# B2B keyword research: from customer questions to a page plan

[Canonical article](https://nirlevi.com/en/blog/keyword-research-b2b/)

By [Nir Levi](https://nirlevi.com/en/about/)

Published: 2026-09-17

Keyword research connects customers’ questions to the pages your website needs. Start with a business need, inspect the search results, then decide whether to improve a page, create one or wait. Search volume informs that choice; it cannot make it for you.

## Article summary

- Define the product, audience and customer decision before expanding the keyword list.
- The largest search-volume estimate can represent the wrong audience. In this example, a broad query also returns courses and jobs.
- Group queries by the reader’s task and check existing pages before proposing new ones.
- Deliver a decision map with a page, action, evidence, owner and condition for reassessment.

Imagine you market work-management software to B2B teams. You receive a spreadsheet containing hundreds of keywords, but the team can work on one page this month. Sorting by volume puts a broad term first. It still does not tell you whether those searchers need your product.

We will follow a fictional company through that decision. The Ahrefs observations were collected for this article; the product, proposed URLs and content decisions are illustrative. This is not a client project or a report of Nir’s SEO results. The Hebrew examples retain their Israel metrics; they are not demand estimates for English markets.

Research and writing were assisted by AI. Nir’s judgment on competition and site authority comes from his personal experience; the examples and checks are presented with their sources and limitations.

## Start with the customer’s decision

A B2B product may be evaluated by an end user, team manager and procurement lead, each with different questions. Before opening a research tool, choose one decision the website needs to support.

For our fictional company, that decision is whether the software can manage tasks and dependencies in a small team. We assume the product supports this workflow. The company does not sell training or operate a jobs board. That boundary will explain why a popular term does not make the first production slot.

| Define | Our example |
| --- | --- |
| Reader | A team manager considering replacing a task spreadsheet |
| Question | Can we assign work, spot delays and track dependencies? |
| Page evidence | A clear workflow, relevant capabilities and limitations |
| Useful next action | Evaluate product fit or request a demonstration |

For a real company, start with sales, support and enquiry records you are allowed to use. Remove identifying information. If those records are unavailable, label the assumptions and plan how to test them rather than attributing invented questions to customers.

## Collect phrases and preserve their context

Your initial phrases should express several ways of describing the same problem. Our Hebrew seeds are ניהול משימות (task management), תוכנה לניהול משימות (task-management software), ניהול פרויקטים (project management) and תוכנה לניהול פרויקטים (project-management software).

Record the source, country, collection date and metric definition beside each phrase. Keep a customer’s actual question distinct from a generated idea or a research tool’s estimate.

Preserve the original wording alongside its assigned group. Hebrew spelling variants such as פרויקטים and פרוייקטים deserve checking; they do not automatically require separate pages. Do not add overlapping variants and present the total as unique people.

AI can help propose wording and organize an initial list. Keep those suggestions marked as ideas until checked. A plausible question is not measured demand.

For an established site, the [Search Console Performance report](https://support.google.com/webmasters/answer/7576553?hl=en) lets you examine queries and pages alongside impressions and clicks. Record the country, device and date filters. These describe your site’s search performance, not the size of the entire market.

## Inspect the intent behind the larger number

Search volume estimates demand for a phrase. It does not tell you how many searchers fit the product, will visit your site or will request a demonstration.

| Hebrew query / meaning | Estimated monthly average: Israel | Latest available month estimate |
| --- | --- | --- |
| ניהול פרויקטים — project management | 800 | 707 |
| ניהול משימות — task management | 350 | 358 |
| תוכנה לניהול פרויקטים — project-management software | 300 | 283 |
| תוכנה לניהול משימות — task-management software | 250 | 215 |

Source: Ahrefs Keywords Explorer, Israel, retrieved by Codex on 17 September 2026. The average uses the tool’s latest known 12 months; the other column reports the latest available month estimate. Neither is site traffic. [Inspect the saved data and definitions](https://nirlevi.com/blog-assets/keyword-research-b2b/research-evidence.json).

Sorting only by the average would prioritize project management. But the Ahrefs snapshot dated 3 August 2026 returned a Campus IL course, Wikipedia, a jobs page and another course among the organic results supplied for positions 1–5.

For project-management software, the snapshot dated 30 July 2026 returned product pages and a software comparison within the same position range. Some products focus on construction and engineering, so even this narrower query still needs an audience-fit check.

These are two historical snapshots from different dates, not an experiment or a verified description of today’s results. They illustrate the decision step; refresh the results before investing in the page. AI citation links were separated from organic results rather than counted as additional organic pages.

The initial decision is to hold the broad phrase and investigate software searches. Those searches can still represent different needs; next, we check whether they belong on the same page.

## Choose a tool for the uncertainty you need to resolve

You do not need every tool for every research task. Choose the source that answers the next question: what people already find on your site, what wording to investigate, or which page types appear in search.

| Source | Useful for | Do not infer |
| --- | --- | --- |
| Search Console | Queries and pages associated with your site’s visibility | Site impressions are not total market search volume |
| Google Keyword Planner | Ideas and planning data under selected settings | Advertiser competition is not organic ranking difficulty |
| Google Trends | Relative interest over time and within a region | A value of 100 does not mean 100 searches |
| Ahrefs Keywords Explorer | Demand estimates, ideas and stored SERP observations | Low KD does not guarantee ranking or traffic |

Google’s [Keyword Planner access instructions](https://support.google.com/google-ads/answer/7337243?hl=en) require account setup, including billing information, for basic features. Do not promise readers complete access without registration. Its [metric documentation](https://support.google.com/google-ads/answer/3022575?hl=en) defines competition in terms of advertisers.

[Google Trends](https://support.google.com/trends/answer/4365533?hl=en) normalizes relative interest to a 0–100 scale; low-volume terms can appear as zero. Use it to investigate patterns, not as a replacement for volume estimates. With [Ahrefs](https://ahrefs.com/keywords-explorer), retain the metric basis and the search snapshot’s date.

This table compares documented roles. Ahrefs was the tool used to collect the article’s observations; this was not a four-tool accuracy benchmark. Check current pricing, limits and account access when selecting a plan.

## Quick answers before you continue

Choose your starting point around the evidence your site already has. Tool roles and access conditions are grounded in the [Keyword Planner instructions](https://support.google.com/google-ads/answer/7337243?hl=en), [Search Console performance report](https://support.google.com/webmasters/answer/7576553?hl=en) and [Google’s metric definitions](https://support.google.com/google-ads/answer/3022575?hl=en).

### What should keyword research deliver?

The output is a reasoned page map: each query group has a reader need, an existing or proposed page, an action and supporting evidence. For example, task-management software may call for improving an existing solution page. Add an owner and a condition for revisiting the decision. A list of keywords and volumes alone is not a content plan.

### Which free tools should you start with?

For a site already earning impressions, start with Search Console to identify existing queries and pages. For a new site, start with customer questions and Keyword Planner ideas, using Google Trends to investigate patterns. Inspect search results for each candidate. Google requires account setup and billing details for basic Keyword Planner features, so check access requirements before relying on it.

### How do search volume, competition and search intent differ?

Search volume estimates demand; search intent describes the task the searcher wants to complete. Keyword Planner competition concerns advertisers, while SEO-tool KD is that tool’s estimate of organic difficulty. None of these measures alone determines which page to build. Check product fit and the results first, then consider demand and your ability to compete.

### Can AI replace a keyword research tool?

AI can suggest wording, clean exports and help make decisions from supplied evidence. A model answer without a data source is not a search-volume measurement. If the system connects to a research tool, ask for its source, country and measurement date, then verify the output against them. Keep missing values unknown rather than treating invented numbers as evidence.

## Choose between a new page, an update and a hold

Similar wording is not enough to assign two queries to one page. In this example, the next question is whether “תוכנה לניהול משימות” (task-management software) and “תוכנה לניהול פרויקטים” (project-management software) represent the same buying decision.

The additional check returned 8 organic results for the task query, from an Ahrefs snapshot dated 4 August 2026, and 9 for the project query, dated 30 July. No exact page URL appeared in both lists. One domain, Projecteam, appeared in both through different pages: its task page and its homepage. The [result lists and comparison method](https://nirlevi.com/blog-assets/keyword-research-b2b/grouping-comparison.json) are available to inspect.

![Two saved result lists: 8 and 9 results, no shared URL, and different Projecteam pages.](https://nirlevi.com/blog-assets/keyword-research-b2b/grouping-evidence.en.png)

Annotated view of Ahrefs data retrieved by Codex on 17 September 2026. This is a capture of a local data report, not a Google screenshot. The result dates differ. [Accessible report with every URL](https://nirlevi.com/blog-assets/keyword-research-b2b/grouping-evidence.en.html).

Opening the Projecteam pages adds context: its [task page](https://projecteam.tools/he/task_mng_heb) also addresses daily organization and personal tasks, while the [homepage](https://projecteam.tools/) presents projects and tasks for teams. One supplier can serve more than one need. Inspecting these pages does not establish that they still hold their recorded rankings.

This is a reason to avoid automatic grouping, not proof that two pages must be created. The snapshots have different historical dates and unequal result counts. A rule such as “zero shared URLs always requires separate pages” would ignore the actual content and audience.

The worksheet therefore keeps two separate rows: start with the page explaining team task management, and hold the project-management group until current results and product fit have been checked. The fictional company already has a suitable solution page, so the first action is to improve that page.

| Group | Illustrative decision | Required before execution |
| --- | --- | --- |
| Task-management software | Improve the existing team solution page | Current results and an actual product walkthrough |
| Project-management software | Separate row on hold; no new page yet | Audience, capability and current-results checks |
| How to define task dependencies | Guide idea, not a writing assignment yet | Demand check and a useful demonstration |
| Project management | Hold at lower priority | Fit with education or employment intent |

To repeat the check, collect results for the same country as close together in time as possible, preserve full URLs and page types, and open the relevant pages. Separate organic results from ads and AI citations. If a refresh shows the same pages serving the same task, reconsider grouping; if the audience or required output differs, plan a distinct answer before creating another page.

For a Hebrew and English site, research each market separately. Translating a query does not transfer its measured Israeli demand or intent.

## Prioritize a decision the team can act on

In my experience, when a site is new or its authority is still low, I reject a highly competitive keyword as an initial target even if its search volume is large. I would not make it the main priority before the site has a suitable foundation to compete. That is a sequencing decision, not a judgment that the topic has no business value.

To turn that position into a practical check, compare the competing pages: how closely they answer the question, the evidence they provide, the links pointing to them and what your site can already contribute. Look for a focused need you can serve better with the resources available.

A metric such as [Ahrefs Domain Rating](https://help.ahrefs.com/en/articles/1409408-what-is-domain-rating-dr) describes the relative strength of a backlink profile. It can inform the comparison, but does not settle page quality or ranking prospects. This guide sets no DR threshold for writing or rejecting a page.

- Fit: will the page help an audience the business can serve?
- Competition: can the site and team address the gap now?
- Evidence: can you provide a useful explanation, dataset or demonstration?
- Execution: would improving an existing page be enough, and what would change the decision?

The example does not measure a real company’s authority, so it cannot label the chosen query easy to rank for. The decision is to scope work around the existing task page and assess feasibility before investing. If serving a query requires inventing a product capability, remove it even when the competition appears low.

## Prompt toolbox: copy a prompt for your next research step

These prompts help turn research into decisions. Open the one that fits your current step, copy it and replace the square brackets with your product details and data. Pass reviewed output to the next step; you do not need to run every prompt each time.

The prompts do not supply search data by themselves. Provide exports or sources you have checked, and preserve missing data as unknown. Before adopting a recommendation, inspect its source rows and product fit. For this article’s example, you can give prompt 3 the two lists in the comparison file; the historical dates should remain in the answer.

1. Turn customer questions into seed queries

Use when you have product context and customer questions, but no research list yet.

```
Help me create a seed-query list for a B2B product.

Product and verified capabilities: [details]
Audience and the decision they need to make: [details]
Target language and country: [details]
Anonymized customer questions: [paste; state if unavailable]

Suggest queries expressing this audience’s problems and decisions. Separate wording from the supplied customer questions from your own suggestions. Do not invent product capabilities, search volumes, competition or questions that customers actually asked. If essential product or audience information is missing, ask before suggesting queries.

Return a table: query | reader need | supplied source or illustrative idea | product fit | required validation.
End with the queries you would investigate first and why. Each idea is a research candidate, not an instruction to write an article.
```

2. Clean an export without losing context

Use when you have keyword-tool data that needs to be prepared for comparison.

```
Clean the following keyword export for content planning.

Data: [paste a table or attach a file]
Tool, country, date range and metric definitions: [known details]

Preserve each original query and source row ID. Add a normalized query column for review only. Flag duplicates and spelling variants, but do not automatically delete or combine rows with different countries, dates or metric definitions. Do not add variant volumes and present the sum as a unique audience.

Distinguish zero from missing data and a yearly average from the latest-month estimate. Do not alter numbers or fill missing values. Ask for the definition of any ambiguous column.

Return a clean table retaining source row ID, query, provenance, measurement conditions and notes, followed by issues requiring my decision. Do not infer search intent from the numbers alone.
```

3. Decide whether two queries need one page

Use with two result lists and content from the relevant pages.

```
Help me assess whether two queries can be served by the same page.

Query A and its results: [full URLs, titles and result types]
Query B and its results: [full URLs, titles and result types]
Country, language, device if known, source and date for each list: [details]
Relevant excerpts from pages that were opened: [content and source URL]
Our product and audience: [details]

Compare organic results only. Separate ads and AI citations. Report shared URLs separately from shared domains and explain any normalization. State how many results each list contains; do not assume ten.

Compare the reader’s task, page type and audience. Where page content was not opened, label the classification as title-based only. Do not describe old results as current or infer causation from overlap.

Return: observations linked to source rows | possible interpretation | recommendation to group, separate or investigate | evidence that would change the recommendation. Do not substitute a fixed overlap threshold for judgment.
```

4. Turn research into a page decision

Use to choose an update, a new page or a hold, then hand the team a task.

```
Turn the attached research into a page-decision map.

Query groups, needs and evidence: [paste checked research]
Existing pages and their roles: [URL, audience, content and desired action]
Product capabilities and limitations: [details]

For each group, first assess whether an existing page can serve the need. Recommend a new page only for a distinct reader task with useful content we can provide. Do not invent existing URLs, performance metrics or capabilities. Label proposed URLs as new. Hold decisions with missing or contradictory evidence.

Return a table: group | reader need | existing or proposed page | update, new or hold | reasoning and source | missing information | next task | responsible role.

For the best-fitting group, add a mini-brief: the reader’s decision, the explanation currently missing, evidence to collect and page acceptance criteria. Do not draft the full article yet or propose a page for every wording variant.
```

5. Prioritize for fit and ability to compete

Use when the page map exceeds the team’s available time and resources.

```
Prioritize this content map by business fit, evidence and ability to execute.

Page map: [paste]
Site maturity, relevant pages and known performance: [data and sources; mark gaps]
Competing results and available backlink or authority data: [data, tool and date]
Team capacity and planning period: [details]

Do not select a target solely for high search volume. For a new or low-authority site, assess whether a competitive target should wait. Compare actual pages and evidence; do not turn DR or KD into an automatic cutoff or ranking guarantee. Without competition or authority data, mark ranking feasibility as unassessed.

Return three groups: work now, investigate before deciding, and hold. For each item, include reasoning, explicitly estimated effort, a dependency and evidence that would change its priority. Distinguish research tasks from writing assignments.

End with one task achievable this week within the supplied capacity. Do not forecast traffic, leads or revenue without suitable data and a model.
```

[Download all prompts as a text file](https://nirlevi.com/blog-assets/keyword-research-b2b/prompts.en.txt). Transfer the decisions you have checked to the page map in the next section.

### Worked response: should the software queries share a page?

Codex prepared a worked response to the third prompt using the article’s two lists and fictional company description. Download the full input to try it and check whether your response preserves the evidence. This is one demonstration, not a model-reliability benchmark or research performed by Nir.

| Output check | What the records support |
| --- | --- |
| List size | 8 organic task results; 9 project results. Missing results were not filled in. |
| Source dates | 4 August and 30 July 2026. A further Ahrefs check on 17 September returned the same snapshots. |
| URL versus domain | 0 shared URLs; one shared domain, Projecteam, represented by two different pages. |
| Content not supplied | No page excerpts were included. Intent interpretation relies on titles only. |
| Practical decision | Two groups to investigate; assess the existing task page first. No instruction to create a new page. |

A separate calculation against the source file confirmed the counts, dates and overlap. Codex also checked the recommendation against the page map. This did not verify current search intent: you still need current page inspection and product checks. If your response says “two pages are required,” ask it to separate the observation from the recommendation.

[Full prompt input](https://nirlevi.com/blog-assets/keyword-research-b2b/prompt-example-input.en.txt) · [Response and next decision](https://nirlevi.com/blog-assets/keyword-research-b2b/prompt-example-output.en.txt) · [Numerical source check](https://nirlevi.com/blog-assets/keyword-research-b2b/prompt-example-check.json)

## Hand the team a usable page map

Give a writer a decision and its rationale, not just a keyword list. The downloads contain a filled example and a blank template with query, reader need, proposed page, action, source, date, responsible role and next check.

[Download the example map in English](https://nirlevi.com/blog-assets/keyword-research-b2b/query-page-map.en.csv) · [Download the blank template](https://nirlevi.com/blog-assets/keyword-research-b2b/query-page-template.en.csv). The product URLs are fictional and are not Nirlevi.com pages. Measured observations are separated from assumptions and untested questions.

Choose one product and one query group. Complete the sheet, mark what remains unknown and assign one task the team can complete this week. After publication, check whether the intended queries surface the page and whether visitors complete a useful action.

Keep impressions, visits and qualified enquiries separate. A before/after change does not establish that the page update caused the difference. If data is absent, first check page access and measurement collection; the [indexing diagnostic guide](https://nirlevi.com/en/blog/why-google-is-not-indexing/) provides a starting point.

When the map spans product, content and engineering decisions, [content strategy consulting](https://nirlevi.com/en/services/content-strategy/) can help turn it into a work plan. For wider search priorities, start with [SEO consulting](https://nirlevi.com/en/services/seo-consulting/).

## Sources and further reading

- [Google Ads: Use Keyword Planner](https://support.google.com/google-ads/answer/7337243?hl=en)
- [Google Ads: About Keyword Planner forecasts](https://support.google.com/google-ads/answer/3022575?hl=en)
- [Google Trends: data normalization and low-volume searches](https://support.google.com/trends/answer/4365533?hl=en)
- [Search Console: Performance report](https://support.google.com/webmasters/answer/7576553?hl=en)
- [Ahrefs Keywords Explorer](https://ahrefs.com/keywords-explorer)
- [Ahrefs Israel: saved keyword and SERP observations, retrieved 17 September 2026](/blog-assets/keyword-research-b2b/research-evidence.json)
- [Ahrefs: What is Domain Rating?](https://help.ahrefs.com/en/articles/1409408-what-is-domain-rating-dr)

[Plan content your team can deliver](https://nirlevi.com/en/services/content-strategy/)