Situation pages for an advisory firm, and a test of what they change

What a services list cannot say

A homepage offering full-service planning for individuals, families and businesses is true, and it answers no one. The couple ten years from retirement, the owner who has had an offer for the company, the surgeon in the first years of practice and the person settling a parent's estate each ask a different question, and each question carries its own situation. A list of services has no sentence that names any of them.

A situation page names one of them in its heading and its first sentence, says what the firm does for people in it, and stops. That is the firm's unique selling proposition written as facts: an answer page in which the firm's name and the client's situation sit in the same few sentences. Whether an assistant uses it is not something the page controls, which is why this article ends with a test. The facts every page of the site carries are on the overview for financial advisors.

Four situations, and what changes from one to the next

Four kinds of situation follow, and each changes something beyond the heading. The table is a template: the brackets are the firm's own details, and a firm writes only the rows it serves.

The right-hand column is the easiest to skip and the most useful to write. A page that says whom it does not serve gives a reader, and any answer that quotes it, a boundary as well as an offer, and it sends the people the firm would turn away somewhere better before a first meeting is booked.

What each page states: the record, the process, the fee, the limits

The record, as counts

State the firm's record in counts it can back from its own files, with the date each one runs from: "Since [year] we have worked with [n] households in or near retirement; [n] of our [n] current client households are in this situation." A count describes the firm and says nothing about how any client fared, and it should not be written to suggest otherwise. The counts go to the compliance officer before they go live, like every other statement the firm makes about itself.

The process for this situation

Describe the steps as they run for this client, not the firm's process in general: how many meetings, over how long, who attends, what the client receives at each step, and what happens once the plan is in place. "[n] meetings over [n] weeks, with [name] at each" is a sentence a client can hold the firm to. "A tailored journey" is a sentence nobody can check.

The fee for it

Give the fee for this situation in the firm's own terms, [fee] for [scope], and any ongoing fee with what it covers, then link the fees page, where the full schedule and the minimum sit. The words about pay on a situation page match the fees page exactly. How that page is written, and the words that carry regulatory meaning in some places, are in stating how an advisory firm is paid.

Who it is not for

Say plainly whom the page does not serve, and describe where they might go by kind rather than by name: "If you need [the thing the firm does not do], a [kind of firm] handles that; we do not." A situation that brings a second service with it, such as disability cover for a young doctor, links to the page for that service, as disability cover questions describes.

Pages that should not exist

Write a page only for a situation where the firm could fill in every bracket above: a count that is not zero, a process with steps of its own, a fee, and a limit. A situation with nothing particular in any of the four is a line on the services page, and a page for it would be a longer version of that line.

Nor does one situation become several pages by swapping the town. Google's spam policies name doorway abuse, which includes pages aimed at particular regions or cities that funnel users to one page, and its guide to generative AI search says that separate content for every variation of how people search, made mainly to manipulate rankings or AI responses, breaks its scaled content abuse policy. One page per situation, with the places the firm serves stated on it, avoids both.

The test: one page, a baseline, and the same questions afterward

The only way to learn whether a situation page changed anything in the answers is to read them before and after, with nothing else changed for that situation in between. Write the test down before the page goes up, in a record like this one:

Situation:        [retirement income | business sale | profession | life event]
Page:             [address], published on [YYYY-MM-DD]
Questions:        [two or three questions a client in this situation might type],
                  each added to the panel as a new line on [YYYY-MM-DD]
Baseline:         [date], four assistants, three runs each
Recorded per run: named or not; what the answer says about the firm for this
                  situation; the sources it links
Other changes:    [anything else changed on the site or the listings, with dates]
Readings after:   [dates, monthly]
Result:           [what moved, what did not, what else changed]

1. Choose the situation and write two or three questions the way a client in it might type them, without the firm's name. What investors ask at each stage is one place to find them. Add them to the panel as new lines with a start date, so each has a history of its own.
2. Take the baseline before publishing, with every question asked the same way, as our methodology sets out for one assistant, Claude, and record what each answer says about the firm for this situation.
3. Publish the page, log the date, and change nothing else about how the firm describes this situation until the readings after it are in.
4. Read the same questions monthly, on the same assistants, the same way. Compare what the answers say as well as whether the firm is named: a fee, a process step or a limit from the new page, repeated in an answer, is the plainest sign the page was read. Open the sources each answer links.
5. Write the result down, other changes included. If the answers change on every assistant and hold for two readings running, and the page is the one thing that changed for the situation, that is as close to cause as this method gets.

The test proves less than it seems, whichever way it goes. Answers vary between runs, the assistants change, and other pages about the firm move in the same months, so a single reading says little and a pattern across several says more. Keeping the record is what lets the firm tell the difference, and how to keep a panel file is set out in setting up prompt monitoring.