---
title: "Offer tests and templates"
description: "Run an A/B test between two versions of a segment's follow-up sequence and keep the winner, and write, preview and test-send the message templates your follow-ups use."
---

Two screens help you improve your abandoned-cart follow-ups. With **A/B tests** you compare two versions of a segment's follow-up sequence on new carts and see which one recovered more. With **Templates** you write the subject lines and messages your follow-ups send, so a change to the words is made once.

A/B tests come with customer segments, which SureBright switches on per store (see [Customer segments](/abandoned-cart-guide/customer-segments)).

## Test two versions of a sequence

**Follow-ups → A/B tests** lists tests in four groups: **Running**, **Drafts** (built by you, not started), **Suggested** (built by SureBright from your numbers, each saying what it expects and why) and **Past tests**. Nothing changes for your shoppers until you start a test.

![A/B tests](/images/abandoned-cart-ab-tests.png)

### How an A/B test works

A test runs on one segment. **Version A** is the segment's steps as they are. **Version B** is the same steps with what you change. From the moment you start, new carts in the segment are split half and half between the two versions; carts already in a sequence carry on as they were. One test runs per segment at a time.

### Build an A/B test

<Steps>
  <Step title="Which segment">
    Choose the segment. A test on a segment with a running test is refused.
  </Step>
  <Step title="What version B changes">
    Choose **One thing**, so the result tells you exactly what made the difference, or **A few things together**, for changes you think work best as a set. Then pick what changes: **Offer**, **Send time**, **Subject line**, **Message** or **Add or remove a step**.
  </Step>
  <Step title="Version B">
    Each step shows what version A sends now. Only the things you ticked can change; a step added in version B is marked **New step, only in version B**. The editor checks version B against your choice and says what is different, or what you ticked but have not changed yet.
  </Step>
  <Step title="When it ends">
    **End when each version has a number of carts**, **End on a date**, or both, whichever comes first. Once each version has its carts, no new carts join and the test waits for those carts to finish their steps; on the date it ends at once. Then the numbers freeze and you choose which version to keep.
  </Step>
  <Step title="Name and start">
    Name it, then **Start A/B test**, or **Save as draft** to keep it as a draft.
  </Step>
</Steps>

**New A/B test** opens the editor. Below, version B changes only the **Offer**, so each step shows what version A sends now and lets you pick a different offer for version B.

![Building an A/B test that changes the offer](/images/abandoned-cart-ab-test-editor.png)

The line beside the buttons tells you what still needs doing, for example *You picked Offer but version B does not change it yet.* **Start A/B test** stays unavailable until version B is really different from version A.

### Read a running test and keep the winner

Each running test shows **Running since** its start date, then one row per version, with how many carts each version has toward its finish line, with **Abandoned carts**, **Contacted**, **Became orders** and **Order value**. The version with the higher share of orders is marked **Ahead so far**. When the test ends, it says which version won, or **No clear winner** and that keeping version A is the safe choice.

While it runs you can end it now and decide later (**End now, decide later**), or end it and decide at once (**End and apply "Version B"**, **End and keep Version A (no change)**). After it ends, **Apply "Version B"** makes it the segment's follow-up steps, and **Keep Version A (no change)** leaves them as they are. The chosen version is marked **Kept** in **Past tests**, and the numbers stay there. A test you built can be copied with **Duplicate and edit**.

### Start from a suggested test

SureBright proposes tests from your own carts and orders, for example sending the first step inside your store's buying hours instead of a fixed number of hours after abandonment. Each says what it expects and **Why** (*N% of your abandoned carts were left outside that window*). **Start test** opens it to check and start; **Dismiss** removes it for good.

<Note>
With automatic follow-ups off, a test's steps go out only when you send them from each cart, so its results build slowly. Turn them on under **Follow-ups → Automatic follow-ups**. A/B tests are off when Omnisend gets offers by **Write to contact properties**.
</Note>

## Write and test your message templates

![Messages → Templates](/images/abandoned-cart-templates.png)

**Messages → Templates** holds the words each follow-up sends. A follow-up step or a cart can pick a template by name; otherwise SureBright uses the template written for the cart's customer segment, then the default for that purpose.

### What each template purpose is for

| Purpose | When SureBright picks it by itself |
|---|---|
| **Cart abandoned (connected tool)** | Given to a connected tool with its *cart abandoned* event |
| **Follow-up without a discount** | A follow-up with no discount, and information campaigns |
| **Discount offer** | The first follow-up with a discount, **Send now** on a cart, and discount campaigns |
| **Reminder (later follow-ups)** | Later follow-up steps and the reminder before a discount code ends |
| **Discount code about to expire** | Given to a connected tool with its *code expiring* event, and last-call campaigns |
| **Comparison campaign** | Comparison campaigns only |
| **Picked by name only** | Never picked by itself; only when a step or cart names it. No longer offered for new templates |

### Write a template

**New template**, or **Edit** on a template card, opens the editor.

![Writing a new template](/images/abandoned-cart-template-editor.png)

| Field | What to enter |
|---|---|
| **Name** | Shown wherever a step or cart picks a template |
| **What it is for** | One of the purposes above. **Picked by name only** stays on older templates that have it but cannot be chosen for a new one |
| **Type** | **Email** |
| **Written for** | **Everyone**, or one customer segment, whose carts use it first |
| **Subject** and message | The words. **Insert a placeholder** adds one of the placeholders below |
| **Use as the default "…" email** | Makes it the fallback for that purpose, for example **Use as the default "Discount offer" email** |

**Insert a placeholder** on the right adds one where your cursor is, in the subject or the message, and says what each one shows when the cart has no value for it, for example *there* for a missing first name. The **Preview** below it fills the template with a sample shopper, product and plan so you can read it as the customer would. Each template card lists where it is **Used in**. A template named by a segment step or by a draft or running A/B test cannot be deleted.

### Placeholders you can use in a template

| Placeholder | Fills with |
|---|---|
| `{first_name}` | The shopper's first name, or *there* |
| `{product}` | The covered product's name |
| `{plan}` | The plan attached, for example *3-year protection plan* |
| `{code}` | The single-use discount code |
| `{expires}` | The day the code ends |
| `{shop}` | Your store's name |
| `{discount}` | What the offer takes off the plan, for example *50% off* |
| `{plan_price}` | What the plan costs with the offer |
| `{checkout_link}` | The checkout SureBright prepared, with the plan at the offer price |

When a step sends no discount, `{code}`, `{expires}`, `{discount}` and `{plan_price}` cannot be filled (and `{plan}` too on a **Cart items only, no plan** step). The template card says so, and SureBright sends its built-in words for that step instead.

### Send yourself a test email

The **Send yourself a test** card sends the Shopify invoice email a shopper would get to your own address, using the template you pick under **Which template** or your top offer rule's words. A $1 test item stands in for the cart and no discount code is made, so `{code}` reads *your code*. The draft order it creates in Shopify is tagged **Test** and safe to delete. Only email templates can be test-sent here.

<Note>
If Omnisend gets the offer by **Write to contact properties**, SureBright's templates are not used and **Templates** is hidden. If your Klaviyo flow or Omnisend automation sends the emails from **SureBright: Offer sent**, the template's subject and message travel with that event, and your email shows them only where its template uses those event properties. Test from the tool's window under **Connected tools**; see [Channels](/abandoned-cart-guide/channels).
</Note>

## Related

- [Customer segments](/abandoned-cart-guide/customer-segments)
- [Automatic follow-ups](/abandoned-cart-guide/automatic-follow-ups)
- [Channels and connected tools](/abandoned-cart-guide/channels)
- [Abandoned Cart Recovery](/abandoned-cart-guide/abandoned-cart)
