Rich-text Reply

Homepage redirect with url targeting to run experiments on same page simultaneously

Ed 12-16-14
Accepted Solution

Homepage redirect with url targeting to run experiments on same page simultaneously

Hi I have a question around page redirects and url targeting. I have a landing page that is currently receiving traffic from various promotions in which I am running an experiment and funneling all traffic to a winning variation. I want to create a separate experiement to a different variation on that page that works only with traffic from our homepage via a redirect. I do not want to mix the two traffic sources. Could you help me understand if this is the correct way to do it:

 

1. Duplicate the original landing page as a different experiment

2. Create a variation on the homepage experiment with a redirect to this duplicate landing page

3. On the duplicated landing page, setup audience targeting with url targeting as simple match (from www.homepage.com)

4. Set up goal targeting on the homepage experiment, including the advanced options to target both pages as instructed here

 

As I see it this should separate the two different experiements by traffic source and mean that I can run 2 different experiments on the same page but with completely different traffic sources.

 

Please let me know if this sounds correct. thanks so much!

Ed
Level 2

Brian_Abad 12-16-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

Thanks for running through your idea with us!

From my understanding you'd like to run 2 separate experiments and ensure that your incoming traffic is entering into ONE of the 2 and not both - is this correct?

If what I mentioned above is what you're looking to do, then the Audience visitor conditions you've set within one experiment will need to conflict with the Audience visitor condition you've set within the 2nd experiment. For example, one experiment will target visitors *coming from* Direct traffic, while the other experiment will *exclude visitors* coming from Direct traffic. This ensures visitors will not get bucketed into both experiments.

From there, if the 2 experiments are targeted to the same page, then the URL Targeting across both experiments will be the same but you will still effectively only bucket visitors into one experiment.

Is there anything I've missed within my explanation above? If you'd like to elaborate further on the type of experiment you're looking to set-up I'd be happy to lend another hand! I wasn't quite sure what you wanted to accomplish with the redirect experiment and the traffic source you're targeting and how that ties in to the experiment set-up.

Thanks in advance for the additional information! Looking forward to hearing back and seeing how we can be of better help.
---
Brian Abad
Manager, Technical Support
Customer Success
Optimizely
Ed 12-17-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

Hi Brian, thanks for the reply -

 



From my understanding you'd like to run 2 separate experiments and ensure that your incoming traffic is entering into ONE of the 2 and not both - is this correct?
Correct.

If what I mentioned above is what you're looking to do, then the Audience visitor conditions you've set within one experiment will need to conflict with the Audience visitor condition you've set within the 2nd experiment. For example, one experiment will target visitors *coming from* Direct traffic, while the other experiment will *exclude visitors* coming from Direct traffic. This ensures visitors will not get bucketed into both experiments.
Ok but how do I do this when the traffic is just a redirect from the homepage? Is that what I said - a url simple match targeting on the homepage url?

From there, if the 2 experiments are targeted to the same page, then the URL Targeting across both experiments will be the same but you will still effectively only bucket visitors into one experiment.

Is there anything I've missed within my explanation above? If you'd like to elaborate further on the type of experiment you're looking to set-up I'd be happy to lend another hand! I wasn't quite sure what you wanted to accomplish with the redirect experiment and the traffic source you're targeting and how that ties in to the experiment set-up.
We are a marketplace so my promotions are only pushing supply side traffic to one landing page. The homepage will obviously cover both and everything else, therefore my messaging needs to be a completely different set of tests. In addition, we are using optimizely to maintain the improvements from prior experiments i.e. we are not making hard coding changes to the original page based on results. Therefore I am effectively using Optimizely as a landing page builder to quickly create completely new versions of pages that target different users and even different products we are offering. Hence the need to segment and isolate traffic by source.

Thanks in advance for the additional information! Looking forward to hearing back and seeing how we can be of better help.

 

Ed
Level 2
Brian_Abad 12-17-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

Thanks for the clear responses!

Just to get an even better understanding - you are running experiments to render changes by utilizing 100% traffic allocation to winning variations. Now, you are running a redirect experiment that is targeted to this same landing page where these other running experiments are currently targeted to - is this correct?

I'm not quite sure where the redirect comes into place and how it affects your traffic source in relation to the other experiments that are already running on the page. Are you targeting visitors who have been redirected from a separate experiment? Is an Optimizely experiment triggering the redirect?

Thanks for clearing this up for me! This part is a bit confusing but I'm happy to help upon receiving a better understanding.
---
Brian Abad
Manager, Technical Support
Customer Success
Optimizely
Ed 12-18-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

Yes the first part is correct. And yes exactly, I am targeting visitors who have been redirected from another experiment. So- an experiment on my homepage has 2 variations, a winning variation from a past experiment and now a new variation which we will use to redirect say, 20% of homepage traffic, to a landing page which has 2 experiments running on it. So I want one experiment on this landing page to pick up the 20% of traffic from the homepage and another experiment to pick up traffic from some email and paid campaigns. Hope that clears it up and you are able to provide an answer!
Ed
Level 2
Amanda 12-18-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

[ Edited ]

Hey there! I've put together a quick drawing of how I understand your scenario. Can you confirm if it looks right? Also, can you let us know what you're hoping to accomplish strategically from these tests on the landing page? Whats your hypothesis?

 

Screen Shot 2014-12-18 at 3.30.20 PM.png

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Here's a couple options that should do the trick for you! The key is finding a way to differentiate the traffic hitting your landing page from the homepage vs. hitting the landing page from a paid/email campaign. 

 

This would probably be the easiest way: 

  • In your homepage experiment, can you set your redirect URL to have a specific query string parameter? For example landingpage.com?source=homepage. Then, in your landing page experiments, you can set one experiment's audience to *only* include traffic with this query string parameter present and in the other experiment you can set the audience to exclude traffic with this query string parameter. The source=homepage parameter would never be present in the URL unless the visitor has been bucketed into the redirect from the homepage. 

Other options: 

  • Identify what traffic looks like coming from the paid and email campaigns and hitting the landing page. Are there specific query string parameters appeded that are unique to those campaigns? If so, then you can use these to set up an audience for one experiment (and exlcude them from the other). 
  • You could also do some cookie de-coding to check whether the cookie from the homepage experiment is present. This cookie could be used to create audiences as described above. 

***One thing I would note -- your traffic to the version of the landing page coming from the redirect is going to be pretty low since you are only sending 20% of the homepage traffic to it, so it could take awhile to reach significance (unless 20% is still a very significant number)

Optimizely
Ed 12-19-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

Hi Amanda thank you for such a thorough and thought out reply, really appreciate it! You are absolutely correct with your diagram. I think the query parameter can work, sure. So can you advise if this is the right way to set it up (I think I'll use medium=homepage)

 

Screen Shot 2014-12-19 at 12.37.43 PM.png

Ed
Level 2
Amanda 12-19-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

Close! The first field is for the parameter name (medium) and the second field is for the parameter value (homepage). It would actually be more like this:

 

URL for redirect in Experiment 1: example.com?medium=homepage 

 

Audience include homepage redirect traffic on LP:

Screen Shot 2014-12-19 at 11.50.25 AM.png

 

Audience to exclude homepage redirect traffic on LP:

Screen Shot 2014-12-19 at 11.50.30 AM.png

 

 

Let me know if you have any additional questions. 

Optimizely
Ed 12-19-14
 

Re: Homepage redirect with url targeting to run experiments on same page simultaneously

Thank you SO much. Can't wait to try it out!

Ed
Level 2