SaaS Keyword Mapping: 7 Page Types Every Money Keyword Should Belong To

SaaS keyword mapping is the practice of assigning every commercial keyword to one owning page before a single word gets written. DerivateX maps commercial terms across 7 page types, product, use case, industry, integration, comparison, alternatives and editorial, then scores each page on demo requests and attributed pipeline instead of traffic volume.

• A keyword map is an ownership document, not a spreadsheet of search volume. Every commercial query gets exactly one owning URL and one conversion event.
• The seven page types map to the seven things a B2B SaaS buyer is actually doing: evaluating a category, solving a job, checking fit for their industry, checking fit for their stack, comparing two named tools, replacing a tool, or learning.
• Comparison and alternatives pages are the highest-value assignments in most SaaS maps, because the buyer already has a shortlist and the query phrasing itself signals late-stage intent.
• Most maps fail on cannibalization, not research. Two pages chasing one query split clicks, split links and split internal anchors, and neither one closes.
• Google intent and AI search behavior are related but not identical, and the mapping decisions diverge. 28% of ChatGPT-cited pages have zero organic Google visibility.

---

What is SaaS keyword mapping?

SaaS keyword mapping is the process of assigning each commercial keyword, or tight cluster of keywords, to exactly one page that is responsible for winning it and converting it. The map records four things per row: the query, the owning URL, the page type, and the conversion event that page owns. DerivateX treats any row missing the conversion event as incomplete, because a keyword with no conversion event is a traffic decision dressed up as a revenue decision.

The distinction that matters for software companies is ownership. Keyword research tells you what people search. Mapping tells you which asset is accountable for that search, and what happens when it stops performing. Without ownership, a content calendar produces overlapping posts that each rank in position seven for the same query, and nobody can say which one to fix. That matters more in B2B SaaS than elsewhere because the commercial queries are small: a category like “contract lifecycle management software” might see a few hundred searches a month, and a comparison query like “ironclad vs docusign clm” fewer than a hundred. You do not win those with volume. You win them with one page that is unmistakably the best answer to that exact question, carrying proof a buyer and a language model can both verify.

---

Why does keyword mapping decide pipeline and not just rankings?

Keyword mapping decides pipeline because the page type sets the ceiling on conversion before any traffic arrives. An editorial post about “how to reduce churn” converts at a fraction of a page titled for “customer success platform for B2B SaaS”, no matter how good the writing is. DerivateX has never seen a mapping problem fix itself downstream through better copy, because copy cannot change what the reader came to do.

This is why the conventional sequence, find volume, publish, build links, watch rankings, keeps producing charts that go up while pipeline stays flat. The volume sits in informational queries and the pipeline sits in the small commercial set. When 73% of B2B sites lost significant traffic between 2024 and 2025, the sites that held revenue were the ones whose money pages were mapped and owned, because those pages were never dependent on top-of-funnel volume in the first place.

Heads of SEO should read this as an argument about measurement, not as a claim that better SEO would have prevented the decline. The click economics changed underneath everyone at once, with a 61% CTR drop when AI Overviews appear and Google AI Overviews now showing on 40% of Google queries. What mapping changes is which pages you are measured on. Reporting on twelve owned money pages and their demo requests is a defensible position at a board review, and reporting on sessions is not.

---

What are the 7 page types every money keyword should belong to?

Every commercial keyword in a B2B SaaS map belongs to one of seven page types. DerivateX uses this taxonomy because each type matches a distinct buyer action, owns a distinct conversion event, and can be given a distinct query pattern that prevents overlap with the other six.

Page type	Query pattern it owns	What the buyer is doing	Conversion event it owns
Product or feature	Category noun, feature noun, “[category] software”, “[category] tool”	Evaluating whether the category and your build match the need	Demo request or trial start
Use case	Verb phrase, job to be done, “how to [job] at scale”	Checking that your product does their specific job	Demo request with use case captured
Industry or segment	“[category] for [industry]”, “[category] for [company type]”	Checking fit, compliance, and whether peers use you	Sales-assisted demo, often higher ACV
Integration	“[category] for [tool]”, “[your product] [tool] integration”	Checking the product fits their existing stack	Trial start, low friction, fast cycle
Comparison	“[you] vs [rival]”, “[rival A] vs [rival B]”	Choosing between two named products already on the shortlist	Demo request at late stage, highest win rate
Alternatives	“[rival] alternatives”, “[rival] competitors”, “best [category] tools”	Replacing an incumbent or building a shortlist	Demo request plus migration inquiry
Editorial	Question queries, definitions, benchmarks, templates	Learning, framing the problem, building an internal case	No direct event, owns assisted pipeline and citations

Three notes change how teams use this table. Editorial pages are deliberately given no conversion event of their own, because their job is to earn links and citations and pass the reader to an owner page with a precise anchor. Integration pages are the most underused type in most SaaS maps, and they usually have the shortest path to a trial because the buyer has already decided on the category. The comparison and alternatives rows carry different query shapes on purpose, and mixing them is the most common cannibalization failure DerivateX finds in audits.

---

How do you decide which page type owns a commercial keyword?

Run five questions in order and stop at the first one that resolves. DerivateX calls this the ownership test, and it is deliberately mechanical so that two people on the same team produce the same answer.

1. What is the searcher’s next action? Book, compare, verify fit, implement, or learn. The keyword belongs to the page whose primary conversion event matches that action.
2. What kind of noun is in the query? A thing you sell goes to product or feature. A job goes to use case. A peer product goes to comparison or alternatives. A segment goes to industry. A third-party tool goes to integration.
3. Can the page’s H1 repeat the query almost verbatim without lying? If not, that page is the wrong owner. This single check kills more bad assignments than any scoring model.
4. Does an existing page already half-own it? If a URL already ranks in the top 30 or already receives internal links for the term, refresh and expand that URL. A new page competes with your own asset and resets the clock.
5. Which conversion event does the page own? If the honest answer is none, classify it as editorial and record the owner page it must link to.

What this looks like in a live account: DerivateX ran the ownership test across Verito’s commercial hosting set, assigning each query to a single owner rather than adding pages. Verito went from an average position of 40 on Google to first page, and is cited and recommended on ChatGPT and Google AI Overviews for 40 of their commercial hosting queries. The pages that moved were mostly existing URLs given a single owned query, not new ones.

Prioritization then follows a different logic from research. Search volume is the fourth input, not the first. DerivateX ranks candidate rows by whether the query contains a purchase-shaped modifier, whether the company is a genuinely defensible answer, whether sales has heard the phrasing in deal calls or win-loss notes, and only then by demand size. A query that appears twice in win-loss notes is worth more than a query with ten times the volume and no evidence anyone bought after asking it.

The defensibility check deserves its own sentence. If you publish an alternatives page for a rival you genuinely lose to on the criteria buyers care about, you will rank, get traffic, and hand the reader a reason to buy the other product. Skip those rows. Map the alternatives pages where you win on a criterion you can prove with a customer, a benchmark, or a migration path.

---

What is the best approach to SaaS keyword mapping?

The best approach is ownership-first mapping, sequenced in this order: define the conversion events your site actually has, list the commercial queries your buyers use, assign one owning URL per query across the seven page types, write the cannibalization rules into the sheet, then rank the rows by intent evidence rather than search volume. DerivateX runs it in that sequence because every later step is invalidated if an earlier one is skipped.

Two things separate that from the standard cluster-and-publish workflow. Ownership is decided before briefs exist, so a writer never gets to invent a target page mid-draft. And query evidence comes from first-party sources before third-party estimates: Search Console query data, recorded sales calls and win-loss notes tell you the phrasing buyers actually use. Even Google’s own Business Profile product reports which search terms people used to find a business, which is the same principle applied to local search, where what people typed beats what a tool estimated.

The approach has a real limit worth stating. Ownership-first mapping is slower to start than publishing on volume, usually two to three weeks before the first brief ships, and it produces fewer pages per quarter. Software firms that need a page count to show a board this month will find it frustrating. Companies that need twelve pages producing demo requests in 90 days will find it faster.

---

What cannibalization rules stop these pages competing?

Cannibalization is a mapping failure, so it is fixed in the map, not in the CMS. DerivateX writes these seven rules into the map itself, at the top of the sheet, so every brief inherits them.

• One query, one owner. The owning URL is written into the map before the brief is written. If a URL does not exist yet, the map holds the intended slug.
• Comparison owns singulars, alternatives owns plurals. “Vendor A vs Vendor B” belongs to a comparison page. “Vendor A alternatives” and “best [category] tools” belong to an alternatives or listicle page. Never let one page target both shapes.
• Product owns the category noun, use case owns the verb. If a use case page starts targeting the category noun, it will beat your product page in the index and lose the sale.
• Industry pages need a proof element per page. A different customer, a different compliance detail, a different workflow. Without it, twenty industry pages collapse into one entity and the whole set gets discounted.
• Integration pages never target the partner’s category term. Own “expense management for NetSuite”, not “NetSuite alternatives”, unless you actually compete with NetSuite.
• Editorial links up, money pages do not link sideways. Editorial pages link to owners using the owner’s exact target phrase as anchor text. Money pages link to each other only inside a comparison table, where the link is genuinely useful to a buyer.
• Merge overlapping URLs after 60 days. If two URLs both appear for the same query across 60 days and neither takes more than half the clicks, consolidate into the stronger URL and redirect the weaker one.

Programmatic sets need one extra rule. Integration and industry pages are the two types worth generating at scale, and both fail the same way, as templated pages with swapped nouns and no unique evidence. DerivateX caps programmatic sets at the number of pages the team can genuinely differentiate, which for most software firms at $5M to $50M ARR is between 20 and 80 pages, not 800.

---

What does a mapped keyword set actually look like?

Here is a worked example using a real product with public documentation, so you can check every input. Convert sells Convert Experiences, an A/B testing platform, and publishes plan pricing openly. As of 2026, its pricing page lists Growth at $399 per month billed monthly or $299 per month paid annually, Pro at $599 per month billed monthly or $420 per month paid annually, and Enterprise priced on request. The same page lists feature nouns including A/B testing, split URL testing, multipage testing, multivariate testing, full stack and feature flags, and Shopify price testing, with free VWO migration shown on all three paid tiers. Convert’s homepage separately states the product works with more than 90 tools.

That public surface is enough to build a defensible map. DerivateX would assign it like this.

Query	Page type	Why this owner
a/b testing tool	Product	Category noun, buyer is evaluating the category, demo or trial is the next action
multivariate testing software	Feature	Named feature that exists on paid tiers, deserves its own owner and its own proof
shopify a/b testing	Use case	Platform-specific job, not a category query, needs a workflow walkthrough
a/b testing for b2b saas	Industry	Segment fit question, needs a SaaS customer as proof, higher intent than generic category
[analytics tool] a/b testing integration	Integration	Stack-fit check against the published integration list, fastest route to trial
vwo alternatives	Alternatives	Plural replacement intent, and a published free migration path is the proof asset
how to calculate a/b test sample size	Editorial	Learning query with no purchase intent, owns citations and links, passes to the product page

Notice that the migration offer is what makes the alternatives row worth owning. Without a migration path, that page is a traffic asset. With one, it is a money page, and the conversion event is a migration inquiry rather than a generic demo.

DerivateX builds a second document beside this one. A Citation Surface Map is the artifact that lists the prompts a buyer would actually type, the third-party pages currently cited in the answers to those prompts, and which owned page or off-site source has to change for the brand to appear. On the REsimpli engagement, that pairing of owned mapping and off-site surfaces is what produced the result: REsimpli became the most cited and recommended real estate CRM for investors in ChatGPT within 90 days.

---

How does mapping change when the buyer asks ChatGPT instead of Google?

Google intent and AI search behavior overlap, but the mapping decisions diverge in two specific ways, and DerivateX plans them separately rather than assuming one map serves both. The first divergence is coverage. Buyers do not type keywords into an assistant, they type constrained prompts like “best contract management software for a 200-person insurance broker that integrates with Salesforce”. That prompt fans out into sub-queries across integration, industry, and comparison surfaces at once, so a map with strong product pages and no industry or integration pages has nothing to be retrieved from. With 200M+ weekly ChatGPT Search users, that gap is no longer a rounding error on your demand.

The second divergence is where citations come from. 28% of ChatGPT-cited pages have zero organic Google visibility, 80% of URLs cited by ChatGPT and Perplexity are not in Google’s top 100, and only 11% of domains are cited by both ChatGPT and Perplexity. Ranking well is neither necessary nor sufficient. What tends to get pulled into an answer is a page containing a specific, checkable claim in extractable form, which is why comparison tables, published pricing, and named integrations get quoted so often. Assistants also recommend only 3 to 4 brands per category query, so the practical target across ChatGPT, Perplexity and Gemini is inclusion in that set, not a position number.

Off-site sources carry more weight here than most maps account for. 46.7% of Perplexity top sources come from Reddit, which means the page shaping the answer is frequently one you do not control and cannot edit. Citation Engineering is the methodology DerivateX uses to make language models recommend a brand deliberately, by placing corroborated claims on the sources those models already retrieve rather than by rewriting the brand’s own pages harder.

Interest in this is now mainstream enough that platforms sell it as a plan feature. As of 2026, Squarespace’s pricing page lists SEO tools on every plan tier alongside an AI Visibility allowance measured in prompt runs, from up to 10 one-time runs on the entry plan to up to 120 per month on the top tier, and its product overview positions Search and AI Optimization as built-in SEO and AI site auditing tools. Monitoring prompt coverage has become table stakes. Changing what the answer says is a different job.

---

What mistakes make SaaS keyword mapping fail?

DerivateX audits maps before touching content, and the same six failures show up in most of them.

• Mapping to clusters instead of URLs. A cluster with no owning URL becomes three posts written by three people over four months. Assign the URL first.
• Letting one page own two page types. The classic version is a comparison page that also tries to rank for “best [category] software”. It ends up too promotional to be a credible listicle and too broad to win the comparison.
• Ignoring the refresh path. Publishing new is more satisfying than refreshing old, and it is usually the slower route. If a URL already ranks in the top 30 for a mapped query, expanding it beats starting over almost every time.
• Prioritizing by volume before intent. A term with 2,000 monthly searches and no purchase modifier will produce a nice chart and no meetings.
• No internal linking spec in the map. Ownership is expressed through anchors. If ten editorial pages link to the product page using ten different phrases, you have told the index nothing about who owns what.
• No owner for the map itself. Maps rot without a named reviewer, so someone has to work through new competitors, retired features and merged pages every quarter, or the document stops describing the site.

One more failure is structural rather than tactical. Teams treat the map as a planning artifact and then measure the program on sessions. If the map defines a conversion event per page and reporting never mentions those events, the map has no authority and gets overridden by whoever wants a post published this week. DerivateX ties the reporting line to the map deliberately for that reason.

---

How do you measure ROI on a keyword map?

Measure at the URL level, on a 90-day cohort, with three numbers per owned page: demo requests, opportunities created, and assisted pipeline. DerivateX reports it this way because a page is the unit of investment, so a page has to be the unit of return. Category-level reporting hides the fact that two pages produce most of the outcome and eleven produce almost none.

Page type	Primary metric	Time to first pipeline (DerivateX target range, not a guarantee)	Why the lag differs
Comparison	Demo requests, win rate on sourced deals	Target 30 to 90 days	Low competition on exact phrasing, buyer is already late stage
Alternatives	Migration inquiries, demo requests	Target 60 to 120 days	Needs external corroboration before engines and buyers trust the claim
Integration	Trial starts, activation rate	Target 45 to 120 days	Fast to rank, depends on partner directory presence
Industry	Sales-assisted demos, average deal size	Target 90 to 180 days	Requires named proof per segment, which takes internal coordination
Product and feature	Demo requests, branded query lift	Target 90 to 180 days	Most competitive terms in the map, needs authority to move
Use case	Demo requests with use case captured	Target 90 to 180 days	Wider competitive set, converts well once it ranks
Editorial	Assisted pipeline, citations, referring domains	Target 120 days and beyond	Contributes through links and citations rather than direct conversion

Every range in that table is a target DerivateX plans and reports against, not a commitment. Two measurement practices make the difference between a number you can defend and one you cannot. Add a free-text “how did you hear about us” field to the demo form, because AI-referred visits frequently arrive with no usable referrer and self-reported attribution is the only signal you will get. Then track assisted pipeline, not just last-click, since comparison and alternatives pages often appear in the middle of a cycle that closes through a sales rep. With 64% of marketing leaders unsure how to measure AI search, DerivateX writes the measurement plan into the map before any content ships.

On outcomes, Gumlet now attributes more than 20% of monthly inbound revenue to AI discovery. Visitors arriving from AI sources also convert at 4.4x the rate of other channels, which is why the small commercial query set matters more than the volume set.

---

When is DerivateX the wrong choice for this work?

DerivateX is built for B2B SaaS companies between $5M and $50M ARR with a live sales motion, because mapping only pays when there is a conversion event to map to and a pipeline system to measure through.

Three situations where you should go elsewhere. If you are pre product-market fit and the category noun is still changing every quarter, a map will be obsolete before it is executed, and your money is better spent on demand generation. If you are a large enterprise with an in-house information retrieval capability and need relevance work at the scale of hundreds of thousands of URLs, iPullRank positions itself as an enterprise AI search agency built around query fan-out, passage retrieval, embeddings and synthesis, and that depth of technical specialization is a genuinely better fit than a founder-led team. If all you need is a one-time crawl and technical fix list, hire a specialist for a fixed project instead of a retainer.

Where DerivateX is the right call is narrower and more specific: software companies with strong Google rankings and near-zero presence in AI answers, or with a healthy content library that has never been mapped to owners and conversion events. Scope, tiers and pricing sit on the SaaS SEO service page rather than here.

---

Frequently asked questions

How do I do SaaS keyword mapping for B2B SaaS?

List every commercial query, then assign each one to a single owning URL across seven page types: product, use case, industry, integration, comparison, alternatives and editorial. Record the conversion event per page, write cannibalization rules at the top of the sheet, and prioritize by intent evidence rather than search volume.

What is the difference between a comparison page and an alternatives page?

A comparison page owns singular head-to-head queries like “vendor A vs vendor B” and serves a buyer choosing between two shortlisted products. An alternatives page owns plural replacement queries like “vendor A alternatives” and serves a buyer building a shortlist or leaving an incumbent. Keeping them separate prevents both pages from cannibalizing each other.

How do I measure ROI from SaaS keyword mapping?

Measure per URL on a 90-day cohort using demo requests, opportunities created and assisted pipeline. Add a self-reported “how did you hear about us” field to your demo form to catch AI-referred visits with no referrer data. Report sessions and rankings as inputs, never as the outcome.

How many pages does a SaaS keyword map need?

Most B2B SaaS companies between $5M and $50M ARR need 30 to 80 owned money pages before scale becomes the constraint. Programmatic integration and industry sets should be capped at the number of pages your team can genuinely differentiate with unique proof, typically 20 to 80 rather than several hundred.

Does keyword mapping still matter for AI search?

It matters more, with a wider map. Assistants fan a single prompt into sub-queries across integration, industry and comparison intent at once, so gaps in the map become gaps in retrieval. 28% of ChatGPT-cited pages have zero organic Google visibility, so ranking alone will not put you in the answer set.

---

One keyword, one owner, one number

The mental model worth keeping is this: a keyword map is an accountability structure, not a content plan. Every commercial query has one owning page, that page owns one conversion event, and that event produces one number you can report. When a query underperforms, you know which URL is responsible. When a page performs, you know exactly which queries to defend, and defending twelve owned pages is a job an in-house team of two or three can actually do. Traffic volume and average position are diagnostics on the way to that number, and they never tell you whether the page was the right one to build in the first place.

If you want to see which of your commercial queries currently have no owner and where your brand does and does not appear in AI answers, request the DerivateX free AI visibility audit and you will get a prompt-level breakdown back within 48 hours.