004Field Note
Give Your Comparison Table Real Row and Column Headers
A practical HTML example for preserving feature and plan context in an accessible product comparison table.
Comparison table semantics
A grid needs relationships, not just bold labels
The relationship is inferred
- Bold cells can still be data cells
- A symbol needs its plan and feature context
- A screenshot cannot retain HTML associations
The relationship is explicit
- Use th cells for actual headers
- Name row and column scope
- Identify the table with a useful caption
A comparison table should preserve the relationship between each answer, its feature and its plan. Bold text and a neat grid are not enough to express those relationships in HTML. For a simple table with both row and column headings, use actual header cells and declare whether each applies to its row or column. Give the table a useful name so someone navigating directly to it knows which comparison they have reached.
This is an accessibility implementation guide for public product-comparison pages. The example is fictional and the guidance comes from W3C Web Accessibility Initiative tutorials. It does not report a screen-reader study, establish full WCAG conformance or promise an increase in AI citations. Better table semantics are a concrete reader-facing improvement; treating them as a ranking trick would claim a result this research did not measure.
A check mark needs a row and a column
Imagine a fictional software comparison with two plans, Starter and Team. One row says whether bulk export is available. Another gives the support channel. A sighted reader can scan the headings and infer what a check mark means, provided the layout stays intact. A developer might build the same picture entirely from ordinary data cells, then make the first row and column bold with CSS.
The comparison-page decision record covers claim verification; this guide addresses the table’s structure. The visual styling does not turn those cells into semantic headers. The document needs to identify which cells supply context. It also needs language that does not leave the meaning of a symbol to color or a separate legend. In the example below, the answers use words: “Not included” and “Included.” The terms are invented for this demonstration and are not actual product entitlements.
Start with a simple, explicit structure
WAI's tutorial for two-header tables recommends th elements for header cells and scope="row" or scope="col" to indicate the direction of their association. In our simple comparison, the top row names the columns and the first cell in each data row names the feature.
| Feature | Starter | Team |
|---|---|---|
| Bulk export | Not included | Included |
| Support channel | Email and scheduled calls |
<table>
<caption>Fictional plan comparison: export and support</caption>
<thead>
<tr>
<th scope="col">Feature</th>
<th scope="col">Starter</th>
<th scope="col">Team</th>
</tr>
</thead>
<tbody>
<tr>
<th scope="row">Bulk export</th>
<td>Not included</td>
<td>Included</td>
</tr>
<tr>
<th scope="row">Support channel</th>
<td>Email</td>
<td>Email and scheduled calls</td>
</tr>
</tbody>
</table>The useful relationship is specific. “Included” belongs to the Bulk export row and Team column. The caption names the comparison. A styled screenshot cannot preserve these HTML associations by itself, and a bold td is still a data cell. Keep appearance and document structure under separate review.
This example is deliberately small. It illustrates one simple row-header and column-header pattern; it is not a universal solution for every table. The supplied HTML can be inspected for its explicit cells and scopes. That inspection does not establish what every browser, assistive technology or AI system will do with it.
Name the table without repeating the page
WAI describes a caption as a table identifier that helps people find it and understand what it covers. The same tutorial distinguishes a summary that explains a more complex table's organization. A caption and summary should not merely repeat each other.
For the fictional comparison, “Fictional plan comparison: export and support” says more than “Features.” It tells a reader which choices and which part of the offer are included. It does not need a long promotional introduction. If the page contains several tables, give each a name that distinguishes its purpose rather than copying the same caption everywhere.
The comparison-page guide provides broader page context. A comparison's audience and scope also matter editorially. If a real page limits an answer to a certain product edition or offer date, put that context where readers can find it. Do not use a caption to hide exceptions that materially change an answer. Header markup makes the relationship explicit; it does not verify the truth of the product claim.
Simplify before adding merged headings
A table can become harder to understand when it combines plan tiers, regions and contract terms under several merged heading levels. WAI has separate guidance for irregular headers. It describes group associations such as scope="colgroup" and scope="rowgroup", alongside markup that defines the relevant column or row groups.
Do not apply those values by guessing from the visual span. First understand the actual data structure. For a small buyer-facing comparison, consider whether two focused tables would communicate the decision better than one enormous matrix. That is a proposed editorial simplification, not a requirement to split every complex table.
If a complex structure is necessary, use the appropriate WAI pattern and review its associations carefully. Do not assume that adding scope="col" everywhere makes a multi-level structure equivalent to the simple example. Complexity should be justified by the reader's task, not by the desire to fit every sales claim into one grid.
Review the generated document as well as the design
The FAQ answer-coverage map addresses answer placement; here, check what the publishing system actually emits. A content editor may enter a “header row” in a CMS while a component renders every cell as td. A responsive layout may replace a table with cards that omit the original column names. Inspect the delivered markup and review the resulting reader experience rather than relying only on a settings panel.
For this example, a structural inspection can confirm the caption, three column headers, two row headers and four answer cells. A meaningful accessibility review still needs the site's real content, responsive behavior and relevant assistive-technology testing. No such full-site or screen-reader evaluation occurred in preparing this draft.
GeoCompanion's audit-to-execution approach can frame a comparison-page issue as a small, owned implementation task: identify the table, describe the lost relationship and verify the correction. This article does not claim GeoCompanion certifies accessibility or predicts citation gains. The practical result to pursue is a comparison whose answers retain their meaning when readers navigate through the document.
Before adding another promotional row, pick one answer cell and ask what supplies its feature and plan context. If the only answer is “you can see it in the layout,” inspect the HTML. Give the comparison a structure that carries the meaning as well as the design.
Continue the GEO Map
Follow the adjacent pages that make the AI visibility model easier for crawlers, LLMs, and buyers to understand.
See how AI sees your brand
See your AI visibility across your site, content, and competitive signal, with the next fixes and priorities mapped for you.
Boost Visibility with AIAlready have an account? Sign inNeed the creator-side next step?
Build your creator momentum on Launchvibes while GeoCompanion stays focused on AI visibility, content structure, and citation readiness.
Build your creator momentumSources
Join the GeoCompanion.ai Community
Connect with founders and marketers building stronger AI visibility, content systems, and next-generation execution.
Join TelegramFound this intelligence helpful? Propagate the signal across your nodes.