Landing page built from your ad data and sales calls

Rebuilding a paid-traffic landing page from evidence moves conversion rate by 15 to 60 percent when the page it replaces was written from a design brief. The range is that wide because the size of the gain depends on how wrong the current page is, not on how good the new one is. If your page was already built from the ad data and the sales calls, expect single digits, and we will say so before you buy rather than after. What we read before anyone opens a design tool: Ninety days of the ad account at search-term and placement level, not the campaign summary. The last forty closed-lost reasons in the CRM, exported with the notes attached. Six to ten recorded sales or demo calls, listened to end to end. Whatever session recording, scroll depth and rage-click data you already collect. For industrial clients we also ask for the last three quotes you sent and the questions the buyer asked after receiving them, because the objection that kills an eighty-thousand-euro machine order is almost never the one the current homepage answers. The page is written before it is designed: You approve the argument as plain text — the claim, the order the objections are answered in, and the proof attached to each one — before any layout exists. Most of the disagreement in this work is about what the page says. Settling that in a document costs a day. Settling it in a design file costs three weeks and a rebuild. What we will not put on a page: A number without a source we can name. If a claim cannot be traced to your own data, a named customer, an audit, a certification body or a published study, the sentence does not ship, whoever on your side wants it. No countdown that resets on refresh, no logo wall assembled from companies that once downloaded a PDF, no testimonial that the named person has not read and approved. In crypto work this rule does most of the heavy lifting: we do not write pages around returns, yield or price expectation, and we do not build for presales. On testing: Detecting a 20 percent relative lift on a 3 percent baseline takes roughly 13,000 visitors per variant at 80 percent power. If the page receives 4,000 visitors a month, that test resolves in seven months, by which time your offer has changed. Below that volume we ship one considered page, instrument it properly and read behaviour, rather than running a test that cannot conclude and calling the result a decision. We will tell you which situation you are in on the first call. What you receive: The written argument and the research behind it. A designed page, built and live on your own stack, with real copy. A measurement plan naming the two events that decide whether it worked, implemented and verified with live submissions. A read of the first 90 days after launch, sent whether or not it flatters us. Not included: Ad creative and media buying. A second language. Offer and pricing strategy. CRM configuration beyond one form endpoint. Page edits after handover — the build goes to your team with its components documented, and we would rather train two of your people than become the queue every change waits in. Not for you if: You have a campaign launching in ten days. You want an existing design improved rather than the page rethought. You are selling something bought on impulse where the page is a shop window; that work is real, it is just not this.