You Can't A/B Test Your Way Out of a Bad Offer
You spent Sunday afternoon on your link in bio page. You rewrote the top button from "Check out my shop" to "Shop the new drop." You tried a warmer accent color. You moved the newsletter signup above the shop link, then moved it back. You read a post about button copy and rewrote everything again.
Monday's numbers came in. Clicks were up four percent. Sales were exactly the same: two.
This is the moment where most people conclude they need to optimize harder. They do not. They have been polishing the sign that points at the thing, when the problem is the thing.
Optimization is a multiplier, and multipliers need something to multiply
Every conversion tip you have ever read is a multiplier. Better copy, better hierarchy, better contrast, fewer choices: all of it takes whatever demand already exists and squeezes a bit more of it through. That is genuinely valuable. A well built page can pull twenty or thirty percent more out of the same traffic, which over a year is real money.
But a multiplier applied to zero is still zero. If nobody wants the thing at the price you are asking, in the form you are selling it, right now, then a warmer orange on the button changes nothing. You will run the test, see a rounding error, decide the test was inconclusive, and run another one. That loop can eat months.
The reason this loop is so sticky is that optimizing feels like work and it is entirely under your control. Reworking your actual offer feels like admitting something. So people redesign the page for the fourth time instead.
I want to be clear about what I am not saying. I am not saying design does not matter, and I am not saying stop testing. We built a lot of Liinks around making those tests easy for a reason. I am saying that testing is the second question, and almost everybody skips the first one.
Two problems that look identical from the outside
From where you sit, a page problem and an offer problem produce the same symptom: not enough sales. That is why they get confused constantly.
They are not the same thing at all.
A page problem means people never reach your offer. They land, they scan, they get confused or bored or distracted, they leave. The offer might be great. Almost nobody actually saw it.
An offer problem means people reach your offer, look at it properly, and decide no. They read the sales page. They opened the pricing. They understood exactly what you were selling. And they closed the tab.
The difference matters because the fixes have nothing in common. A page problem is fixed with layout, clarity, and hierarchy. An offer problem is fixed by changing what you sell, who you sell it to, or what you charge. Applying the first fix to the second problem is the loop I described above.
How to tell them apart in about ten minutes
You do not need a funnel diagram or an analytics degree. You need two numbers and one honest look.
Number one: how many people who land on your page click the offer link at all? Open your Liinks analytics and compare page views to clicks on the specific button that leads to the thing you are selling. Not total clicks across every link. That one button.
Number two: how many of those clickers buy?
Now read the pattern:
- Low click rate on the offer button, and the few who click convert fine. Page problem. Your offer is fine and buried. Fix the page.
- Healthy click rate, almost nobody buys. Offer problem. People are seeing it and saying no. Do not touch the page.
- Low on both. Start with the page, because you do not yet have enough people looking at the offer to judge it. You are drawing a conclusion from six visitors.
- Healthy on both. Nothing is broken. You just want more traffic, which is a different post.
What counts as "healthy" varies wildly by audience and platform, so do not chase a benchmark you read somewhere. Compare against your own other links. If your free playlist link gets clicked by a quarter of visitors and your paid course gets clicked by two percent, that gap is the signal, whatever the absolute numbers are. If you want to go deeper on reading these numbers without spiraling, we wrote about that in CTR in real life.
The honest look is the third input, and it is the one people skip: go click through your own offer as a stranger would. Open it on your phone, from the link in your bio, the way an actual visitor arrives. Not from your bookmarks. Not while logged in. If you cannot tell within about eight seconds what you get, what it costs, and who it is for, you have found something no A/B test was going to surface.
Four ways an offer goes bad
"Bad offer" sounds harsh, and it usually is not about quality. The work is often excellent. Here are the four failures I see most, roughly in order of how common they are.
It is aimed at everyone. A course called "Grow Your Instagram" competes with every free thing on the internet. A course called "Grow Your Instagram as a Wedding Photographer Who Hates Being on Camera" sells to a smaller group who feel personally addressed. Narrow is not a limit on your market. Narrow is the thing that makes people believe you are talking about their situation specifically.
It asks for too much, too early. Somebody discovered you eleven seconds ago and the first thing you ask for is 400 dollars and a six week commitment. The offer is not bad, the timing is. People buy after trust and the offer arrived before any existed.
It is vague about what actually arrives. "Coaching" is not an offer. "Four one hour calls over four weeks, plus a written plan after each one, plus voice note access between calls" is an offer. Vagueness reads as risk, and risk loses to the close button every single time.
The price does not match the story. Too high with nothing to justify it and people stall. Too low and people quietly assume it is not serious. Both are offer problems, and both are invisible to a button test. We wrote a whole piece on the adjacent version of this, which is refusing to show prices at all: nobody is going to DM you for prices.
Notice that not one of these is fixed by page layout.
Fixing the offer instead of the page
Changing your offer sounds like a quarter of work. It usually is not. Four things that take an afternoon each:
Say the outcome, not the format. People do not buy a PDF, a call, or a template pack. They buy the state they will be in afterward. "A 40 page guide to lighting" is a format. "Stop shooting in bad light and know exactly where to stand in any room" is an outcome. Same product, completely different answer to the only question the buyer is asking, which is what changes for me.
Cut the offer in half and charge less. Not forever. As a test. If the six week program is not selling, sell the first two weeks as its own thing. Smaller commitments give you actual data in weeks rather than a year, and everyone who buys the small thing is a warm lead for the big thing.
Add proof rather than adjectives. One screenshot of a real result outperforms three paragraphs of description. A single specific sentence from a real client beats a wall of stars. If you have no proof yet, that is itself the finding: your next move is getting one result you can show, possibly for free, not rewriting your headline again.
Change who it is for and keep everything else. The cheapest experiment on this list. Same product, same price, new audience. Coaches selling to beginners often find the identical program sells three times better positioned for people who already started and stalled.
Run one of these at a time and give it two weeks. If you change the offer, the price and the page all in the same week, you will learn nothing about which one mattered. This is the same discipline as good page testing, which we covered in our guide to running A/B tests without losing your mind, applied one level up.
Be fair to the page, though
There is an obvious failure mode in the other direction: deciding your offer is doomed when actually nobody ever found it.
If your page has fourteen links, your offer is link number nine, and the top three slots go to your Instagram, your TikTok and your YouTube, then your offer does not have an offer problem. It has a real estate problem. People arrived from Instagram and the first thing you offered them was a link back to Instagram.
Some honest checks before you blame the product:
- Is the offer above the fold on a phone, or does it require scrolling?
- Does the button say what happens next, or does it say "Link"?
- Is the offer competing with five free things sitting right next to it?
- Has the page changed at all since you launched the offer, or is the offer just quietly sitting in the pile?
If the answer to any of these is uncomfortable, fix that first. It is a one afternoon job with Liinks: reorder, put the offer in its own section with breathing room around it, write a button that names the outcome, and move the social links to the bottom where people who want them will still find them. Then wait two weeks and look at the offer click rate again. Now you have a clean read on whether the offer itself works, because you know people are actually seeing it.
That order matters. Page first when nobody is reaching the offer. Offer second, once they are.
The uncomfortable part
Sometimes you do all of this and the answer is that the thing you built is not something enough people want to buy right now.
That is not a failure of nerve or effort, and it is not permanent. Products get repositioned constantly. Audiences shift. Timing is real. The people who get through it are the ones who found out in six weeks instead of eighteen months, because they went and looked instead of redesigning the page one more time.
The page is knowable. Traffic is knowable. Whether people want the thing is knowable too, and it is the cheapest thing on the list to check. It just requires asking a question that might come back with an answer you do not like.
TL;DR
- Optimization multiplies existing demand. It cannot create demand that is not there.
- A page problem and an offer problem look identical from the outside: not enough sales.
- Two numbers separate them: what share of visitors click your offer button, and what share of those clickers buy.
- Low clicks and good conversion means fix the page. Good clicks and no conversion means fix the offer. Do not fix both at once.
- Offers usually fail for four reasons: aimed at everyone, asking too much too early, vague about what arrives, or priced out of step with the story.
- Fix an offer by naming the outcome, shrinking it, adding proof, or changing who it is for. One change at a time, two weeks each.
- Be fair to your page first. An offer buried at link nine never got a chance to fail.
Try it on your own page this week
Open your page. Find the one link that is supposed to make you money. Check what share of your visitors actually click it, then go through the checkout flow on your phone as a stranger would.
Ten minutes gets you an answer that four more rounds of button copy would not.
Liinks gives you the page and the click data in the same place, so the diagnostic above takes minutes rather than an afternoon of exporting spreadsheets. Build your page free, put your real offer at the top of it, and let the numbers tell you which problem you actually have.
