Zero-Party Data
Zero-Party Data
What is zero-party data?
- Preference and intent data customers share themselves
- The key is data customers volunteer, not data you observe
- Accurate raw material for consented personalization
Once you start using zero-party data in earnest, your marketing team’s workflow flips. Analysis used to come first: collect logs and clicks, then infer that “this person is probably interested in this topic.” Now designing what to ask, and on which screen, comes first. A single checkbox on a subscription form or one dropdown on a seminar registration becomes a data field, so whoever builds the landing page effectively becomes a data designer. The trade-off is that you can never know what you didn’t ask, but the answers you do collect stay on file as facts you can use without further interpretation.
In practice, responses usually fall into three groups: preferences, such as topics of interest; intent, such as adoption timing and budget range; and context, such as job function and the number of people to be trained. The line between this and first-party data is whether the input was required to complete a transaction. A name and company entered only to download a resource are hard to call zero-party data, while a topic someone chose when they didn’t have to is zero-party data. Keep it to three questions or fewer, and show what the customer gets in return on the same screen, or responses will drop off. B2B interests also shift with reorganizations and budget cycles, so treat any answer as stale after 6 to 12 months.
The most common misconception is that simply running a survey builds up zero-party data. Problems actually arise when you collect it and don’t use it. If a customer clearly picked “AI at work” and the following week gets the same mass email as everyone else, the more carefully they answered, the more betrayed they feel, and they unsubscribe. If you can’t show what changes right after a response, it is better not to ask at all. Also, a stated interest is not purchase intent. If you follow an “I’m interested” response with an immediate sales call, customers start avoiding your calls and response rates on later surveys fall too. Using data beyond the purpose you stated at collection is also overstepping the consent you were given.
John KimInstructor John Kim, author of B2B Marketing Design and CEO of Nanum Consulting
Jaryong KooMarketing PhD and a leading authority on AI, data analysis, and personal brandingThe term was coined by the research firm Forrester to separate data customers “voluntarily provide” from broader first-party data, and it spread alongside the shift toward privacy-centered marketing.
Consider a hypothetical case. Company A, a B2B software company with 210 employees, had 14,000 newsletter subscribers. But its open rate stayed at 14%, and the newsletter generated only 3 training and consulting inquiries a month. The marketing team decided to drop “inferred tagging” and ask customers directly, on a six-week timeline.
- Kim, the marketing team lead, added one line and five checkboxes to the bottom of the newsletter: “Pick up to three topics and we’ll send you our case studies on them first.”
- Over two weeks, 1,180 of 14,000 subscribers (8.4%) responded. Interest split clearly: 470 chose “Applying AI at Work” and 390 chose “B2B Sales Training.”
- When the same question went to non-responders a second time, only 2% answered. Kim changed the wording to “Get our 30-page case study instantly” and resent it.
- Lee, who runs CRM, stored the responses in fields separate from inferred tags and recorded the source and date with each one, such as “self-reported, 12 Mar,” so refresh timing could be managed.
- The following month, they stopped the all-subscriber send and split it into two topic-based newsletters. A training manager at one client company replied, "I opened it because it was a topic I chose."
- After three months, open rates rose from 14% to 31%, monthly sales inquiries grew from 3 to 11, and the unsubscribe rate fell from 0.9% to 0.4%.