&answer=139394

That will let google know that all of the variation pages are the same page.

If your goal is a landing page on another site, and you don’t have cross domain tracking with your tracking code on that other site, you will have to add some other code to your page(s). What you’ll want to do is have Event Tracking on your external links to this other landing page, and then set a goal for the event that someone has clicked on that external link. Then you can use that goal for your content experiment.

Sam says:

Hi Sayf,
sorry for bothering.
When I set up the experiment,I enter the URL but the snapshot doesn’t show (google says “We could not find the URL you entered”)

But it shows perfectly in the browser.
could you help me to figer out?
if there is any additional imfromation you need ,please feel free to let me know, thanks

Sayf Sharif Sayf Sharif says:

There are a number of reasons that could happen. I’ve seen it before for instance where you can go to a site in a browser, but the Google spider cannot (like if you have a no robots tag or something like that on the page or it’s otherwise redirected or improperly tagged).

However you can always just force the experiment on past that point. I’ve seen those errors before, and when you force the experiment it works. If you can’t figure it out, try running the experiment, and see if data gets collected. Most likely you’re ok, and it’s just an issue with Google being able to see those pages at the time.

satya says:

when i added experiment code on original page ,exactly after the opening head tag.That original page not after execution of project. And gives an error for that page-javax.servlet.ServletException: Error Parsing /pages/login.xhtml: Error Traced[line: 21] The content of elements must consist of well-formed character data or markup.
javax.faces.webapp.FacesServlet.service(FacesServlet.java:606)

so pleese give me the solution.

Sayf Sharif Sayf Sharif says:

Sorry Satya, that goes beyond the level of support I can offer in these comments.

Mazen says:

Hi Sayf;
I added an CE on my google analytic account then I disabled it but it’s still change the content of my page… please how can I remove it completely… I removed the script from the page but even it’s still change the content per visit.
thanks

Sayf Sharif Sayf Sharif says:

Mazen,

if you’ve stopped the experiment in GA it should no longer deliver that alternate page. If you’ve removed the actual script from the page and are still seeing your alternate, I would assume you’re seeing a cached version of that page on your browser. Try clearing your browser cache.

Paola says:

Hello Sayf!
I’m trying to understand if it is possibile to analyze individual users (I can see only general graphics and numbers). I wanna run a test that I need to understand the individual behavior because I’ll send the URL variations to small groups of users.
Thanks in advance
:-)

Sayf Sharif Sayf Sharif says:

Paola, not directly through content experiments, but you could assign user-id’s from your own system into another custom variable, and then use those custom variables to parse out specific users and data.

You could also try using Visitors Flow under the Audience section.

Adriano says:

Hi, Sayf
You mention :
Test results don’t even show up for 2 weeks or more,
Does this mean that I will see no conversions on the experiments ???
Cause I see conversions on the goals ( when look at the goal itself ) , but I don’t see it on the experiment ( that using the goal )
I read in te foogle forum that a lot of people are having the same issue

Sayf Sharif Sayf Sharif says:

That was when it was first implemented. Now they show you the results much sooner. You should see conversions on experiments within a day or two generally. Of course that depends on your traffic and whether people actually convert.

The algorithm that GA uses will cause the data you see in the conversions to be different than what you see in the main account for conversions. I think what you are seeing might be caused by the multi armed bandit model they use.