Opt-in Form Best Practices: Design That Converts
Twelve rules with the evidence behind each, and the two that contradict each other so you know which context decides.
On this page
The short version
- Twelve rules, and two of them contradict each other. Which one applies depends on whether the visitor arrived thirty seconds ago or has read three of your articles.
- The single largest lever is what the form offers, not how it looks. A form asking for an address in exchange for nothing specific cannot be styled into converting.
- Every field you add costs signups. Some earn it back in segmentation; most do not.
Opt-in form advice tends to arrive as a list of tactics with no indication of which matter and which are rounding errors. The rules below are ordered by effect size, which means the first four are worth acting on today and the last four are worth knowing about.
They also come with the evidence behind each, because "best practice" without a reason is how a rule survives long after the thing that justified it has changed.
Twelve rules, ordered by how much each moves
| Rule | Why | Effect |
|---|---|---|
| Offer something specific | An address is traded, not given | Large |
| Ask for the minimum | Each field costs completions | Large |
| Match the offer to the page | Relevance beats prominence | Large |
| State what happens next | Removes the main unspoken objection | Moderate |
| Button says what you get | "Submit" describes the system, not the reader | Moderate |
| One form per view | Two asks compete rather than add | Moderate |
| Show the frequency | Prevents an unsubscribe later | Moderate |
| Validate helpfully, not strictly | Over-strict rules reject real addresses | Moderate |
| Label above the field | Placeholder-only labels vanish on focus | Small |
| Give the button real size | Tap targets on a phone | Small |
| Privacy line near the button | Read at the moment of hesitation | Small |
| Autofocus only on a dedicated page | On a scrolling page it hijacks the view | Small |
The top three are about what the form is, not what it looks like. Most published lists start at row nine.
The two that contradict each other
Rule six says one form per view. Rule three says match the offer to the page. Those pull in opposite directions the moment a page could reasonably carry two different offers — an article about deliverability with both a general newsletter form and a deliverability checklist.
The resolution is not a compromise. It is that the visitor's state decides. Someone who arrived thirty seconds ago from a search result has no relationship with you and will take the specific, narrow offer if they take anything. Someone on their fourth article has already decided you are worth reading and is a candidate for the general one.
In practice that means one form, chosen by context, rather than two forms shown to everyone.
The same page, two visitors
First visit, from search
No relationship yet
- Offer the thing this page is about
- One field
- Say exactly what arrives and when
- No general newsletter pitch
Fourth visit, direct
Already reading you
- Offer the ongoing thing
- One field, possibly two
- Frequency stated, because it now matters
- The specific checklist is a distraction here
State what happens next
The unspoken question at every form is what this costs. Not money — attention, and the risk of a mailing list that turns out to be daily.
One line answers it: what arrives, how often, and that leaving is easy. "One email a fortnight. Unsubscribe in one click." That sentence does more than any styling decision on this page, and it is absent from most forms.
It also pays off later. A subscriber who was told the cadence is far less likely to treat the third email as a surprise, which is where unsubscribes and complaints come from.
Labels, and the placeholder mistake
Using the placeholder as the only label saves vertical space and costs completions. The label disappears the moment someone starts typing, so anyone who pauses, gets distracted, or comes back to check has nothing telling them what the field is.
It also fails for screen readers unless the field is separately labelled, and it usually is not. A visible label above the field costs one line and removes both problems.
Form review
- The offer is specific enough to name in four words
- Every field is justified by something you will actually do with it
- The button says what the visitor gets
- One line states what arrives and how often
- Labels are visible, not placeholders
- The privacy line sits next to the button
- It has been completed on a phone, by someone who did not build it
Frequently asked questions
Should the form ask for a name?
Only if you will use it in a way the subscriber would notice. A name collected for a greeting merge tag costs completions and buys a personalisation that most readers now register as automation.
Does form design matter at all, then?
It matters once the offer is right. A good offer in a bad form loses some conversions; a bad offer in a beautiful form loses almost all of them. Fix them in that order.
How many forms should a site have?
As many offers as you have, but only one visible per view. The number of distinct offers is the real question, and for most sites the honest answer is two or three.