Can I schedule an update to an existing Adobe Target A/B test?

[aaronlng_92]

I'm working on setting up a test that is swapping out different graphics on landing pages. These graphics change at different intervals of time and I'm wondering if there is a way that I can go into the test while it is live and schedule on a certain day and time for the test to update to match the live site? Or does this have to be done manually every time something changes on the live site? 

[Gokul_Agiwal · Accepted solution]

Hi @aaronlng_92 

You can use the option called Time Frame while creating audience and from there you can add start and end dates and times in Adobe Target to target users who visit your site during a specific time frame. You can also set Week and Day Parting options to create recurring patterns (frequency) for audience targeting.

More https://experienceleague.adobe.com/en/docs/target/using/audiences/create-audiences/categories-audien... 

Hope this helps. 

[henrikk]

Hello! What if I want to do it without using audiences? Because if you use audiences then you would need to use experience targeting, correct?

What if there's just a content update you want to schedule, for example only affecting variation 1 (B) that you want to schedule. Is that possible without some sort of convoluted method of creating audiences that match certain days/hours and a separate audience that matches a different set of days/hours?

[Gokul_Agiwal]

Hi @henrikk 

The Time Frame options you can use for A/B testing too. 

On the another question regarding the content update scheduling - As far as I know there is no direct option available based upon the content/variation/experience you want to trigger/schedule at certain time however you can do this on activity level itself.   

You can schedule the activity, specify a date and time for delivery of the activity. Scheduling controls the delivery time frame of the activity; however, the activity must be explicitly activated before it can be delivered according to the specified schedule. 

You can see this option in Goals and Setting page like below 

Hope this helps. 

In case if there are other options are available, people in CC might share some lights in this thread. . 

CC: @kandersen   

[henrikk]

Hi @Gokul_Agiwal, 

Thanks for replying!

I'm not clear how the time parting audience would solve things. Imagine the below scenario:

Monday we launch a simple A/B on a landing page. On Tuesday we change some content on the live experiment and only in the variation but we can't show it to the public until Wednesday 3am. So we change the variation content and want to schedule it in advance to go out on Wednesday at 3am. However we want to keep the original experiment running throughout this time and we want the people in the variation to continue to stay in the variation throughout the time.

My understanding is that if you tried to do the above, by scheduling the change on Tuesday to go out on Wednesday at 3am, this causes Target to DEACTIVATE the currently live experiment and schedule it to go back out live on Wednesday at 3am. But I would not want the experiment to deactivate and I would also not want the initial variation audience to change at all from Monday through to Wednesday (except of course on Wednesday after 3am they would now see the content updated on the landing page, which is expected).

How would the above the possible without using experience targeting and instead a simple standard A/B test?

Thanks!

[kandersen]

Hi @henrikk 

My understanding is that you have an A/B test running - at some point you need to change the content of variant B. This needs to happen at a certain day and time - up until this day and time, you still want the 'original' content of variant B to show up until this time. Is that correct understood?

There's no feature in Target that supports this. However, the way i would approach it would be creating variant B using code. So in the VEC I would add a modification using custom code. 

Here's a simple script that changes the src at a certain day and time, that is defined:

(function() {
    // Define the date and time limit (format: YYYY-MM-DD HH:MM:SS)
    var deadline = new Date("2024-12-01 14:00:00").getTime();

    // Define the image element selector
    var imageSelector = '#image-id'; // Change this to your image's selector (ID or class)

    // Define the source URLs for the image
    var imageBeforeDeadline = "https://example.com/image-before.jpg";
    var imageAfterDeadline = "https://example.com/image-after.jpg";

    // Get the current date and time
    var now = new Date().getTime();

    // Find the image element on the page
    var imageElement = document.querySelector(imageSelector);

    if (imageElement) {
        // Check if the current time is before or after the deadline
        if (now < deadline) {
            // Before the deadline
            imageElement.src=imageBeforeDeadline;
        } else {
            // After the deadline
            imageElement.src=imageAfterDeadline;
        }
    } else {
        console.error("Image element not found");
    }
})();

 Using the above you can avoid having to manually go in at the relevant time and update the variant. 

Having said this, I would be cautious when evaluating the results if you've been changing the variant during the test period. So make sure you note down the dates/time when you're changing the content so you can take it into considerations when looking at the results.

[henrikk]

Hey @kandersen thanks for getting back to me. Yes, you understood the scenario perfectly. 

We thought about using code to do it but wanted to first see if there was the possibility of using a built-in Adobe Target feature to achieve this. Can we add this to a feature wish list?

Yes, I agree it's not good to change content during the experiment. However, in this case it's a promotion that is changing (the experiment would not explicitly be about the promotion but rather other collective changes on the landing page).

Thanks for sharing the code snippet example!

[Gokul_Agiwal]

Thanks @kandersen  

Yes completely agree with @kandersen  - in the scenario the code would be the ideal option but as Kasper mentioned, make sure you note down the timings for the results. 

@henrikk Good to know that you found solution useful. 

[henrikk]

Can we add this to a feature request @Gokul_Agiwal ?

[kandersen]

@henrikk post it here in the ideas section for it to be reviewed: https://experienceleaguecommunities.adobe.com/t5/adobe-target-ideas/idb-p/adobe-target-ideas