Search Engine Optimisation
Programmatic SEO
The dividing line is data. If every page differs only by a name, this is the wrong service and we will say so before quoting.
Programmatic SEO generates pages at scale from structured data — locations, products, comparisons. The Nexclick builds them only where each page has genuinely different data behind it, because a template with a variable swapped in is a doorway page with extra engineering.
Is this you?
What usually prompts the call
- You have a structured dataset — products, routes, specifications, coverage — and no pages built from it.
- Competitors rank for hundreds of long-tail variations you have no pages for.
- You generated pages at scale previously and traffic fell afterwards.
- Your catalogue supports thousands of genuine comparisons and you are publishing manually.
What we do
The actual deliverables
Things that appear on an invoice, not adjectives.
- Data viability assessment first
- Whether your dataset genuinely differentiates each page. This assessment ends roughly a third of enquiries before any building starts, which is the correct outcome.
- Demand validation before generation
- Confirming search demand exists for the pattern. Generating 4,000 pages for queries nobody makes produces index bloat, not traffic.
- Template design with real variation
- Templates built so the data genuinely drives the content — different figures, different comparisons, different recommendations — rather than a fixed paragraph with a variable inserted.
- Quality thresholds enforced in code
- A page only generates when the underlying data meets a defined completeness bar. Everything else is excluded rather than published thin.
- Internal linking architecture
- Hub pages, sensible cross-links and a crawl path that reaches deep pages within three clicks. Generated pages nothing links to will not be indexed.
- Staged rollout
- Released in batches with indexation and ranking monitored between them. Publishing 5,000 pages in one deployment gives you no way to tell what worked.
- Pruning built into the process
- Pages that gain no impressions after a defined period are removed. Programmatic sets need pruning as much as generation, and almost nobody plans for it.
Comparison
When programmatic works, and when it is a doorway page
The distinction that decides whether this technique earns traffic or damages the site. Real examples of both, with what makes the difference.
| Page pattern | Verdict | What decides it |
|---|---|---|
| Product specification pages from a real catalogue | Works | Every page has genuinely different data |
| Route or journey pages with real timetable data | Works | Times, prices and connections differ per page |
| "X vs Y" comparisons from structured attributes | Works | The comparison itself is unique to each pair |
| Property or vehicle listings | Works | Each record is genuinely distinct |
| Job listings by role and location | Works | Real vacancies, real data |
| Statistics pages from a proprietary dataset | Works | Original data nobody else has |
| "[Service] in [Town]" with only the town changing | Doorway | No differentiating data — the classic failure |
| "Best [product] for [use case]" with generic text | Doorway | Same recommendations, different heading |
| Pages generated for keyword variants of one topic | Doorway | One page would serve all of them better |
| AI-written pages from a topic list | Doorway | No underlying data at all, just generated prose |
| Currency or unit conversion pages | Borderline | Works only with genuine calculation and demand |
How it works
Step by step, with timeframes
Timeframes are typical rather than guaranteed, and they assume we get account access and approvals when we ask.
- 01Week 1–2
Assess the data
Whether the dataset can carry genuinely distinct pages, and whether demand exists. Both have to be true or the answer is no.
- 02Week 2–4
Prototype a small set
Twenty to fifty pages built and published, then monitored. This tells you more than any amount of planning.
- 03Month 2–4
Scale in batches
Released in tranches with indexation checked between each. Scaling stops if the prototype set is not being indexed.
- 04Ongoing, quarterly
Prune and refine
Pages with no impressions after a defined window removed, and templates improved from what the performing pages have in common.
What you get
Reporting and ownership
- A written viability assessment before anything is built, including when the answer is not to build it.
- A template with quality thresholds enforced in code, so thin pages cannot generate.
- Staged rollout with indexation data at each stage, rather than one deployment and hope.
- A pruning process documented and scheduled, not left as an intention.
- The generation pipeline handed over so your team can extend it.
Tools and platforms
- Google Search Console (bulk export)
- Screaming Frog
- Next.js or equivalent static generation
- Structured data pipelines
- Ahrefs or Semrush
- BigQuery for large-scale analysis
Timeline
How long this actually takes
Assessment takes two weeks, the prototype set another two, and full rollout runs over two to four months in batches. Indexation is the constraint that surprises people: Google indexes large generated sets slowly and selectively, and a substantial share of pages may never be indexed regardless of quality. Expect three to six months before the pattern is readable. The blunt warning: this technique has a bad reputation because it is usually done badly, and doing it badly can damage the whole site rather than just the generated section. If your data does not genuinely differentiate the pages, the correct answer is not to do it.
Pricing model
Fixed-price project
Fixed-price assessment first, quoted small so you can get an honest answer cheaply. Build is quoted separately once viability is established.
Questions
Programmatic SEO questions
Is programmatic SEO against Google’s guidelines?
Generating pages at scale is not, in itself. Google’s policies target scaled content abuse — pages produced primarily to manipulate rankings with no value to a reader. A generated page backed by genuinely distinct data serving a real query is fine. The technique is neutral; the execution decides it.
How do we know if our data is good enough?
The test is whether a reader would find each page meaningfully different from its neighbours. If the only difference is a name in a heading, it is not. We run this assessment before quoting a build, and a fair proportion of the time the answer is that it should not be built.
How many pages will actually get indexed?
Fewer than you generate, often considerably fewer. Google indexes large generated sets selectively, weighing perceived value and crawl budget. Half being indexed is a reasonable outcome on a first rollout, which is why staged release and pruning matter more than volume.
Can we use AI to write the content for these pages?
For assembling data into readable sentences, yes. For generating the substance itself, no — that produces pages with nothing behind them, which is exactly the pattern Google’s scaled content policies target. The data must exist first; AI can present it.
What happened to sites that did this badly?
Scaled content abuse enforcement has removed large generated sections from the index, and in some cases affected whole sites. The risk is not confined to the generated pages, which is the important part — a bad programmatic rollout can drag down content that was performing perfectly well.
How is this different from location pages?
Location pages are a specific application, and the most commonly abused one. Our location page service exists separately precisely because the honest answer there is usually "build five good ones", not "generate two hundred". Same principle, applied to the case where it most often goes wrong.
Last reviewed 28 July 2026.
Tell us what you are trying to fix
A 20-minute call, no pitch deck. The Nexclick will tell you what we would do, roughly what it costs, and whether we are the right people for it.