Summarize with AI
If your conversion rate is not improving after several landing page rewrites, the problem is probably not the page. Look at what happened to signups after each rewrite. If signups went up and paying customers stayed flat, the page did its job, and the leak sits in the product, somewhere between the signup button and the first moment the product is worth it.
This matters because a rewrite is not free. Each one takes a few weeks of the team, and an early company counts its runway in months. A founder who rewrites the page three times in a quarter is not being careless. He is acting on the only evidence he has, and that evidence is collected almost entirely above the signup button.
A sheet called Things we tried
Picture an AI tool for small machine shops. It reads a customer's request for quote, with the drawing attached, and drafts the quote: material, machine hours, finishing, and a red flag on any tolerance the shop cannot hold. This is an example, not a client. The founder keeps a shared sheet called Things we tried. It has four columns: Date, Change, Signups per week after, and New paying accounts the month after. The numbers below are illustrative.
- 12 Aug. New headline, Quote every RFQ the same day. Signups per week went from 38 to 51. New paying accounts stayed at 6.
- 2 Sep. New pricing. The free plan went from 20 quotes a month to 50. Signups went from 51 to 64. Paying accounts went from 6 to 5.
- 23 Sep. New hero. A product video replaced the illustration. Signups went from 64 to 70. Paying accounts went back to 6.
- 14 Oct. A testimonial carousel under the hero, with three logos. Signups went from 70 to 72. Paying accounts stayed at 6.
Now sign up. The first screen is called Quotes. In the middle it says No RFQs yet. Under that is a line that reads Forward RFQs to this address, then a long generated email address with a Copy button. There is nothing else on the screen. The person who just signed up is the shop's estimator, at a desk next to the floor in the middle of the afternoon. RFQs land in the owner's inbox, and the owner is out on a machine. So the estimator copies the address, closes the tab, and the drafted quote the video promised never appears. The design file for that Quotes screen was last edited in February.
Read the sheet as evidence, not as a to-do list
Most founders read a sheet like this as a record of effort. Read it instead as data about where the problem is. Two things stand out. First, every row lives above the signup button: headline, pricing, hero, proof. Not one row touches a screen a customer sees after signing up. Second, the signup column moved every single time, and the paying column did not move at all.
That second pattern is the answer. A page that lifts signups from 38 to 72 a week is a page that persuades. It brought nearly twice as many people to the Quotes screen. The Quotes screen then turned that larger crowd into the same six customers. Here is the claim most conversion advice gets wrong: each of those four rewrites was a success at its own job. Testing the page harder will keep producing small wins in the signup column and nothing in the column that pays the bills.
Why every tool points you back at the page
The founder in this example is not being lazy. He is being rational. His website has page analytics, a heatmap of the hero, scroll depth, and an opinion from every advisor who visits it. His product has two events. One fires when signup is completed. The next fires when a quote is drafted. Between them sits the whole Quotes screen, and it records nothing. Nobody emails support to say the RFQs go to someone else. They just leave.
So watch what he does on a normal week. He opens the analytics tool on Monday, and every chart in it is about the page. He books a designer for a new hero. In sales calls he shares his screen and drags a sample drawing and RFQ from his desktop into the product, and those trials tend to turn into customers. He reads that as proof the product is fine. What it really proves is that the product works once someone has an RFQ in it, and he is the only person who always arrives with one.
The upstream or downstream test
Before the next rewrite, answer three questions from numbers you already have. They take an hour, not a sprint.
- Did signups move after the page changes? If the signup column rose after at least two rewrites, the page is persuading people. That half of the funnel works.
- Did paying or active accounts move with them? If signups rose and paying accounts stayed flat, the extra people are being lost after signup. The problem is downstream.
- Do the accounts you hand-hold convert better than the ones who serve themselves? If people convert when you walk them through it and stall when you do not, the product needs you to work. That is downstream too.
If the answer to the first question is no, and no version of the page ever lifted signups, then the page may really be the problem. In that case it is worth reading what separates pages in the SaaS website conversion rate range. But if signups moved and revenue did not, stop editing the page. Every further rewrite is spending the team on the half of the funnel that is already working.
What a quarter of rewrites costs
Each row on the sheet took the team a few weeks of copy, design, review and launch. Four rows is most of a quarter. For a seed company, a quarter is a real share of the runway, and across all sectors the median gap from seed to Series A is around 616 days. Roughly 70 percent of seed companies never raise a Series A at all.
The money cost is quieter. Every one of those extra signups was paid for, in ads, content or the founder's own time, and the cost lands at signup whether or not the person ever sees a drafted quote. More signups into the same leak means each customer costs more to win and takes longer to pay back. The upside is large in the other direction. Measured on SaaS and product-led companies, activation of 20 to 40 percent is normal, and a ten point improvement in activation typically drives a 15 to 25 percent increase in free-to-paid conversion. That is the lever the sheet never pulled.
One week of product evidence, collected on purpose
You do not need a research project. You need one week where the evidence comes from inside the product instead of the page. Here is a plan for it.
- Monday. Sign up with a new email on a laptop that has never seen the product. Do not use the sample files on your desktop. Write down every screen in order, and the exact text on each button. Note the moment you would close the tab if this were not your company.
- Tuesday. Add the missing events between signup and first value. For the quoting tool that means Quotes viewed, Forwarding address copied, and First RFQ received. Three events are enough to show where people stop.
- Wednesday. Email ten people who signed up this month and never sent in an RFQ. Ask one question: what were you trying to do when you signed up? Read the answers as a list of situations, not as feedback on the design.
- Thursday. Change one thing below the button. In this example, the Quotes screen opens with a sample bracket drawing already quoted and one tolerance flagged in red, with a button under it that says Quote one of your drawings, which opens an upload box. Upload comes first, forwarding comes later.
- Friday. Add a new column to Things we tried called Where. Every past row gets the word Page. Thursday's change gets the word Product. Put the next three changes in the Product column before any new Page row is allowed.
If you want a fuller method for finding the exact step where people stop, our guide to fixing SaaS onboarding drop-offs counts users at each step and shows how to cut, defer or pre-fill the one that loses the most.
What to stop changing, and for how long
Freeze the landing page until the product change has had a full month of signups. If you change both at once, you will not know which one moved the number, and you will be back to guessing. The page is not finished forever. It is finished for now, because it already sends more people than the product can keep.
Then make the first screen do what the page promised. The question to ask is simple: what did this person come for, and how many steps stand between the signup button and that thing? In the Green Glass wine app, Scan leads from day one and everything else earns its place after it. The taste profile only opens up after three bottles. The thing a beginner came for, one bottle they will like, is the first thing they can do. The quoting tool needs the same order: a drafted quote first, the forwarding setup after the estimator has seen why it is worth asking the owner for.
For a self-serve product, the first run comes before the website. Studio Maydit is a web and product design studio for AI founders in the US, UK and Europe. A fixed-scope project runs three to four weeks and ends with a diagnosis of what is leaking in the product, so the next row on your sheet lands in the right column. If you have rewritten the page more than twice and the paying number has not moved, see what our product design work covers, or book a call and bring your Things we tried sheet.
