Begin with the decision the form must support
A marketing form is a small service interaction, not just a box at the end of a landing page. Before choosing fields, write down what a person is trying to accomplish and what the organization needs in order to respond. For a hypothetical request for a website consultation, a name, reply address and short project question might be enough to start; budget, phone number and a long qualification questionnaire may not be needed at that point. This is a design exercise, not a report about a real client or observed user behavior.
The useful test is whether each field has a stated purpose and whether the person can understand that purpose before answering. W3C’s WCAG 2.2 guidance says inputs need labels or instructions that identify the expected information. It also warns that too much instruction can create confusion, so the goal is enough context to complete the task without unnecessary detail. [1] In practice, a plain label such as “Work email” can be paired with a short explanation of how the address will be used, if that affects someone’s choice to submit.
Make the expected answer visible before submission
A placeholder is a poor place to hide the only instruction: it disappears when a person types and may not remain available as a persistent label. Put the field name where it remains visible, then add a hint only when the answer has a non-obvious format or rule. If a project reference number must contain a prefix, show an example. If a question accepts free text, do not imply a rigid format that the system does not require. W3C distinguishes the presence of a label or instruction from whether it is clearly worded and correctly associated with the control; teams need to check both the visible content and the implemented relationship. [1]
Required and optional fields should be identifiable in text, not only by color or an asterisk whose meaning is left unexplained. Group related choices under a concise question, and label each radio option or checkbox in language that makes the selection clear. These are accessibility and content-structure decisions as well as visual design decisions. The exact implementation still needs review against the applicable standard and the site’s assistive-technology support.
Treat validation as a repair path
An error state should answer three questions: what went wrong, where it happened and what the person can do next. GOV.UK Design System guidance recommends an error message beside the affected field and an error summary at the top when a form has validation errors. It advises retaining the answers already entered so people can correct a problem without starting over. [2] Its validation pattern also recommends waiting until the person tries to continue before showing most errors, rather than interrupting every keystroke. [3]
For the hypothetical consultation request, “Enter an email address in a format like name@example.com” gives a repair direction; “Invalid” does not. If an optional phone field is empty, the system should not reject it. If a service cannot use a supplied value, explain the constraint and preserve the rest of the form. A service outage or eligibility decision is not a data-entry mistake, so it should have its own explanation and next step rather than being disguised as field validation. These are practical interpretations of the cited guidance, not a promise that one error pattern fits every service.
Keep the form and the promise in agreement
The call to action should describe what happens after submission. “Request a consultation” is more informative than “Submit” when a person is asking for a conversation; a confirmation state can then say what was received and what response to expect, provided the business can actually honor that expectation. Avoid urgency language or privacy assurances that the organization has not verified. The page should also make the route back to revise information obvious before sending.
A compact design review can trace one fictional request from the landing-page promise to each field, its help text, its error message and the confirmation. Ask a reviewer to complete the form with a normal answer, omit a required answer, enter an ambiguous value and correct it. Check keyboard order, visible focus, label association, error announcement and whether earlier answers remain. Record issues and retest after changes. This exercise can expose mismatches between copy, interface and validation logic; it does not establish conversion lift or substitute for testing with representative users.
A practical lesson for marketing teams
Design the form around the first useful exchange, then justify each additional question against a real operational need. Put labels and necessary instructions in view, make requirements understandable, and build an error path that helps the person continue. Finally, test the complete journey with the actual implementation and confirm that the organization’s follow-up matches the promise on the page. A shorter form is not automatically better, and a well-styled form is not evidence of better performance. The quality of the experience depends on whether the requested information, the stated purpose and the next action fit together.
Sources & useful references
- W3C: Understanding WCAG 2.2 Success Criterion 3.3.2, Labels or Instructions
- GOV.UK Design System: Error message component
- GOV.UK Design System: Recover from validation errors
Results are attributed to the original publisher. Proposed tests and creative recommendations are our analysis.