Compliance first: We never take HIPAA/PHI data or personal customer data. Everything is de-identified before it leaves your control.
Home / Insights / Explainer
Explainer

What Is a Header Sample? Why It Gets You a Data Bid Faster

What Is a Header Sample? Why It Gets You a Data Bid Faster

A header sample is a short description of your data's structure: field names, record counts and date ranges, with no actual content. It is the fastest way for AI data buyers to judge fit and price, and it never exposes customer details.

Most data deals stall at the same point: the buyer wants to know what the data looks like, and the seller does not want to show the data. A header sample solves that. It describes the shape of a dataset precisely enough for a buyer to bid, without revealing a single email body, customer name or transaction detail.

By OpenDataPolicies Editorial · November 15, 2026 · 6 min read

The short definition

A header sample is the list of column or field names in each data source, plus a few facts about volume and time span. Think of it as the top row of a spreadsheet, plus a note on how many rows sit below it and what dates they cover.

For example, a support system's header sample might say:

  • Fields: ticket_id, created_at, updated_at, status, priority, category, channel, assignee_role, first_response_minutes, resolution_minutes, satisfaction_score, message_count, subject, body.
  • Record count: about 180,000 tickets.
  • Date range: March 2016 to last month.
  • Notes: subject and body are free text; customer names and emails would be removed in de-identification.

That is it. No rows of actual data. For a step-by-step walk-through by system, see our guide on how to create a header sample.

Why buyers want it

AI and machine-learning buyers have a short list of questions before they can put a number on any dataset. A header sample answers most of them at once:

  • What kind of data is it? Email, chat, tickets, CRM, projects, accounting, phone logs or industry software.
  • How much is there? Volume drives what a model can learn and whether a deal is worth the buyer's fixed costs.
  • How far back and how recent? Long, continuous history up to the present is ideal.
  • How structured is it? Fields like status, category and outcome act as natural labels.
  • How much free text? Text-heavy fields are where workflow knowledge lives, and also where de-identification work concentrates.
  • Is it unique? An industry-specific system with specialized fields signals data the buyer cannot get elsewhere.

These map directly onto how buyers set prices. See how buyers price data.

Why it gets you a bid faster

Without a header sample, the conversation goes in circles: the buyer asks for details, the seller asks what they need, someone schedules a call, someone else asks for a sample export, legal gets involved over what can be shared. Weeks go by.

With one, the buyer can evaluate fit on day one. According to founder Jeff Cline, bids typically arrive within 72 hours if the data is correct and a header sample is provided. Results vary; not a guarantee.

A header sample also makes competition possible. Because it contains no sensitive content, it can be shown to several prospective buyers at once. According to Jeff Cline, ODP seeks offers from multiple buyers so a seller can take the highest bid. Results vary; not a guarantee.

What a header sample does not contain

This is the part that makes owners comfortable. A header sample does not include:

  • Any actual email, chat, ticket or document content.
  • Customer names, contact details or account numbers.
  • Employee names.
  • Transaction amounts or individual records.
  • Anything from patient records or protected health information.

We never take HIPAA/PHI data or personal customer data. Everything is de-identified before it leaves your control. The header sample is simply a description of structure; de-identification happens later, before any actual records move, and only after you have accepted an offer. Our de-identification techniques explainer covers how that works.

One caution: field names themselves occasionally contain sensitive information, for example a custom field named after a client. Review the list and rename or drop anything like that.

What to include, system by system

A strong header sample covers each system you are offering. For each one, include:

  1. System name (for example Microsoft 365 / Outlook, Slack, HubSpot, Zendesk, Asana, QuickBooks).
  2. Object or table names (contacts, deals, tickets, tasks, invoices, jobs).
  3. Field names for each object, including custom fields, with obviously sensitive names cleaned up.
  4. Approximate counts in the system's native unit: emails, messages, records, tickets, tasks, transactions, call hours or GB of files.
  5. Oldest and newest record dates.
  6. Seats or users, which signals how many people's workflows are represented.
  7. Known gaps: a migration from another system, a year of missing data, a field that changed meaning.

Each of our tool pages explains where to find these counts inside that specific system.

Header sample examples by data category

Every category of operational data has its own typical fields. These illustrative examples show the level of detail that helps a buyer.

Email

Fields such as message_id, thread_id, sent_at, sender_role, recipient_count, has_attachment, subject and body; plus total messages, number of mailboxes and the date range. Sender and recipient addresses are not part of the sample and would be replaced with role tokens during de-identification. See email data.

CRM

Objects such as accounts, contacts, deals and activities. Deal fields such as created_at, stage, stage_changed_at, amount_band, close_date, outcome and loss_reason are particularly useful because they carry natural labels. See CRM data.

Projects and tasks

Fields such as task_id, project_type, created_at, due_date, completed_at, status, assignee_role, comment_count and description. The gap between due and done dates tells a model a lot about how work really flows. See projects data.

Accounting

Fields such as transaction_date, type, category, amount_band, payment_terms and paid_date. Exact amounts and counterparty names are typically generalized or removed. See accounting data.

Industry software

Vertical systems often have the richest fields: job type, dispatch time, arrival time, technician role and outcome in a field-service platform, or RFI status and response time in a construction platform. See industry software.

In each case, the header sample lists what exists; it never includes the values themselves.

Accuracy beats optimism

The fastest way to slow down a deal is an inaccurate header sample. If you say 2 million emails and the export shows 400,000, or you list fields that turn out to be empty, the buyer has to re-evaluate and trust takes a hit.

Some practical tips:

  • Pull counts from admin reports rather than guessing.
  • Say "approximately" when you are estimating, and give a range if unsure.
  • Flag fields that are rarely filled in.
  • Note systems you are not sure you have the rights to license, so they can be discussed rather than discovered. The data rights checklist helps here.

Remember the "if the data is correct" part of the 72-hour expectation. Accuracy is what makes speed possible.

From header sample to closed deal

Here is where the header sample sits in the overall process:

  1. Inventory your systems (data readiness checklist).
  2. Header sample for each system you want to offer.
  3. Bids from multiple buyers.
  4. Offer review: compare price, exclusivity, license term and permitted use. You can say no to any offer.
  5. Agreement: the purchase agreement sets payment terms and delivery.
  6. De-identification and delivery.
  7. Close. According to founder Jeff Cline, deals typically close in about 90 days. Results vary; not a guarantee.

Requesting a bid is free. If you have your counts handy, start with the valuation calculator for a rough range, then request a bid and attach your header sample.

How to prepare a header sample

  1. List your systems. Write down every system you might offer: email, chat, files, CRM, support, projects, accounting, phones and industry software.
  2. Export field names. For each object or table, copy the column or field names, including custom fields. Rename or drop any field whose name is itself sensitive.
  3. Add counts. Record approximate counts in the system's native unit, pulled from admin reports where possible.
  4. Add date ranges. Note the oldest and newest record dates and any known gaps or migrations.
  5. Submit with your bid request. Attach the header sample to a free bid request so buyers can evaluate fit immediately.
Our compliance promise: We never take HIPAA/PHI data or personal customer data. Everything is de-identified before it leaves your control.

Frequently asked questions

What is a header sample?

A description of a dataset's structure: field or column names, approximate record counts and date ranges, with no actual records. It lets buyers judge fit and price without seeing sensitive content.

Does a header sample include customer data?

No. It contains field names and counts only. We never take HIPAA/PHI data or personal customer data. Everything is de-identified before it leaves your control.

How fast can I get a bid with a header sample?

According to founder Jeff Cline, bids typically arrive within 72 hours if the data is correct and a header sample is provided. Results vary; not a guarantee.

What if I do not know my exact record counts?

Give your best estimate and say so. Each ODP tool page explains where to find counts in that system's admin reports.

Updated November 19, 2026.

OpenDataPolicies.org connects businesses with buyers of de-identified operational data. We are not a law firm, broker-dealer or investment adviser, and nothing on this site is legal, tax or financial advice. Statements about amounts, timing and outcomes are Jeff Cline's statements based on past transactions. Results vary; nothing here is a guarantee of any bid, price, buyer or closing. Every deal depends on the data, its quality, its rights and buyer demand at the time.