Stage 2 · Programmatic SEO

One page for every query, built from one template

We list the queries your customers actually search, by location, service or product, and generate hundreds of pages at once from a single template filled with real data for each combination. Customers search with specifics, like “kitchen remodel cost for a 3-bedroom in Austin,” yet most sites have only a handful of pages that answer. The pages go on a subdomain separate from your main brand, so they are evaluated on their own.

  1. 1Website migration
  2. 2Programmatic SEO
  3. 3Citable documents

The three stages are a sequence, not a menu. Programmatic SEO builds on a finished Website migration.

Two or more of these means it’s time to scale out

  • Large competitors seem to own every search term in your industry
  • You know you need a page per location or product, but hundreds of pages sounds out of reach
  • Every blog post you write targets one query at most

53 sites published in the same sequence all appeared in search after an average of 33 days (Naver Search Advisor)

What we do

How the pages get built

We work in order: each step finishes before the next begins, and you can also start with a single step.

  • 01

    Pick from four combination axes

    Service, location, condition and situation. We use one axis or multiply two. Each of our 5 first-party case studies was built this way, from a single axis to two multiplied together. We only choose axes that show demand signals.

  • 02

    Check that each combination has its own data

    300 pages that only swap the city name look like one page to search engines. A page becomes distinct when it carries real data for that place, so we open only as many pages as we have data for.

  • 03

    Hundreds of pages from one template

    Instead of designing every house, we draw one blueprint. One blueprint puts up 300 houses at the same quality, and only the lot conditions change. Update the design once and all 300 pages change together.

  • 04

    Built apart from your main site

    If hundreds of pages sit on your main domain and one goes wrong, your main site is evaluated along with it. So we build on a separate subdomain. Search engines treat it as its own site, which also means it earns trust from scratch.

Process

How we build, in four steps

Build the query list, collect the data, design the template and publish by cluster. Before kickoff, you can see where the data goes and where the process pauses.

  1. 01

    List the entities

    Take what people add to their searches, such as field, need or industry, as axes and multiply them into a query table. The number of rows is the number of pages to build.

  2. 02

    Keep only combinations that can be filled

    Combinations are only distinct when the values filling them differ. Any combination that can’t fill three data points is filtered out here and stays unpublished.

  3. 03

    Design the template

    Design one page and it unfolds into as many pages as the list has rows. Structured data and internal links live in the template, so no page needs fixing by hand.

  4. 04

    Publish by cluster and check indexing

    We publish cluster by cluster and check indexing status in between. If the previous cluster’s indexing stalls, we fix that cluster before moving to the next.

Standards

Four standards every page must meet

What decides the outcome is less how many pages you build than which pages you hold back. Here are the thresholds we use and what we do when a page misses one.

3

Three data points unique to every page

Pages that only swap a name look identical to search engines. So the template reserves three data slots, and only combinations that fill all three go live.

If the three slots can’t be filled, that combination stays unpublished. We choose fewer pages.

3 clicks

Every page within three clicks of the home page

Pages buried deep get crawled late or not at all. Hub pages in between keep every page within three clicks.

Past three clicks, we add another hub page to reduce the depth.

4–6

Links to sibling pages on the same level

Each page links to its neighbors on the same axis. People move on to the next option, and crawlers find sibling pages without going back through a hub.

If a page can’t get four siblings, the axis is split too finely, and we regroup it.

Clusters

Publish in clusters

Publishing hundreds of pages in a single day backs up the crawl schedule. We publish cluster by cluster and check indexing in between.

If the previous cluster’s indexing stalls, we fix that cluster before publishing the next.

ExampleEvery page is reachable within three clicks of the home page (click depth 3). The lines on the side are internal links to sibling pages on the same axis
1 click2 clicks3 clicksHome pageGwacheonSanbonGunpoBundangSeongnamGimpo

Terms

The terms behind this service

The words we use, next to the industry terms. Seeing them side by side makes it clear what you are buying.

TermWhat it means
Entity-based design and expansionOur name for the full process: list the queries first, then build a page for each one.
Entity SEOGetting search engines to identify a brand, place or concept as one distinct entity.
Programmatic SEOGenerating many pages at once from one template and a data list.
Keyword clusterThe unit we build in: one axis that splits queries (a whole region). An entity, by contrast, is a single item (one neighborhood).
Index coverageHow many of the pages you submitted a search engine has indexed.
Crawled - currently not indexedA quality judgment rather than a technical error. It changes when the page improves, and resubmitting leaves it where it is.

Policy line

Where scaled page building crosses Google’s line

This is how Google’s spam policies draw the line, in the policies’ own terms. The four standards above exist to stay clearly on the right side of it.

Scaled content abuse

Primary purpose

The test Google applies to pages generated at scale. The policy looks at the primary purpose, meaning what the pages were made for, rather than the page count or how they were produced.

Doorway abuse

Similar search queries

The condition for doorway abuse: pages made to target similar search queries. When each page answers a different question, it falls outside that definition.

Google Search updates in 2026TypeStartEnd
February 2026 Discover updateDiscoverFeb 5, 2026Feb 27, 2026
March 2026 spam updateSpamMar 24, 2026Mar 25, 2026
March 2026 core updateCoreMar 27, 2026Apr 8, 2026
May 2026 core updateCoreMay 21, 2026Jun 2, 2026
June 2026 spam updateSpamJun 24, 2026Jun 26, 2026
August 2026 spam updateSpamAug 18, 2026Aug 20, 2026

Sources: Google Search spam policies, reviewed directly on Aug 24, 2026 · Update calendar: Google Search Status Dashboard (status.search.google.com), reviewed directly on Aug 24, 2026

Fit

When we recommend skipping Stage 2

Building more pages isn’t always the right move. If any of these apply, we’ll tell you first and point you to a better route.

You have fewer than ten target queries

Designing a template costs more than writing a few pages by hand.

Going straight from Stage 1 to Stage 3 is faster.

Every query would carry the same data

If different locations or products fill in the same values, you get the same page many times over.

We find an axis where the values differ first, then scale.

Your site is still built one page at a time

Without pages generated from data, there is nothing to scale from.

We do Stage 1 first, then scale on top of it.

You need results within two weeks

In our first-party data, reaching search impressions took 33 days on average.

If that timeline doesn’t fit, another approach will serve you better, and we’ll say so up front.

Preview

What these standards look like in practice

Described in words, this can sound like mass-producing thin pages. Here is how we count queries and how one template produces pages that genuinely differ.

Original dataThe head term “hair salon” (미용실) multiplied by two axes gives 30 cells. We searched 240 queries from our first-party data for ones that earned impressions in each cell, and 16 cells were filled
Areamen’swomen’spermcutrecommended
Gwacheon과천 근처 남자머리 잘하는 미용실···과천 미용실 추천
Sanbon산본 남자 미용실산본역여자컷잘히는곳산본 여자펌산본 단발컷 미용실 추천·
Gunpo군포 남자 미용실··군포 남자 컷·
Bundang분당 남자 미용실·분당 남자 펌 싼것분당 쇼컷트 잘하는 미용실·
Seongnam성남시 남자컷 싼대·성남 히피펌 미용실경기도 성남태평동리프컷·
Gimpo···김포 레이어드컷김포 고촌 미용실추천
ExampleThere is one template, and only the variables marked below change from page to page. The query on top of each page is a filled combination from the grid above (first-party data); the filled values are examples
TemplateEvery page uses this

{Area} {Modifier} hair salon

The title, URL and meta description are generated from these variables
  • {Area}Row variable
  • {Modifier}Column variable
  • {Unique data 1}Data specific to that area
  • {Unique data 2}Data specific to that condition
  • {Unique data 3}Data that exists only for that area × condition combination

Edit the template once and every generated page updates. There is one place to make changes.

Pages generated from the same template

산본 남자 미용실

Men’s hair salons in Sanbon

Each page answers this one query
  • Area coveredSanbon-dong · around Sanbon Station
  • Services for this conditionMen’s cuts · down perms · iron perms
  • Question the page answersIs there a salon that takes walk-ins?
분당 남자 펌 싼것

Men’s perms in Bundang

Each page answers this one query
  • Area coveredJeongja · Seohyeon · Yatap
  • Services for this conditionDown perm · hippie perm · volume perm
  • Question the page answersWhat drives the price?
김포 고촌 미용실추천

Hair salons in Gochon, Gimpo

Each page answers this one query
  • Area coveredGochon-eup · Hyangsan-ri
  • Services for this conditionCut · color · treatment
  • Question the page answersWhat should you look for when choosing?

The queries in the grid above are real searches, shown verbatim in Korean rather than written by us. They were counted within the top 30 rows of each site’s query report, though, so an empty cell doesn’t mean there is no search demand. On a real project, we start by rebuilding this grid for your industry.

Deliverables

What you have after the build

These are the exact deliverables you receive. What you get, how many and to what spec are all settled before kickoff.

  • The list of chosen entities and the reasons they were chosen
  • Data filled in for each combination, with sources
  • One keyword cluster · one subdomain · 300–500 pages · built to all four page standards
  • Registration status by search engine and indexing progress after launch

Scope

What programmatic SEO includes

We open one keyword cluster (300–500 pages) first and decide the next step from its indexing. The prices below are for one cluster; taking several clusters at once costs less than adding them one by one.

  • Entity-based design

    One keyword cluster · 300–500 pages · includes template design

    $660

  • Entity-based expansion

    One more keyword cluster · 300–500 pages · reuses the first template

    $510

  • Enterprise

    Page count and data sources sized together with you

    Custom quote

Questions

Common questions before you start

If your question isn’t answered here, send it to us as it is and we’ll answer it the same way.

What does entity-based design and expansion mean?
An entity is something a search engine recognizes as one distinct thing, such as a neighborhood or a treatment. We build the list of entities first, then create a separate page for each row. Designing the entities comes first and the page expansion follows from it. The industry calls the expansion half programmatic SEO; the difference is the order. Entities and data are settled before any page is generated.
Won’t search engines treat this as spam?
They will if you mass-produce thin pages. Google’s spam policies judge scaled content by its primary purpose, not its page count: whether it was made mainly to manipulate search rankings. The doorway abuse policy looks at whether pages target similar search queries. So every page answers a different question, and we only open queries where three data points unique to that page can be filled. Three hundred data rows make three hundred pages.
Isn’t this just AI-generated filler?
The values come from collected data, and the writing follows a spec. Rotating the same sentences through different combinations would turn the whole set into duplicates. So before publishing, we count whether every page actually carries different values.
When will the pages start showing up?
Getting indexed and getting impressions are different things. The first is something we submit and verify; the second is up to the search engine. Our case studies record how many days the same process took, so use that as your reference range. With every cluster we publish, we check that indexing is progressing, and if it stalls we fix that cluster before publishing the next.
Our business only covers a few locations. Does this still work?
Location is just one axis. Service type, condition, situation and comparison targets all work as axes. What matters is not how many axes you have but whether people actually search those combinations.
Why a subdomain instead of our main domain?
Mass pages on your main domain get your main site evaluated along with them. A subdomain keeps that risk separate. In exchange, it starts without your main site’s authority, so this stage stands on its own from day one.
How many pages do we start with?
Start with one keyword cluster (300–500 pages), then expand. Confirm that the first cluster gets indexed, and let those numbers decide the next step.

Next stages

What comes after the build

Each stage pays off once the one before it is done. You can hire us for a single stage, and we start by checking where you are now.

  • 01

    Website migration

    We move hand-built websites to a data-driven structure, with body text in the HTML crawlers receive and sitemaps and structured data built in.

  • 02You are here

    Programmatic SEO

    We list the queries customers actually search and build one data-backed page per query from a single template, starting with one cluster of 300–500 pages.

  • 03
    Cited

    Citable documents

    Citable documents with one direct answer per question, submitted to search engines and placed where they can be verified outside your own site.

We start by mapping which query axes are open in your industry

We open your site, check what is indexed and which queries it appears for, and reply with what to fix first.

Free audit