This won't always be an easy investigation. We have locations, segments, test/target delivery, and then what your code is actually doing. Each of these can fail. We will try to reach an answer as quickly as we can.
What does the Force Experiment Widget show?
We find this to be a great place to start as it describes many issues in one click.
My test isn't showing in there.
Make sure you're looking for the right ta_ value. It's not always obvious, and doesn't update if you ever change the project name.
If it's still not there at all, the problem is probably your location targeting.
Check your locations, including using a unique site object and making sure the URLs/Regexs match.
Not sure what your location is?
On the Experiences list screen, go to Metadata:
You'll find the name of your location in the first column:
My test is there, but I'm not falling in.
Check the state of the test and of yourself
If it's in staging mode, are you?
If it's live, are you in "normal" mode?
If the states match, the problem is segmentation or throttling.
If the states don't match, update your user's state and try again.
If the problem appears to be segmentation:
You can check your entry criteria by adding the debug query string onto the URL:
?_wt.debug=vvv
Filter the console logs by "segment" and you'll see these:
control_response: JSON Control Response Log:
[SERVER LOG] : Lookup OTS Package (GUID: 2100000-ta_UKCartSocialProofing) ---
[SERVER LOG] : Found packages in cache. (count: 1)
[SERVER LOG] : Manage Packages ---
[SERVER LOG] : Start Segmented Run Evaluation : (GUID: 2100000-2522727-2522733) ---
[SERVER LOG] : --- Segment GUID : /item/2100000/84001/2271296/
[SERVER LOG] : --- Start Segment Execution
[SERVER LOG] : --- Segment Output: Does 'something1' match the regex 'something2' ?
[SERVER LOG] : --- Segment Output: no it doesn't
[SERVER LOG] : --- Segment Output: Does 'something1' match the regex 'something3' ?
[SERVER LOG] : --- Segment Output: no it doesn't
[SERVER LOG] : --- Segment Output: Does 'something1' match the regex 'something1' ?
[SERVER LOG] : --- Segment Output: yes it does
[SERVER LOG] : --- Segmented Run Evaluation (Result: true)
[SERVER LOG] : --- End Segment Execution
[SERVER LOG] : --- End Segmented Run Evaluation : (GUID: 2100000-2522727-2522733)
This can be read as:
We're evaluating ta_UKCartSocialProofing
We have failed a few checks
We passed the last one
The result was true, and we were issued an experiment.
Failures will be called out in this logic too, as well as being throttled-out, so all entry criteria scenarios can be understood there.
For more details on interpreting these logs, see: Debug Mode > Test evaluation
My test is there but I'm in the wrong variation.
Variations are assigned at random for tests based on your % split. Click on the variation you want to see and try again.
If you feel the percentage split isn't behaving as expected, clear your cookies and you will be re-evaluated.
Note that 50/50 split doesn't mean you will enter every other time. It's a 50% chance, and while the probability is less likely that you'll be served the same variation across 10 tries, it is a slim possibility. Moreso if you're on a 90/10 split, where you may need many attempts to fall into the 10% variation.
My test is there and it says I'm in the right variation, but I don't see my changes.
Your code may be throwing errors.
Add ?_wt.debug=v to your URL and reload the page. If you see any blocks of red, you'll likely need to deal with them especially if/where they relate to your test.
Check if your code is reaching the page.
If you're writing CSS to the page, check the devtools > elements tab. If your code is there, you know the test is being delivered to the page.
If it's just JS, try adding a console log to the top of pre-render. If this fires, you again know your test is trying to run and the problem is likely in your code.
If you see your code reaching the page, it will be up to the developer to understand why the code isn't doing what you hope it to.
My scenario isn't covered here
When asking for support, it is very helpful to walk through these steps yourself and tell us the outcome of these. Quite often it will guide what comes next, and we won't get stuck in a back-and-forth looking at basic troubleshooting.
We are happy to help solve the above problems, but many are simple problems you can resolve without external help, and in a much shorter timeframe (5 minutes, instead of a few hours).