Criteria-Combination Answer Architecture
Publish structured answers to specific combinations of buyer criteria
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 96%
This architecture converts long-tail question research into a manageable set of pages organized around combinations of buyer criteria. Instead of generating a page for every conceivable prompt, the company groups needs by persona, feature, integration, language, and use case. It then publishes authoritative lists, capability pages, and help-center articles that clearly state what the product supports. A persona page might enumerate all features relevant to HR, while an integration page can answer several related questions at once. Internal links connect these resources so crawlers and answer engines can locate them. The mechanism captures many detailed prompts with dozens of substantial pages rather than millions of thin variants, while keeping every claim useful to human visitors and grounded in the actual product.
Origin
Extracted from Marketing Against The Grain
Core principles
- 01Answer engines excel at matching multiple criteria
- 02Specific answers create opportunities where no incumbent exists
- 03A useful list can cover many combinations without one page per prompt
- 04Help centers naturally explain what a product can and cannot do
How to run it
- 1
Extract combinations of criteria
Identify questions containing several constraints, such as a feature, use case, integration, language, and persona.
Pro tip Start with combinations repeatedly observed in sales, support, and community data.
Watch out Do not create content for hypothetical combinations with no evidence of demand.
- 2
Group questions into durable page types
Organize the questions into persona pages, feature lists, integration lists, language resources, use-case pages, and help-center articles.
Pro tip Choose page types that can answer several related prompts without becoming vague.
Watch out One page per exact prompt can create a large amount of thin content.
- 3
Write explicit product answers
State whether and how the product satisfies each criterion, using concise explanations and structured lists where useful.
Pro tip Include concrete setup instructions or limitations when they help users evaluate fit.
Watch out Do not imply support for a feature or integration the product does not provide.
- 4
Use the help center as evidence
Publish operational details in the help center because it naturally documents how the product works.
Pro tip Update help articles whenever product capabilities change.
Watch out Outdated documentation can cause answer engines to repeat obsolete claims.
- 5
Connect related resources
Add internal links among persona, feature, integration, and help-center pages so each answer is discoverable in context.
Pro tip Use descriptive anchor text that identifies the relationship between resources.
Watch out Orphaned pages may remain invisible to both users and crawlers.
- 6
Expand by observed demand
Add or deepen pages as new recurring question clusters appear instead of attempting exhaustive generation in advance.
Pro tip Review new support and sales questions on a recurring schedule.
Watch out Publishing at mass scale without quality control can clog the site and dilute trust.
In the wild
Rippling serves personas including HR, IT, finance, and operations. Rather than creating a separate page for every possible question, it maintains pages listing the features available to each persona, allowing one structured resource to answer many detailed capability prompts.
→ A limited set of substantive pages covers a broad range of persona-specific questions.
A buyer wants a website builder for a design portfolio that also supports SEO, AEO, and a Stripe integration. Webflow can answer the combination through connected use-case, capability, integration, and help-center content rather than relying on a generic product page.
→ The answer engine has explicit evidence for matching the product to the buyer’s detailed criteria.
Common mistakes
Creating millions of thin pages
Exhaustive prompt permutations can clog the site and produce low-value content instead of authoritative answers.
Publishing generic feature copy
Broad marketing language does not resolve the specific combinations of criteria users give an answer engine.
Neglecting the help center
The most credible details about how a product works often reside in operational documentation.
Is it for you?
Best for
Products with many features, integrations, personas, languages, or specialized use cases.
Not ideal for
Simple products with few meaningful variations or companies unable to verify detailed capability claims.
From the transcript
“And so a repeatable thing is to have answers to questions that people are asking.”
“So you don't necessarily need a page for every single feature or use case, but you can create these lists of features or lists of…”
“You see help centers showing up a disproportionate amount of time in LMs.”
From the episode
‘My Data Proves SEO is NOT Dead’ + How to Rank #1 on Google & AI