Value First

For a long time the order was fixed. A page explains the product, anyone interested creates an account, and then the product appears. Receiving before giving — and handing an email address and a company name to a service you have never used is a bigger decision than it looks.

The numbers show exactly where that costs. If 40,000 visits produce 5,000 people who reach the first screen of the signup flow and 1,200 who actually create an account, the signup screen converts at 24 percent. Even after the browsers have been filtered out, three of every four people who arrived intending to create an account walk away. The friction here isn't only the number of input fields. The request itself is the obstacle: hand something over before you know what you get back.

So the order got reversed. Instead of explain, collect, reveal, it became reveal, let them build, then collect. It's the difference between tasting something at a market and buying it from a description, and it costs the seller one sample to find out who was serious. The same request weighs differently depending on whether the person is empty-handed or already holding something they made.

Only one thing changes: where in the sequence the request sits. Not at the front, but at the moment the user tries to save or share what they just made. Five decisions follow from that.

1. Pick one slice that can be done without any request. Writing one document, converting one image, calculating one quote — something that completes on its own. The test for the slice is whether finishing it teaches the user what the product is worth.
2. Hold the work without a login. With no account there is nowhere on the server to file it, so the state gets kept in the browser and attached to an account later. This is where most of the engineering effort in the design goes, and it is also where it most often breaks — a closed tab or a cleared cache takes the work with it.
3. Attach the request to the result. Saving, sharing, and downloading — actions that protect the thing or send it outward — are the natural spots. A modal that appears the moment the page loads has simply moved the request back to the front.
4. Preserve what they made. The work has to continue exactly where it was the instant the account exists. A flow that finishes signup and lands on an empty screen is the worst case here, because it demonstrates value and then takes it away.
5. Reduce the request to one field. Ask for an email and nothing else, and collect team size or role later through progressive profiling while the product is being used.

One caution belongs with this. Signups can go down after the change, because someone who accomplished what they came for in one pass leaves without making an account. So the decision doesn't get judged on signup count but on paid conversion per visit. If signups fall from 1,200 to 900 while teams actively using the product in their first week rise from 340 to 520, the 300 who disappeared were never going to stay.

The design lands in three places.

1. The first screen. Instead of a button requesting a trial, the input or the workspace itself is the first thing on the page. It shortens the pitch, so the decision usually arrives together with a rewrite of the landing page.
2. Tool-shaped products. Converters, calculators and generators — anything where one pass produces a result — fit best. Put the result on screen and say that saving it needs an account. Products with a longer setup can still borrow the idea by shipping a prefilled example instead of an empty workspace.
3. The point of payment. The same principle applies to money: raise price after the value is demonstrated. The moment someone wants a high-resolution download of what they just made is that point.

Other devices in this group work alongside it rather than instead of it. Social proof persuades with the signal that other people use it and urgency pushes with time, while this one persuades with something the user built. All three often sit on the same screen.

Some products can't do it. When producing any result requires connecting the customer's own data first, or when regulation puts identity verification ahead of everything else, this ordering cannot be built. A costly computation per attempt makes it expensive rather than impossible, which usually turns into a limit on how many free passes one visitor gets.

"Let people write one document without logging in, and ask for signup when they hit share."

Moving the request from the first screen to the far side of the result.

"Signups are down but active teams in week one are up. We're keeping it."

Judging the change on teams that actually use the product rather than on account count.

"For this product, an input box on the first screen beats a trial button."

Replacing the screen that explains with the screen that does.

Pagedeck's marketing page is the document editor itself. A first-time visitor can write a page immediately, and the email question appears when they press share. When Noa built that screen, the explanatory copy came down to three lines and the workspace took the space it left.

The longest part was the draft-holding Theo owns. If the document a visitor was writing disappears while the account is being created, the whole design runs backward: the product proved it was useful and then charged for the proof.

The first assumption is that opening more converts better. The goal is demonstrating value, so more of it should help. But if everything is usable without any request, there is no reason to create an account at all. The request has to sit at the first moment the result needs to persist. One page free, and an account required to save that page or show it to a colleague, is the shape that works. Set that line wrong and visitors leave satisfied while no accounts remain.

The second is reading a drop in signups as failure. It's the easiest number to see and the easiest to compare week over week. But this design deliberately filters people before the signup step. Someone who tried it, found it wasn't for them, and left would previously have created an account and vanished, so much of the decline isn't a loss. The values to check are paid conversion per visit and the first-week conversion rate after signup. When those two rise together, fewer signups is not a problem.