Draft three versions of an email, send them to real prospects and compare the replies.
Mei sells to cafe groups and writes to their founders. She wants to know whether her cold email works, so she tests three versions of it on real prospects over two weeks and keeps the one that gets the most replies.
Most sellers have sent the same cold email for about a year without knowing whether it works. It has gone out hundreds of times. Rewriting it feels like a gamble, so it stays, and nobody learns anything from all those sends. A small, honest test gives you a result to go on instead of a guess. Writing the versions takes about thirty minutes. After that, sending and recording take a few minutes each day.
Start with version A, the control. This is your current best email, and the other versions are measured against it. If you don't have one you trust, use the email you wrote for lesson 4.1, Anatomy of a cold email that gets a reply.
Versions B and C are challengers. Each one changes exactly one element of version A: the subject line, the first line, or the ask. The rest of the email stays word for word the same. Under the one-change rule, any difference in replies can be traced to the element you changed. Suppose a challenger changed the subject and the ask together and got more replies. You would not know which change helped, and next month you could keep the wrong one and drop the right one.
Mei's version A is the email from lesson 4.1. Its subject, "new outlet in Tampines", is adapted to each prospect's trigger. Version B changes only the subject, to the question "month-end with four outlets?" Version C changes only the ask, from "Worth a short call to see if that would help?" to "Is month-end a headache for you too, or have you got it sorted?"
Save your three versions side by side in one document and mark the element each challenger changes.
All three versions go to people who fit the same persona from lesson 1.2. Each version needs at least ten people, and twelve to fifteen is better if your list allows it. Match the groups as closely as you can on business size, trigger and source. If version A goes to your warmest leads and version C goes to a cold list, a difference in replies could come from the people and have nothing to do with the wording.
Mei has thirty-six cafe groups that fit the founder persona. She sorts them by number of outlets, then deals them out like cards: one to A, one to B, one to C, and repeat. Each version ends up going to twelve similar businesses.
Before you send anything, run the module 5 checks on your list, especially lesson 5.4, Build a compliance checklist for your outreach.
Send all three versions across the same two weeks and at similar times of day, so one version doesn't get all the Monday mornings. Personalise each prospect's first line using lesson 4.3, Personalise in five minutes, not fifty. Leave the element you're testing exactly as written.
Count replies. Don't count opens. Many email tools report opens, and the numbers look precise. An open is usually tracked through a tiny image inside the email, though. Some email apps load images automatically to protect privacy, and that registers as an open whether or not anyone read the message. Other apps block images, so some real reads are never counted. A reply is unambiguous because someone read the email and wrote back.
Every reply counts, including "not interested" and "not now". A polite no still shows the email was read and answered, and the test measures exactly that. Write down the number of positive replies separately. Record the counts in your scorecard from lesson 2.4, Build a lead source scorecard, with one row per version.
At the end of the two weeks, turn each count into a rate. Mei's results are:
Version A: 2 replies from 12, about 17 percent Version B: 4 from 12, 33 percent Version C: 1 from 12, 8 percent
With groups of about twelve, one or two replies either way could change the order of the versions, so act only when the gap is clear. A gap of four replies against one is worth acting on. Two against three is too close, and with a result like that the honest conclusion is that you don't know yet.
Mei keeps version B because the gap is clear enough to act on and the new subject line is cheap to keep. She doesn't decide that question subjects always win. She has one result, from one month, with one persona, and one result does not prove a general rule.
The winning version becomes next month's control. Write one new challenger that changes a different element, and run the test again. Mei's next challenger keeps the question subject and tests a new first line built around Google reviews instead of job ads. Over several months, each small win builds on the last one, and your email improves based on evidence about replies, not on opinions about what sounds good.
Now run the full two-week test on your own list. When it ends, you should have three versions saved side by side with the changed element marked, and a reply count for each version in your scorecard. Then write one sentence naming the version you'll keep and the change you'll test next.
Over two weeks, send three versions to similar groups, record the replies for each in your scorecard, and write down which version you will keep.
Junxiong-WFG Organisation is an authorised representative of AIA Financial Advisers Private Limited (Reg. No. 201715016G).