When you're testing e-mail promotions to determine which approach will produce the best results, you'll want to be testing a number of things, including:

Same offer; different subject lines. Test the exact same e-mail promotion in the exact same format and change only line of each.

Same subject line; different letters. Use the same subject line And the same format , but alter the body of your letter slightly. This doesn't mean you have to write an entirely new letter; instead, try tweaking just one section so you can determine which change makes the difference.

Different opening paragraphs. It's a good idea to create a letter that works, then try a couple of different opening paragraphs. This is area of the letter that will have the biggest impact on your sales, because this is where people make their "read it or delete it" decision.

Different offers. You don't always have to be testing the effect of different sales copy – you can also try testing the appeal of offer versus the interest in another.

Different price points. You should always test a few different prices for a new product or promotion. It's amazing but true – a product will often sell more at a higher price then at a lower, discounted price…though often somewhere in between is your best bet. Test at least three price points for any product.

Different angles. Try a couple of different angles in your promotions.

Different sales letters. If your e-mail promotions link your customers to a sales letter at your web site, try using the same mailing and the same subject line, but linking to different sales letters or web page (with different headlines, angles, etc).

Different formats. While we used to send almost all of our e-mail promotions in HTML format, we have recently discovered that our audience is more responsive to mailings in plain text format. Again, if we weren't constantly testing and tweaking our promotions, we never would have learned this.

Short copy versus long copy. Depending on your promotion, the length of your e-mail can play a large in its success. Typically, e-mails that are more than two pages in length don't get read – but this is not always the case. Sometimes you need that extra, persuasive text to convince your potential customers that your offer really is the best. Try testing a three-quarter-page email promotion versus one that is closer to one and half pages in length.

The single most important thing you can do to guarantee the long term success of your e-mail marketing program is to constantly test every aspect of your promotions.

So what, exactly, should you be testing? Anything and everything. You should try testing things like:

Subject lines
Offers
Opening paragraphs
Price points
Angles
Sales Letter
Formats