Third-party cookies are on the way out, and with them the oldest shortcut in marketing: guessing what a visitor wants from where they have been. The replacement is older than the cookie itself, which is asking.
Zero-party data is what a visitor hands over on purpose: what they are looking for, when they need it, which size or plan fits, and how they want to be contacted. Nobody infers it, buys it, or has to clean it up later.
That is also why it converts. A visitor who has just said they need a gift for a teenager in the next two weeks is not a segment, they are a person with stated intent, and the next message can be about that alone.
What is zero-party data and how does it differ from first-party data?
Zero-party data is data a customer intentionally and proactively shares with a brand. First-party data is behaviour you observe, such as pages viewed or orders placed. Third-party data arrives from someone else's audience, and the three are not interchangeable.
| Type | Where it comes from | What it answers | Main weakness |
|---|---|---|---|
| Third-party | Data brokers, ad networks | Who they resemble | Deprecation, consent, accuracy |
| First-party, observed | Analytics, clicks, orders | What did they do here | Needs volume to say much |
| Zero-party, declared | Forms, popups, quizzes, preference centres | What do they want, and why | Only as good as the ask |
Observed data tells you a visitor looked at three pairs of running shoes. Declared data tells you they run trails twice a week and want a wide fit. The second can be acted on the moment it arrives, which is why the ask belongs near the moment of interest, not in a profile you plan to enrich for six months.
Which questions are worth asking a visitor?
One test settles it: if every possible answer leads to the same follow-up, the question is decoration.
- Intent: what are you trying to solve, how soon does it matter, is this for you or for someone else.
- Constraints: budget band, size, team size, deadline, compatibility.
- Preference: how often you want to hear from us, and about what.
Anything outside those buckets has to justify itself. Asking for a phone number when the sale happens by email costs completions and buys nothing.
Where should the ask live on your site?
Placement decides who answers.
- Exit-intent popup: best for a single intent question, asked as someone is about to leave.
- Quiz funnel: best when the answer needs context, because the result gives the visitor a reason to finish.
- Preference centre: best for frequency and topic choices, and the place every email footer should point to.
- Post-purchase survey: best for constraints you only learn after delivery.
- On-site poll: best for one low friction question that warms a cold visitor.
Start with one ask in one place: five popups competing for declared data collect less than one the site can act on.
How do you make the exchange feel fair?
The visitor pays with attention and personal information, so say what they get back. State the benefit in the same sentence as the question, show the answer being used, and always offer a way past, because a question with no skip turns a conversation into a checkpoint.
Never gate something the visitor already has. A size guide or a delivery estimate held hostage by a form shows up in your bounce rate long before it shows up in your reports.
How often can you ask before it becomes annoying?
Treat every ask as a budget. One question per visit, not one form with four fields. Never interrupt checkout or a form the visitor is already filling in, the pattern covered in popup frequency capping. Record the answer and stop asking for it: a visitor who has stated their goal twice will not state it a third time.
If you need several answers, spread the asks across visits instead of stacking them into one wall.
What do you do with zero-party data once you have it?
Unused answers are worse than unasked questions, because the visitor remembers telling you. A stated deadline moves someone into a short sequence, a stated constraint reorders what they see, and a stated preference sets the send frequency. The on-site half is covered in website personalisation: a declared fact should beat an inferred guess.
Keep the record small, dated and editable. Preferences change, and a preference centre that lets someone revise their answer is also a fresh piece of zero-party data.
How do you measure whether the capture works?
Four numbers, read together rather than one at a time.
- Answer rate: visitors who answered the ask, out of those who saw it.
- Completeness: answers that are usable, not blank or evasive.
- Action rate: how often the answer changed what the visitor saw or received.
- Downstream result: conversion or revenue per segment, compared with visitors who did not answer.
The last two decide whether the ask stays. A high answer rate that never changes a follow-up is a survey habit, not a data strategy.
What are the trust and compliance basics?
Say what you will do with the answer, then do that. Marketing consent is separate from storing the answer, and a preference centre needs a real deletion path. Ask for what the next step needs, not for what might be interesting in three years: the less you hold, the less you have to explain.
Underneath the compliance layer sits something simpler. Zero-party data works because the visitor chose to tell you. The moment the choice is fake, the data is worthless.
FAQ
Q: Is zero-party data the same as first-party data? A: No. First-party data is behaviour you observe on your own properties. Zero-party data is what a visitor deliberately tells you, such as a goal or a preference.
Q: How many questions should a capture popup ask? A: One to three. Every extra field lowers the answer rate, so keep the essential question and collect the rest across later visits.
Q: Can zero-party data replace third-party targeting? A: It replaces the intent signal, not the reach. You get better answers about your own visitors and no audience from anywhere else.
Q: How do I know the data is any good? A: The action rate is the honest test. If answers never change what the visitor sees or receives, the quality question stays academic.
Start with one question that changes the next message, ask it where intent is highest, and act on the answer the same day. heycustomer.byako.dev gives you the popup, the quiz and the segments to put those answers to work.