Most redesigns begin with a moodboard. Ours begin with a sentence. It is the sentence you would say about your business if you had ten seconds and no slides. That is what positioning before design means. It is also the first two weeks of every engagement we run. This post explains the full reason for that order. It covers what the sentence has to contain, the tests it has to pass, why you cannot write it after the layout, and how we find it.
It is also the standard behind every teardown we publish. When we score a homepage weak on the sentence, this is the standard we are scoring against.
The site is downstream of the story
When a homepage feels vague, the problem is rarely the design. The real problem is that nobody decided what the business sells. Good layout makes a clear sentence clearer. It cannot make a vague sentence clear. Put a strong grid under a weak sentence and you get a good-looking page. The buyer still cannot repeat what it says to their co-founder.
We see this most often on B2B SaaS homepages, meaning software sold to other businesses, where the site has grown along with the product. The company launched with one feature and then added six. Now the top of the page says "the platform for modern teams" because nobody wanted to leave anything out. The Pickcel repositioning started in exactly that place. The old site described screens and software as a list of features. The buyer was really purchasing an operating layer for the in-store experience. Nothing about the old site was badly designed. It accurately described an older version of the company.
If you cannot state what you sell in one line, the design team cannot state it in a full screen either.
The reverse is also true, and that is why the order matters. A page built on a sentence that holds can be plain and still convert. Every section has a clear job, so every section can be checked. The design has a smaller job to do, and it does that job better.
What the sentence has to contain
A working sentence carries three things and nothing else. It says who the product is for, what changes for them, and what kind of thing it is. Take this example: "Digital signage software for retail chains that want every screen updated from one place." It names the buyer (retail chains). It names the change (every screen updated from one place). It names the category (digital signage software). The line is not clever. A buyer can check it and repeat it.
Compare that with the lines we see most often on B2B SaaS homepages. "Better decisions, faster." "The platform for modern teams." Each one is a vague promise in the place where a claim should be. Neither names a buyer, so neither can be wrong. That is also why neither of them works. The H1 is the main headline on the page, the large line a visitor reads first. The scorecard marks the sentence weak when the H1 is an adjective with no buyer and no category. That single check catches more homepages than any other.
Two of the three parts are usually missing for the same reason. Naming a buyer excludes someone. Naming a change means someone can prove you wrong. Both feel like risks. Both are what make the line work. A sentence that cannot exclude anyone is a description of the category, and your buyer already knows what the category is.
The engineering studio version has the same shape. Prodios Labs builds systems for ministries and state agencies and needed to sell to private companies. The line that carried the site was "We build the systems states run on". It names the category (systems, built) and the proof (states run on them). It also implies the buyer: anyone who needs software that cannot fall over. It does not say "digital transformation partner", which is what the competitors say. That is why it works.
The three tests every line has to pass
We run every candidate sentence through the same three checks. They take five minutes and they catch almost everything.
The co-founder test. Read the line to someone once. Then ask them to tell a colleague what the company does. If they cannot, the line is not carrying the meaning. If what they say is very different from what you meant, the line is not carrying the meaning either. Buyers forward homepages, so the sentence has to survive being forwarded.
The competitor test. Could your nearest competitor put the same line on their site without changing a word? If yes, the line describes the category and not the company. "AI-powered sales assistant" fails this test for every AI sales tool at once. The fix is usually to add the buyer or the change. Those are the parts a competitor cannot copy without also copying your strategy.
The search test. Open the page source and read the meta title. The meta title is the title line that appears in a Google result. Then read the H1. If the meta title is a better sentence than the H1, someone on your team already wrote the positioning. They put it where only Google reads it. This happens more often than you would expect. In our SuperOps teardown the meta title said "Autonomous IT operations platform for MSP and IT teams". The page itself led with a slogan about the age of AI. The sentence existed, but it was in the wrong tag.
Why the sentence cannot come after the design
The usual counter-argument is that positioning and design can run at the same time. A related version is that the copy can be "dropped in" once the layout is approved. In our experience that leads to two redesigns instead of one. The reason is that the sentence decides four things the design depends on.
The proof stack. The proof stack is the set of logos, case studies and numbers that back up your claim. Once the sentence names a buyer and a change, the proof has to support that change for that buyer. A logo wall of companies in the wrong sector stops being proof. A case study with the wrong metric stops being proof. You cannot design the proof section until you know what it has to prove.
The section order. A homepage makes a case for the product one section at a time, and the order matters. If the sentence claims a change, the next section has to show that change. If the sentence names a buyer, the page has to speak to that buyer before it speaks to anyone else. Our guide to the B2B SaaS homepage walks through the order section by section. Every section in it is placed relative to the sentence.
The lead form. When the sentence names the buyer, the form can ask what the buyer sells instead of only asking for an email. The wrong buyers then stop filling it in. A form designed before the sentence is a generic form.
The pricing frame. A pricing page is only a table until the sentence tells you what the buyer compares the price against. That comparison is the frame. Examples are "what you would pay to do this yourself" or "what the delay costs per month". The frame comes from the change the sentence promises.
Design any of these first and the sentence will contradict them when it arrives. The team then has two choices. Either they rewrite the design, or they water down the sentence to fit it. The second choice is more common. That is how homepages end up saying "the platform for modern teams".
How we find the line
The method is short and it does not look impressive. It is also the part most agencies skip, because it does not produce a deliverable that looks like design.
- Strip the deck. We take your sales deck, your current site and any positioning document. We cut every sentence a competitor could also say. What survives is usually two or three paragraphs. The outcome your customer is really buying is somewhere in them. Look past the feature list and the category name. Look for the thing your customer would describe to a peer over coffee.
- Name the moment they reach out. Find the event that made them reach out, rather than describing who they are in general. It might be a fundraise, a sales hire, a redesign that stalled, a competitor that just raised, or an audit that failed. The sentence should speak to that moment. That is the state your buyer is in when they land on the page.
- Read the last ten sales calls. Use recordings if they exist and notes if they do not. Buyers describe what they are buying in their own words. Those words are almost always better than the marketing copy. The final line usually contains one phrase lifted straight from a call.
- Write three candidates. Each one names the buyer, the category and the change. Each one puts the emphasis in a different place. One leads with the buyer, one with the change, and one with the category. The differences make the choice a real choice.
- Test and recommend. Every candidate goes through the three tests above. Then it goes to you with one recommendation and the reasoning behind it. You sign off the line in writing before anyone opens a design tool.
That is week one and part of week two. Only once the line holds do we touch type, grid or color. The result is a site that says one clear thing, because the team decided what that thing was before the design started.
The objection we hear most
"Our buyers already know what we do." Sometimes that is true. It still does not change the answer. The buyers who already know are not the ones the homepage is for. The homepage is read by the person who arrived from a search, a referral, a cold email or a teardown. It is also read by the colleague they forward it to. Both are strangers. The sentence is for them.
The second objection is that a specific sentence will lose deals outside the named buyer. That is true. Those are the deals that would have cost the most to close and the most to serve. When the sentence names the buyer, the wrong buyers stop writing. The right ones write with their problem already described. That is the cheapest qualification step a company can build. The cost is one line of copy.
What changes when the sentence comes first
The page order changes. Proof moves above the ask. The form asks what the buyer sells. The pricing page gets a frame instead of a plain table. None of these are design decisions. They follow from a decision made in week one. That is why our SaaS website design engagements run positioning and build as one team instead of two vendors.
The inquiries change. You get fewer of them, and they are more specific. More of them are close to a scoped conversation on the first reply. Once you can see the inbox change, you stop asking whether the redesign worked.
The next redesign gets cheaper. A site built on a clear sentence can be refreshed one section at a time. The argument still holds, so only the evidence needs updating. A site built on a moodboard has to be rebuilt from the top every time the product changes.
Where to start
Run your homepage through the scorecard. It takes a minute. It will tell you whether the sentence is the weak point, or whether the problem is further down the page in the proof or the path. If the sentence scores weak, read the guide for what the first screen has to do. Then read the Mutiny teardown to see what happens when a well-funded company skips this step.
If you want the line written for you, that is the first two weeks of a positioning engagement. The sentence, with three candidates and the reasoning, is the first thing you sign off. The design starts the day after.