How to write a data listing that buyers trust

A buyer arriving on your listing has one question: is this the data I need, and can I trust it? They usually decide within a minute whether to contact you or move on. A trustworthy listing answers practical questions before they are asked, shows evidence instead of adjectives, and is honest about its limits.

This playbook covers each part of a listing, with examples of what works and a checklist to run before you publish.

What buyers check first

Most buyers scan in roughly this order:

1. Title and summary: is this the right kind of data for my problem?
2. Coverage: does it include my countries, sectors and time period?
3. Schema and sample: are the fields I need there, and do they look usable?
4. Freshness and delivery: how often is it updated, and can I get it into my stack?
5. Price and licence: can I afford it, and can I use it the way I plan to?

If any of these is missing, the buyer either leaves or sends a message asking for it, which slows your sale. Write the listing so a buyer could shortlist you without talking to anyone.

Write a title that says exactly what the data is

A good title names the subject, the data type, the geography and, where it matters, the granularity or update frequency. Compare:

• Weak: "Premium global business intelligence"
• Strong: "German company registry data with officers and filings (daily)"
• Weak: "Location insights"
• Strong: "Aggregated foot traffic for retail places in Spain and Portugal"

Specific titles match the words buyers type into search and set expectations correctly. Avoid superlatives such as "best" or "most accurate": they cannot be checked, and they make careful buyers suspicious.

Make the summary answer "is this for me?"

Use one or two sentences covering what the dataset contains, how much of it there is and how often it changes. For example: "Every ocean import manifest into the United States since 2010, with shipper, consignee, HS code and ports, updated daily." A buyer can tell immediately whether to keep reading.

Describe coverage honestly

Geography, history and volume

State the countries or regions covered, the year your history starts and how many records or entities the dataset contains. If coverage differs by country, say so. A list of countries with notes on depth is more useful than a world map with everything shaded.

Known gaps

Every dataset has gaps: missing fields for small companies, countries with a longer update lag, a period when a source changed format. Naming them shows you know your data. It also prevents the most damaging kind of complaint, the one where a buyer discovers a gap after signing.

Publish a data dictionary

List every field with its name, type and a plain-language description. Add an example value where the meaning is not obvious, and note the fill rate for important fields if you measure it. Buyers' data teams read this section closely because it tells them how much work integration will take.

A short dictionary might look like this:

Field	Type	Description
company_id	string	Stable identifier that persists across updates
legal_name	string	Name as registered with the official registry
status	string	Active, dissolved or in liquidation
incorporated_on	date	Date of incorporation

Explain where the data comes from

Buyers, and their legal teams, want to know the source and how you obtained it. Cover:

• Source: official registries, public websites, partner feeds, panels, surveys or your own sensors
• Collection method: licensed bulk feed, automated collection, manual research or survey fieldwork
• Processing: cleaning, normalization, entity matching, de-identification or aggregation
• Legal basis: the rules or agreements that allow you to license the data, and any consent involved

For sensitive categories such as location, transaction or health-related data, explain how the data is aggregated or de-identified and what minimum thresholds you apply. DataStoa does not accept data covered by its prohibited data policy, and explaining clearly why your data is compliant saves a round of questions.

Show quality with numbers you can defend

"High quality" means nothing on its own. Useful quality statements say what was measured and how:

• Fill rate: the share of records in which a field is populated
• Accuracy: checked against what ground truth, on what sample, and when
• Match rate: for enrichment products, the share of a typical customer file you can match
• Latency: the time between a real-world event and its appearance in your data

Only publish figures you can explain if a buyer asks. A modest number with a clear method earns more trust than a perfect score with none.

Add a sample that represents the real file

A sample is the strongest trust signal on a listing. Use the same schema and format as the production delivery, include typical rows rather than only your best ones, and mask or replace any personal data. A few hundred representative rows usually tell a buyer more than a polished slide deck. If you cannot share real rows, explain why and offer a call to walk through the data.

Be specific about delivery, licence and price

• Delivery: list the methods you support, such as SFTP, cloud storage, a data share or an API, and the file formats.
• Update frequency: daily, weekly, monthly or on request, and the typical lag.
• Licence: which uses are allowed, such as internal analytics, use in commercial products, redistribution or AI training.
• Price: a starting price and what changes it. Our guide to pricing a data product covers this in detail.

A pre-publish checklist

Before you publish, check that your listing:

• Has a specific title with subject, data type and geography
• States coverage, history start year and volume
• Names known gaps and differences in coverage
• Includes a complete data dictionary
• Explains source, collection method and legal basis
• Gives quality metrics with a short note on method
• Has a representative, privacy-safe sample
• Lists delivery methods, formats and update frequency
• States usage rights and a starting price
• Has been read once by someone outside your team

If you are new to selling on DataStoa, start from the seller overview, then compare your draft with other listings in your category using search.

Frequently asked questions

How long should a listing description be?

Long enough to answer the buyer's practical questions, usually a few hundred words plus the data dictionary. Use headings so buyers can jump to coverage, methodology or delivery. Cut marketing language before you cut facts.

Can I list a dataset without a sample?

You can, but expect more questions and slower decisions. If the data cannot be shared, describe the schema in detail, publish quality metrics with methods, and offer a walkthrough call or a sample under a confidentiality agreement.

Should I mention my data's weaknesses?

Yes. Buyers will find gaps during evaluation anyway, and naming them first shows that you know your data. It also reduces disputes and churn after the deal.

How often should I update my listing?

Whenever coverage, schema, delivery or pricing changes, and at least every few months to refresh volume figures and samples. A listing that says "updated daily" but shows a year-old sample undermines its own claim.