The fixed-fee ledger: build notes from a family law site

The ledger: three cells between two hairlines

The band sits immediately below the hero, and in the stylesheet it is commented as exactly what it looks like, a hairline ledger. A one-pixel rule above, a one-pixel rule below, three equal columns divided by lighter rules, no box, no shadow, no card. Below 760 pixels wide the three columns stack into one and the dividing rules move from the sides to between the rows, so the ledger reads the same on a phone at midnight as it does on a desktop.

The three cells say, in order: $250: first meeting, one hour, fixed, credited to your file if you retain us. In writing: every fee estimated before the work, revised in writing if scope changes. 72% settle: most matters resolve by agreement or mediation, not a courtroom.

Two of those three are not numbers. That was deliberate. A price on its own invites the question “and then what?”, which is the question that stops a frightened person from booking. The middle cell answers it before it is asked, and the third one answers the fear underneath it, which is not really about money at all.

The same figure appears again inside the primary button: Book a first meeting — $250 flat. A reader who never scrolls past the headline has still been told what it costs.

Paper, ink and pine: a palette with one dark moment

The whole page runs on four colours, declared once at the top of the stylesheet: a warm paper white, a near-black ink that is brown rather than grey, a soft grey for secondary text, and a deep pine green. Nothing else. There is no accent colour for “calls to action”, because the buttons are already pine and the page already has only one action.

Pine appears in small amounts almost everywhere: the eyebrow labels, the buttons, the italic phrase in the headline, the step numbers, the little dot next to the privacy line. It is used at full strength as a background exactly once, on the band about how the firm works. That is the only dark moment on the page, and it is placed there so the promises land with weight, and so the reader gets a change of air roughly two-thirds of the way down without a single decorative element being added.

Type is two faces: an editorial serif for headlines, the fee figures and the lawyers’ names, and a plain grotesque for everything else. The serif is set at unusually light weights; the headline sits around 355 where a default would be 400, which is what keeps a 48-pixel-tall sentence from shouting. Big and quiet at the same time is a weight decision, not a size decision.

Writing for someone who has not decided anything

The build carries its five-question contract in a comment at the top of the stylesheet, and the audience line is specific: the reader is lying awake at 1am, quietly researching a separation. Every word on the page is chosen against that sentence. The headline is “You don’t need to have decided anything to talk to someone”, which removes the commitment the reader is afraid of rather than asking for it.

Under the hero buttons there is a single grey line: this page does not track you, and nothing here creates a record of your visit. Someone researching a separation is often doing it on a shared computer. No stock photograph addresses that; one sentence does. The fine print at the bottom of the page does the same job for a different fear: everything you say is privileged, including the first meeting.

The steps section is numbered, and the comment in the stylesheet says why: it is a real sequence, so numbering earns its place. Numbers on a list of features are decoration. Numbers on “what happens next” are the whole point, because the reason this feels impossible is that nobody tells you the order of things. Step three ends with “there is no follow-up sequence and no one calls you”, which is a design decision as much as a copy one: the page is built to be left.

Writing for people in the worst week of their year is a thread across several of these builds. The emergency version of the same problem is in our notes on designing for the 2am call, where the reader has water coming through a ceiling and forty seconds of patience.

The ban list: what was deliberately left out

Every build here starts with a named ban list, and for a law firm the bans write themselves. There is no gavel, no scales, no marble column, no Corinthian anything, no skyline at dusk, no handshake, no photograph of a leather-bound book spine. Open the file and search for them: they are not there.

The favicon is the test of whether that ban held, because it is the one place a legal cliché is easiest to smuggle in. Alderwood’s is drawn inline in the page: a vertical stroke with a shallow arc across it, pine on paper. A stem and a branch, or an arch, depending on what you bring to it. Not a gavel.

There are no stock photographs on the page at all. The two portrait spaces in the lawyers section are toned gradients, and the stylesheet comment states their purpose plainly: a real photograph replaces them without any layout change. A concept build should not pretend to have assets it has not been given, and it should not be built so that receiving them breaks it.

The practice areas are set as two columns of quiet prose with a rule under each heading, not cards, not tiles, not icons. Four services do not need four boxes. And nothing on the page moves unless the reader has asked for motion: the reveal script checks the reduced-motion setting first and returns immediately if it is on, and the stylesheet flattens every transition to a thousandth of a second as a second line of defence.

The honest part: every number on that page is fiction

The $250, the 72%, the two bar-call years under the lawyers’ names: all invented, because the firm is invented. The page says so itself, in a badge fixed to the bottom-left corner on every screen size: a concept build by Droom, this firm isn’t real. The file also carries a noindex instruction, so it is not competing with any actual firm for any actual search.

That matters more than it might sound, because when we build for a real business, a number like that cannot be written by us at all. Our validation stage scans the finished page for exactly these claim patterns (founding years, tenure, review counts, awards, prices) and fails the build when it finds one that did not come from the client’s own material. A demo containing a plausible lie loses the only reply you were going to get, and a live site containing one is worse.

So on a real family law build, the ledger has the same three cells and the same rules around it, and the firm supplies the three answers. The design holds either way. What the design cannot do is guess.

What this would cost, and what you would get

Alderwood is a compact build, and a site of this kind sits in our $1,500 tier, which covers up to five pages. The $2,900 tier adds the rest of the pages a firm actually needs, a page per practice area written to be found, and $4,900 covers a larger site with booking, intake and content structure. Ongoing care is $99 a month, or $349 with local search handled too; nothing is bundled into a price you cannot see, which would be a strange way to run a business whose demo argues for stating the fee.

If you are a family lawyer reading this, the version of this argument written for your practice rather than for designers is on our page about websites for family law firms. The other builds, each with its own decisions and its own ban list, are on the concept builds shelf.

One last thing, unrelated to design and more important than all of it: whoever builds your site, make sure the domain is registered to you. We wrote the five-minute version of that check in who owns your domain. A firm that cannot recover its own web address has a bigger problem than its typeface.