# Product led SEO

> Turn product data and free tools into pages that rank and convert: opportunity sizing, three build models, engineering cost, and conversion instrumentation.

Source: https://saas-marketing.net/guides/product-led-seo/
Topic: SaaS SEO
Type: guide
Published: 2026-09-11
Last updated: 2026-09-11
Publisher: SaaS Marketing (saas-marketing.net)
License: CC BY 4.0. Quote or republish with attribution and a link to https://saas-marketing.net/guides/product-led-seo/

## Short answer

Product led SEO builds pages out of product capability or product data rather than out of articles about the product. A free calculator, an integration directory generated from your app graph, or public pages rendered from in-product objects are all examples. It differs from product led content, which is blog writing that mentions the product. The distinction matters because a language model can write your blog cluster in an afternoon and cannot reproduce your data.

## Key takeaways

- Product led SEO builds pages from product capability or data. Product led content is writing that mentions the product.
- The defensible asset is data or function a competitor cannot copy by writing, which rules out most blog strategies.
- Three build models cover almost every case: free standalone tool, data directory, and indexable in-product objects.
- Ahrefs free tools and HubSpot Website Grader needed engineering. Canva's template gallery is user generated supply.
- Instrument tool use to signup as a funnel, or you will kill a working asset because the blog attribution model cannot see it.
- Every product led page set carries permanent maintenance cost, and abandoned tools decay faster than abandoned articles.

---

An LLM can write every blog post in your topic cluster in an afternoon, and so can your competitor's intern using one. Nobody can write your integration graph, your aggregated usage data, or a tool that actually performs the calculation. That asymmetry is the entire argument for product led SEO, and it has got sharper every quarter since 2023.

The term comes from Eli Schwartz's 2021 book, and Kevin Indig has done most of the practical work since on how these programmes get built and measured. The definition has since been muddied by people using it to mean blog posts that mention the product, which is a different thing with different economics.

## Product led SEO and product led content are not the same thing

Product led content is editorial: an article that teaches something and demonstrates the product while doing it. Product led SEO is structural: the page is the product, or is generated from its data. One is a writing discipline. The other needs engineering.

Ahrefs is usually cited as the canonical example of both, which is why people conflate them. The Ahrefs blog is product led content, and it's excellent. The free Backlink Checker, Keyword Generator and Website Authority Checker are product led SEO, because they answer the query by running the job against Ahrefs' index rather than explaining how to do it. The second set is the one a competitor cannot replicate with a content budget, because replicating it means building a web-scale crawler first. The full breakdown sits in our [Ahrefs product led SEO teardown](/examples/ahrefs-product-led-seo/).

That's the filter to apply to any idea in this space: if a well-funded competitor could match it by hiring two writers, it isn't a moat. It might still be worth building. Just be honest in the business case about what you're buying.

## Finding the product data that maps to existing demand

Start with an inventory, not with keyword research. List every object your product creates, stores or connects: integrations, templates, file formats, currencies, job titles, regulations, locations, metrics, schemas, categories, benchmarks. Then find which of those objects people already search for.

**Opportunity sizing for product led pages**

The demand check is where most ideas should die, and they usually don't because the idea is exciting. Sample the actual combinations, not the head term. "Project management templates" has volume. "Project management template for a mid-market construction subcontractor" does not, and a set built on the second pattern produces eight hundred pages that nobody searches for and Google declines to index. That's the failure mode documented in our [programmatic page indexation study](/research/programmatic-page-indexation-study/), and it is far more common than the success case.

Eight thousand apps in a directory implies roughly 32 million ordered pairs. Zapier does not publish a viable page for every pair, because most pairs have no demand and no users. The count is the ceiling, never the plan.

## The three build models, and which ones needed a product team

Almost every product led SEO programme is one of three shapes. They have very different costs and very different odds.

| Model | What it is | Real examples | Engineering cost |
| --- | --- | --- | --- |
| Free standalone tool | A utility that performs a job, gated or ungated | Ahrefs free SEO tools, HubSpot Website Grader, Shopify business name generator | High. A real build with a real backend |
| Data directory | Pages generated from a dataset you own or aggregate | Zapier app and integration pages, G2 category and comparison pages | Medium. Templates plus data pipeline |
| Indexable in-product objects | Public rendering of things users create in the app | Canva template gallery, public Miro templates, public Typeform examples | Medium to high. Privacy and permissions dominate |

Ahrefs' free tools and HubSpot's Website Grader both needed genuine product engineering, because both run real analysis against a real index. Website Grader has been running since 2007 and remains one of the longest-lived examples of the model. These are not marketing site projects. They sit on infrastructure, they have uptime expectations, and they need an owner in engineering.

Zapier's directory is the cheapest of the three in engineering terms and the hardest in editorial terms. The data already existed in the product: every app in the directory, every trigger and action, every popular workflow between pairs. What made it work was demand-matched page selection and internal linking, not the generation itself. If you build a directory, budget most of your time for the [internal linking](/guides/saas-internal-linking/) architecture and the indexation controls, because that is where these sets succeed or quietly fail.

Canva's template gallery is the model people most want and least often can have. The supply is user generated and the SEO value comes from the volume and variety of designs people make in the product. You cannot bootstrap that. If your product does not already create public-shaped artefacts at volume, this model is not available to you this year, and pretending otherwise produces a gallery of forty in-house templates that nobody links to.

You do not need Zapier's scale. A 60-page set covering "connect [your product] to [each of your 60 integrations]", each page carrying the real setup steps, the actual field mappings and a working screenshot, is a legitimate product led set. It is also the one most B2B SaaS companies should build first.

## Scoping the build with engineering without losing the argument

Engineering teams reject SEO requests because the requests are usually vague and the payoff is usually unfalsifiable. Fix both before you walk into the planning meeting.

Write the ask as a product spec. Name the data source and the exact fields. Name the number of pages in batch one. Name the template, including which parts are generated and which parts a human writes. Name the success metric and the date you'll evaluate it. Name what you'll do if it fails, which should be "we delete the set", and mean it.

Then make the first ask small. One template, 30 pages, four weeks. The point of batch one is not traffic, it's evidence: do these pages get indexed, do they rank, do the sessions convert. Bring that evidence back and the second ask is easy. Ask for a 5,000-page build up front and you'll get a polite no, deservedly.

Two scoping details that cause most of the rework:

- Decide who writes the unique content per page and when. Every viable template needs some genuinely unique input, and a plan that says "we'll add unique content later" means a set of near-duplicates goes live first and poisons the whole directory's reputation with the crawler.
- Decide the index policy per variant before launch, not after. Which combinations get an index tag, which get noindex, how a page graduates from one to the other when real data arrives. The rollout mechanics in [launching a programmatic page set](/playbooks/programmatic-page-launch/) cover the batch sizing and sitemap segmentation that go with it.

For enterprise organisations where the marketing site is owned by a different team than the app, add a quarter to every estimate and involve security review early, particularly for anything exposing in-product objects. The realities of that environment are in [enterprise SaaS SEO](/playbooks/enterprise-saas-seo/).

## Instrumenting the path from tool use to signup

Product led pages break most analytics setups, because the valuable event happens on the page rather than after a click to a form. Measure the tool's own funnel and stop comparing it to the blog.

Four events, tracked as a sequence:

1. Session on the tool or generated page.
2. Tool completion, meaning the user actually ran the thing. For a calculator, a result rendered. For an integration page, a click on the setup CTA.
3. Account creation attributed to that session, with the tool identifier stored on the user record.
4. Conversion to paid, reported against the same identifier at 30, 60 and 90 days.

Store the tool identifier on the account, not only in the analytics session. This is the single most useful thing you can do, because it survives the attribution window and lets you answer the only question that matters in the renewal review: what is a customer acquired through this tool worth, compared to one acquired any other way.

Expect a specific and counterintuitive pattern. Free tools frequently show poor session-to-signup rates, often well under 1 percent, and strong signup-to-paid rates, because the people who bother to run an analysis tool have the problem the product solves. Blended reporting hides both halves, and teams kill working assets on the strength of a low top-line conversion rate. Report the full chain or don't report it at all.

Gate carefully. Requiring an email before showing any result usually raises immediate signups and lowers links, rankings and repeat use, which is a bad trade for an asset whose value is compounding. The pattern that generally holds up: show the result, gate the export, the saved version, the historical comparison or the full dataset. That's also how these pages earn the links documented in our [free tool link magnets](/examples/free-tool-link-magnets/) collection.

## What good looks like at 90 days

Set the evaluation bar before launch, because after launch everyone negotiates it. These are the numbers I use to decide whether batch one earns batch two, based on B2B SaaS sets in the 30 to 500 page range.

| Signal | Weak | Working | Strong |
| --- | --- | --- | --- |
| Pages indexed at day 30 | Under 50% | 70 to 85% | Above 90% |
| Pages with any impressions at day 60 | Under 40% | 60 to 75% | Above 85% |
| Median position at day 90 for target queries | Outside 30 | 11 to 25 | Top 10 |
| Session to tool completion | Under 20% | 35 to 55% | Above 60% |
| Tool completion to account creation | Under 2% | 4 to 8% | Above 10% |
| Referring domains earned per 100 pages | 0 to 2 | 5 to 15 | 25+ |

Read the rows in order, because they fail in order. Low indexation at day 30 is a template and internal linking problem and no amount of waiting fixes it. Good indexation with no impressions means the demand check was wrong and the pages target phrases nobody types. Good impressions with a terrible completion rate means the page is ranking for an intent it does not serve, which happens constantly with calculators that rank for a definition query.

The referring domains row separates the two asset types cleanly. A genuine free tool that does real work should attract links without outreach, and if it earns nothing in three months you built a page rather than a tool. A generated directory set rarely earns links at all, which is fine, because its job is coverage and internal link equity rather than authority.

Generated page sets commonly sit flat for eight to twelve weeks and then move together once the crawler has processed the cluster. Judge indexation at 30 days and rankings at 90, but do not delete a set at day 45 because the chart is flat.

## Maintenance cost, and the ones that rot

Every product led asset carries permanent cost, and this is where the honest accounting happens. A blog post that goes stale is mildly embarrassing. A free tool that returns wrong numbers, or an integration page describing a setup flow you shipped a redesign over eight months ago, actively damages trust with the exact buyers you were trying to reach.

Budget for it explicitly:

- An owner in engineering, named, with the tool in their service catalogue.
- A quarterly accuracy review for anything producing numbers or instructions.
- A hook in the release process so a product change flags the pages describing it.
- A decay check on the generated set: which pages have lost impressions, which have never been indexed, which should be consolidated or removed.

The abandoned free tool is a genuinely common outcome. A marketing team ships something well, the champion leaves, nobody owns it, and two years later it's returning stale data with a broken layout on mobile, still ranking, still the first impression thousands of buyers get. If you cannot name the person who will own it in year three, build something smaller.

## When product led SEO is the wrong call

Three situations where I'd tell you not to do this.

You have no bottom of funnel pages yet. If you don't have comparison, alternatives and use case pages ranking, a free tool is an expensive detour. Build the pages that catch buyers who are already shopping, using something like the [bottom of funnel page sprint](/courses/saas-seo-sprint/02-bofu-pages/), and revisit this in two quarters.

You have no proprietary data and no engineering capacity. A calculator built from public formulas that ten competitors also publish is not a moat, it's a Tuesday. It can still be worth building as a link asset, but call it that in the business case rather than dressing it up as strategy.

Your buyers want a vendor, not a utility. In enterprise security, regulated finance and healthcare, the purchase runs through procurement and risk review, and a free tool rarely intersects that path. Content that helps a champion build an internal case does more. The choice between templated page sets and editorial is worked through properly in [programmatic versus editorial content](/comparisons/programmatic-vs-editorial-content/), and the honest answer for many companies is editorial first.

## What to do next

Run the object inventory this week. It takes two hours with a product manager and a whiteboard, and it either produces three credible candidates or it tells you plainly that you don't have the raw material yet, which is a useful answer and a cheap one.

If you do have candidates, sample the demand before you write a spec, then scope one template and 30 pages with a named success metric and a kill date. Ship it, measure the four-step funnel, and bring the evidence back before asking for the rest. The production mechanics live in [programmatic SEO for SaaS](/guides/programmatic-seo-for-saas/), and where this fits against the rest of the programme is laid out in the [SaaS SEO](/saas-seo/) hub. Build the thing a competitor would need an engineering team to copy, and keep it working.

## Frequently asked questions

### What is product led SEO?

Product led SEO is the practice of building search landing pages out of product capability or product data rather than out of written content about the product. A free tool that performs a job, a directory generated from your integration graph, or public pages rendered from objects users create in your app all qualify. Eli Schwartz named the approach in his 2021 book of the same title.

### How is product led SEO different from product led content?

Product led content is editorial writing that demonstrates the product inside a useful article, which is what Ahrefs and Zapier are best known for in their blogs. Product led SEO produces pages that are the product, or built from its data. The first is a writing discipline and any competitor can copy it. The second requires engineering and data a competitor does not have.

### Do free tools still work for SEO in 2026?

Yes, better than most content does, because a tool answers a query by doing the job rather than describing it. Free tools also attract links at rates editorial rarely matches. The trade is cost: a real tool takes weeks of engineering and never stops needing maintenance, and a broken free tool damages trust in a way a stale blog post does not.

### How long does a product led SEO build take?

A single free calculator with no backend typically takes two to four weeks including design and QA. A directory generated from existing product data takes four to eight weeks, most of it spent on templates, internal linking and indexation controls rather than on the data itself. Exposing in-product objects publicly is the longest build because of privacy and permission work.

### How do you measure whether a free tool is working?

Instrument four steps: page sessions, tool completions, account creations from the tool, and paid conversions from those accounts. Report the tool's own funnel rather than blended blog metrics. Tools often show low session-to-signup rates and excellent signup-to-paid rates, which looks like failure in an aggregate report and is actually the opposite.

### Is programmatic SEO the same as product led SEO?

They overlap but are not the same. Programmatic SEO is a production method: generate many pages from a template and a dataset. Product led SEO is a strategy about what the pages are made of. A programmatic set built from scraped third party data is programmatic but not product led, and a single free tool is product led but not programmatic.

### When is product led SEO the wrong choice?

When you have no proprietary data, no engineering capacity, and no ranking bottom of funnel pages yet. A company with ten indexed pages and no comparison content should build those first. Product led SEO also fails in categories where the searcher wants a vendor rather than a utility, which is common in enterprise security and regulated industries.