There are a few different ways to preview experiments. We will go through them in the order of our preference, and will explain why.
Force Experiment Widget
What is it?
This is a bookmarklet for your browser. You can click it when on your test page, on your website, and it will list out all experiments that are being served, which you've fallen into, and give you the option to change your view.
This is a significant improvement on Preview Links, which say "Show me Exp2 for Test 20" but give you no context of what else is on the page, and where potential overlaps may occur when you go live. It's for this reason that we recommend the FXP widget over all other methods of previewing.
It is possible to get this working across all desktop, tablet and mobile devices, as well as simulation tools like Browserstack.
Installation
This is just a bookmark you can keep in your browser. You'll find a link for it when on the Preview Links pane in the dashboard:
Or, you can browse through to this link:
Usage
Browse to any page you wish on your website, and then click the bookmark.
It will launch after a couple of seconds, and show you all experiments that are running on that page. For example, in our UI, we see:
This shows that I'm treated as a "normal" user, i.e. not in Staging Mode and not previewing anything actively. It also shows that I'm in a VE Widgets, Welcome Video and Intercom target, and a New Navigation test where I'm in the control group.
To become a staging user, simply click "normal" in the top-right corner and it'll change to staging. The page will refresh and you'll be evaluated for staging tests.
Or, to see any specific experiment, click that number - e.g. "2" for the New Nav test to see the new navigation variant. Again, the page will refresh, cookies adjusted, and you'll see that variant.
Limitations
For tests:
Able to preview a specific variation - yes
Able to check Segmentation works - yes
Able to check Location works - yes
Able to check if Metrics track - yes
For targets:
Able to preview a specific variation - no
Able to preview if there's only 1 variation - yes, as long as you meet the segmentation criteria set for that rule if any.
Able to check if Segmentation works - yes
Able to check Location works - yes
Able to check if Metrics track - yes
Leveraging Staging Mode
What is it?
Sometimes, it's enough just to say "I'm a staging mode user" and let the system do the rest.
For example, let's say you have built an all-user target for your homepage for an upcoming promotion, and want to make sure it works.
In this scenario, it's sufficient to just identify yourself as a staging user, and let the system serve you the right experience like it should do.
Usage
There are two key ways to enter staging mode:
1. Staging Mode Query String
You can add _wt.mode=staging into your query string to identify as a staging user. For example:
becomes
2. Force experiment widget
As mentioned above, the top-right corner of the FXP widget allows you to change your User Mode from "normal" to "staging" and back.
Pinning a variation
This is a useful feature to have alongside staging mode.
When users are evaluated for an experiment, they are randomly allocated variation A or B based on the % split you have, e.g. 50/50.
You have the ability to bypass this, and "pin" an experiment such that all users being assigned a test will be given a specific variation if they get past segmentation.
Once saved, you'll only be given the experiment you select. This applies to all users, so can help your colleagues too e.g. when sending the test to the QA team.
Just remember to set it back to "At Random" once you're done.
Limitations
For tests:
Able to preview a specific variation - yes, with pinning
Able to check Segmentation works - yes
Able to check Location works - yes, by inference
Able to check if Metrics track - yes
For targets:
Able to preview a specific variation - no
Able to preview if there's only 1 variation - yes, as long as you meet the segmentation criteria set for that rule if any.
Able to check if Segmentation works - yes
Able to check Location works - yes, by inference
Able to check if Metrics track - yes
Preview Links
What is it?
Preview links are shareable URLs that allow you to see a specific variation that you've built.
They don't offer any context on what else is on the page, just very specifically request a given variation.
Note:
The preview links regenerate on change. This is a known confusing part of the feature that we are improving. Old links will still function, even if not displayed.
Usage
You'll find preview links in the Preview Links pane in the dashboard. They are also found in the Advanced Editor, using the Preview button at the bottom. We are also looking for other useful places to insert them.
The contents of the text box comes from the Editor URL of your location, if you've specified one. If you haven't, the URLs might look broken as per the screenshot - in that situation, just put a correct URL into the box and all of the links will adjust accordingly.
You can share these with anyone - they don't need to be a user in the UI to preview your experiments using these links.
Limitations
For tests:
Able to preview a specific variation - yes
Able to check Segmentation works - no, the check is skipped
Able to check Location works - no, the check is skipped
Able to check if Metrics track - no
For targets:
Able to preview a specific variation - yes
Able to check if Segmentation works - no, the check is skipped
Able to check Location works - no, the check is skipped
Able to check if Metrics track - no
Note that we are working on improving our preview links to overcome some of these challenges, or better embed the Force Experiment Widget into the preview workflow.