> For the complete documentation index, see [llms.txt](https://docs.cymbal.co/help/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.cymbal.co/help/text/a-b-testing-text-campaigns.md).

# A/B testing text campaigns

Test 2 to 6 versions of a text campaign on an even split of your audience and compare how each one performs.

Not sure which message will land best? Add variants to a text campaign to send different versions of your message to your audience. Cymbal splits everyone in the campaign evenly across your variants, then shows you which version had the best delivery, click, conversion, and opt-out rates.

Every recipient gets one variant at send time. There's no test round first, and Cymbal doesn't automatically send the winning version to anyone afterward.

## How to add variants to a text campaign

You can only add or change variants while the campaign is still a draft.

{% stepper %}
{% step %}
Go to [Text](https://manager.cymbal.co/text) and create a new campaign, or open a draft campaign.
{% endstep %}

{% step %}
Click **Create A/B test** next to the campaign title.
{% endstep %}

{% step %}
Use the variant pills (**Variant 1**, **Variant 2**, and so on), under the **A/B test** label, to switch between versions. Each variant has its own message, dynamic data, and optional image.
{% endstep %}

{% step %}
Add more variants or remove one until you have between **2 and 6**. Each message follows the usual [text message length](https://docs.cymbal.co/help/pricing/sms-pricing) limits.
{% endstep %}

{% step %}
Pick your audience, then send or schedule the campaign as usual. Once it leaves draft, the variants are locked.
{% endstep %}
{% endstepper %}

## How your audience is split

Cymbal divides the whole audience evenly across your variants. It mixes in contacts who usually engage alongside those who rarely do, so each variant gets a similar mix and the results are a fair comparison. If the audience doesn't divide evenly, the first variants get one extra contact each.

Because variants can have different lengths or images, the cost estimate before you send breaks segments down per variant. See [SMS pricing](https://docs.cymbal.co/help/pricing/sms-pricing) for rates.

## How to compare results

{% stepper %}
{% step %}
Open the sent campaign from [Text](https://manager.cymbal.co/text) and go to **Analytics**.
{% endstep %}

{% step %}
Use the **All variants** pill to see the whole campaign, or pick a **Variant** pill to see one version on its own.
{% endstep %}

{% step %}
Check the comparison table. It shows each variant's delivery rate, click rate (when your message has links), conversion rate (when conversions are tracked), and opt-out rate.
{% endstep %}

{% step %}
Look at the **Winner** column to see which variant did best on each metric. For opt-outs, the lowest rate wins. If two or more variants tie, all of them are listed as winners, and a variant that sent no messages can't win.
{% endstep %}
{% endstepper %}

Each variant also shows its sends, link clicks, conversions, and revenue, and your recipients export lists which variant each contact got.

{% hint style="info" %}
**Email campaigns too.** You can also A/B test an email campaign with 2 to 6 variants. The steps, audience split, and results on this page are for text campaigns.
{% endhint %}

## Related

* [How to launch an effective SMS campaign](https://docs.cymbal.co/help/text/sms-best-practices/how-to-launch-an-effective-sms-campaign)
* [How revenue attribution works for Text](https://docs.cymbal.co/help/text/text-settings/how-revenue-attribution-works-for-text)
* [Text business hours](https://docs.cymbal.co/help/text/text-settings/text-business-hours)
* [SMS pricing](https://docs.cymbal.co/help/pricing/sms-pricing)
