You find a winning offer by changing one thing at a time and counting what happens after the click. For one week, keep your creative style the same, test one variable (price, bonus or deadline) and measure sales or sign-ups per 100 link clicks. Likes tell you about the post. Only that ratio tells you about the offer.
TLDR: Day 1 sets the baseline. Days 2 to 6 change one variable per day. Day 7 picks the winner by sales per 100 clicks and runs it again.
Decide what you are testing before you post
The most common failed test changes everything at once: a new video, a new price and a new caption. When sales move, nobody knows why.
Pick one variable from this list and write it down:
- Price or discount: 20% off vs 10% off plus a gift.
- Bonus: free shipping, a template, a free call.
- Deadline: 24 hours vs a vague "this week".
- Proof: a customer quote vs a result number.
Then fix the control: the same post format, the same length and roughly the same posting hour. If you usually post carousels, test with carousels.
You also need one number that answers the question. Use sales (or sign-ups) per 100 link clicks. Reach and likes are fine as extra context, but they are not the verdict.
The 7-day plan
| Day | What you do | What you record |
|---|---|---|
| 1 | Post your current offer as the baseline | Clicks, sales |
| 2 | Offer A: change the one variable | Clicks, sales |
| 3 | Re-post the baseline in a new format or caption | Clicks, sales |
| 4 | Offer B: a second version of the variable | Clicks, sales |
| 5 | Re-run the stronger of A and B | Clicks, sales |
| 6 | Reply to every comment and DM about the offer | Questions asked |
| 7 | Add up the table and decide | Winner and next test |
Why re-run on days 3 and 5? Because one day can be a fluke. A winner that repeats is far more believable than one that appeared once.
Make the click-to-sale path measurable
If you cannot see what happens after the click, you are guessing. Give each offer its own link, or at least its own link inside your link-in-bio page, so you can tell which post sent which visitors.
Keep the landing page the same across variants. Otherwise you are testing the page, not the offer. Our guide on turning clicks into sales from a link in bio covers the page side of this.
Read the results like a skeptic
Small numbers fool people. If Offer A got 4 sales from 60 clicks and Offer B got 3 from 55, that is a tie, not a winner.
Use these rules:
- Need a gap of at least 30% in sales per 100 clicks before calling a winner.
- Need the same direction on both days the offer ran.
- If it is a tie, pick the cheaper offer. It costs you less to give.
Also remember benchmarks. In Mydrop's data, a link in the caption cost 7% of an account's typical engagement. That is fine: you are paying a small reach price to get a trackable click. Just do not read a slight engagement dip on offer posts as failure.
Know what to do with a losing week
Sometimes no version wins. Every offer lands at the same low ratio, and that is useful news. It usually means the problem sits upstream: the audience is not a buyer, or the product page raises doubts the offer cannot fix.
Before you test again, read the comments and DMs from day 6. People tell you what stopped them: "too expensive", "not sure it fits me", "how does shipping work?". Fix the most repeated objection on the page, then rerun the plan with the strongest offer as the new baseline.
Plan and schedule the whole week in one sitting
Offer tests fail when someone forgets to post on day 4. Build all seven posts on Sunday and schedule them. In Mydrop, you can draft them in the content calendar, keep a note on which variable each one tests, and publish without remembering anything midweek.
Then reserve 20 minutes on day 7 to fill in the table. Write the result in one line: "Free bonus beat 20% off, 9 sales per 100 clicks vs 5." Next week, test the next variable against the winner.
To start, create a free Mydrop account, schedule your baseline post for tomorrow, and write down your single variable before you go to bed.
Google review
Trustpilot review
Product Hunt comment
