Since you're looking at testing two segments over time, this isn't something you can do using the built-in A/B testing functionality. However, you can accomplish this goal by creating the groups, one for each segment. To start off with, create two empty groups for the segments. You can name them whatever you like, such as Group A and Group B, but I'd recommend something descriptive so that there is no confusion as you go through the test. Note the Group ID for each group.

The basic process will be to run a query that returns everyone in your house file and use a mail merge to export the results to Excel. In Excel, you'll assign a random number to each row and use that random number to assign each constituent to a group. Then, you'll do a custom constituent import to update the group membership of those constituents.

Here's the step-by-step details:
  1. Determine the criteria of the constituents that will be a part of the segments you will use to test. This will be the basis for your query. At the very least, you will want to restrict your test to active constituents with valid email addresses that are opted-in to emails.
  2. Once you have determined your criteria, create a query (help document on working with queries). If you want select everyone, then your query would look like this:
    • System Active Status equals Active
    • AND Accept Email equals True
    • AND Email Status equals Good
  3. Click the Use Query link and select Create a Mail-Merge list.
  4. Complete the mail merge setup. The only required field is the constituent ID, but you can include as many as you like.
  5. Once the mail merge has been generated, download the results and open the results in Excel.
  6. Add two new columns to the left of your data; label them Group_ID and Random.
  7. Enter =RAND() in the top row. This will generate a random number between 0 and 1. Copy the formula to all rows in your data set. Tip: Copy the top cell with the formula, then scroll down to the last row and hold down the Shift key while clicking the last cell. This will highlight all of the cells between the first and last row. When you paste the formula, it will paste to all of the highlighted cells.
  8. Now all of the rows should have a random number assigned. Please note that the random number will change each time there is a recalculation (like when rows are sorted). As a result, we need to change this from a formula to a value.
  9. Highlight the entire Random column by clicking the column header (probably B). Select Copy, then right-click to select Paste As Value. Now these numbers are fixed values and can be sorted.
  10. Sort your data by the random number field. Your data set is now randomized!
  11. Add the Group ID of the first test cell to the top half of the rows, and the other Group ID to the remaining rows. If you want to adjust the test split, such as 25-75 instead of 50-50, this is where you would do so.
    (If you're comfortable with using Excel formulas, an alternate method of adding the Group ID is to use a conditional statement and paste it to all rows. If you do that, you don't need to sort the data. For a 50-50 split, the conditional would look something like this: =IF(B1<=0.5, "GroupID_1", "GroupID_2").)
  12. Now that each row has a Group ID, save your work.
  13. In Luminate Online, go to Import/Export and click Create Custom Constituent Import. Map your fields, set the import to update existing records and run it. The number of records with updated group membership should match the number of records in your file, and your two groups should now be populated.
Now that your two test groups are created and populated, you can use them as the target audiences for your different email messages. In order to maintain the integrity of your test, do not make any changes to the groups until the test is complete.  When you're ready to report on the results of the test, you can use the Email > Group Performance Report to compare them directly, or any other report (such as donations) to evaluate any other differences between the two groups. If you have specific metrics in mind, you can run a few practice reports with other groups to make sure that you'll be able to analyze the results properly.