Content and search
Content earns its place by answering a question a buyer is asking, nearest to the purchase first. Plan clusters, write for the query, link new posts from old ones, and refresh what decays.
54 steps, about 178 hours of work · 40 plays
SaaS SEO is the work of making a software product the page Google shows when a buyer searches for the problem it solves. For a new site, that means passing the technical checks, getting indexed, building comparison and use-case pages first, briefing each post from the live results page, and earning the links its authority lacks.
The concept, explained in full: SEO.
Do it
Every step in the free launch plan that this topic needs, in the order you need them. Each keeps its plan number.
- How do you see what Google sees?3 steps
- What does Google need before anything can rank?13 steps
- How do you get every page indexed?7 steps
- Which pages should a new site build first?9 steps
- How do you choose and brief each post?10 steps
- How do you grow the queries Google already shows you for?8 steps
- How do you get cited by AI search as well?20 steps
- Where do the links come from?67 steps
- How do you scale it once it works?27 steps
- How do you own the category term?29 steps
How do you see what Google sees?
Two free consoles and one report. Without them every decision below is a guess.
1.1.1.5Connect Google Search Console
The only free source of truth for what you rank for and what Google refuses to index.
1.1.1.6Connect Bing Webmaster Tools
One console for the second-largest reach you have: Bing’s index also serves Yahoo, DuckDuckGo, Copilot and most of Ecosia. It no longer feeds ChatGPT, which runs its own crawler — but it is the only console that reports your Copilot citations.
Done when: Open Bing Webmaster Tools and confirm the site is a verified property. The URL of the property page is the evidence.
4.1.6Attribute signups to the page that earned them, not just the channel
Sessions rank your posts one way and signups rank them another. Carry it through to activation and retention too — pieces that pull the wrong reader show up as churn months later, not as a bad traffic number.
2h
What does Google need before anything can rank?
A short technical baseline, done once. Past it, a deep audit can wait until the site is large.
1.3.1.1One H1 per page, sequential heading hierarchy
Sequential headings appear on 68.7% of ChatGPT-cited pages against 23.9% of Google’s top results, and 87% of cited pages have a single H1 (AirOps, 2025). It is a correlation, not proof that headings cause citation.
1.3.1.2Generate sitemap.xml automatically
Hand-maintained sitemaps rot within a month. Generate it from your routes.
30m
1.3.1.3Allow crawling in robots.txt and name the sitemap
A staging Disallow shipped to production is the single most common launch-week disaster.
10m
1.3.1.4Add canonical tags to every page
You will syndicate to Medium and Dev.to in Phase 2. Canonicals are what stop that hurting.
30m
1.3.1.5Add OG and Twitter card images
Every share for the next two years renders through these. Make them readable at thumbnail size.
45m
1.3.1.6Pass Core Web Vitals on mobile
Test on a throttled 4G profile, not your laptop.
1.3.1.7Read your own page the way a crawler does
Fetch it with JavaScript disabled: one audited site was serving hundreds of Lorem Ipsum pages to crawlers while looking finished in a browser.
30m
1.3.1.8Confirm Google has indexed every page that takes money
Homepage, pricing, signup and your best feature page. A page Google has not indexed cannot rank and cannot appear in an AI Overview, and a nosnippet tag keeps it out of the AI answers even when it is indexed. Until these four pass, a deeper technical audit can wait.
Done when: Run URL inspection in Search Console on the homepage, pricing page, signup page and one feature page. For each, say whether it reads "URL is on Google" and whether the page carries a nosnippet or max-snippet:0 directive.
1.3.2.1Add Organization schema
Name, logo, URL, sameAs to every social profile. This is what entity resolution keys on.
30m
1.3.2.2Add SoftwareApplication or Product schema
Category, operating system, price. Review sites and AI answers both read it.
30m
1.3.2.3Add FAQPage schema to the homepage and pricing page
The highest-leverage schema for answer extraction, and you already wrote the FAQs.
30m
1.2.3.2Add an FAQ block with FAQPage schema
Kills the top objections and feeds answer extraction in AI search at the same time.
45m
How do you get every page indexed?
A page Google has not indexed cannot rank, so indexing is a number to watch, not a box to tick.
3.3.1Submit the sitemap to Google and Bing
Then check the coverage report for what was excluded and why.
Done when: Open Search Console and Bing Webmaster Tools and confirm the sitemap is submitted in each, then read the coverage report for what was excluded and why. The exclusions are the point — a green tick with 40 pages excluded is not done.
15m
3.3.2Set up IndexNow
Push new URLs on publish instead of waiting for a crawl. One ping reaches Bing, Yandex, Naver, Seznam and Yep — and, through Bing, Yahoo, DuckDuckGo and Copilot.
Done when: Fetch the key file the site serves at its published path and confirm it matches the key the app submits with. Then show the submitter has actually run — a log line, or a recorded submission. A key file nobody pings is decoration.
45m
3.3.3Make every page reachable in three clicks
Orphan pages do not get crawled. Internal links are the cheapest ranking factor you control.
Done when: Walk internal links outward from the homepage and count how many indexable pages are reached within three hops. Every page in the sitemap must be among them. Report the number reached and the number of pages that were never reached — and retry a failed fetch before calling it unreachable, because a timeout looks exactly like an orphan.
2h
3.3.4Clear the crawl errors in Search Console
Especially “Discovered — currently not indexed”, which usually means thin or duplicate.
2h
3.3.5Give every page four links from the body of other pages
Nav and footer links sit on every page and separate nothing. Links written into the prose are the ones that carry — and vary the words you link on, because the same anchor repeated reads as a template. Read each anchor inside its own sentence: one that promises something the destination does not contain costs you the click and the reader.
Done when: Count, for each indexable page, how many DISTINCT other pages link to it from their body — strip nav, header and footer first, or every page scores the same and the number means nothing. Report how many pages fall below four. Re-run before believing it: a transient fetch failure reads exactly like a missing link.
2h
3.3.6Track indexed as a share of submitted, not as a count
A hundred indexed pages reads as progress until you notice you asked for five hundred. One audited site had roughly one sitemap URL in five indexed.
30m
3.3.7Claim the national consoles wherever a local engine competes with Google
Read your country split first, then claim only where buyers are: Yandex (Russia, Turkey, CIS), Naver and Daum (Korea), Seznam (Czechia), Cốc Cốc (Vietnam), and Baidu, then Sogou, 360 and Shenma (China, which also wants Chinese pages and an ICP-filed host). All free.
Done when: Quote the sessions by country from analytics, then for each of Russia, Turkey, Korea, Czechia, Vietnam and China either name the console the site is verified in or state that the country sends no sessions.
45m
Which pages should a new site build first?
The pages buyers search when they are already choosing: alternatives, comparisons, use cases.
3.1.1.1Write /alternatives/[competitor] for each of your top 4
One page per competitor, targeting “[competitor] alternative”. Nothing else ranks this cheaply.
Done when: Open each alternative page and confirm it names a real competitor, states what that competitor charges from the competitor's own pricing page, and says plainly when a reader should choose them. List the pages and the competitor each covers. A page that cannot honestly recommend the competitor to anyone is an ad, not a comparison.
12h
3.1.1.2Add an honest feature comparison table to each
Answer engines cross-reference competitor claims and de-rank pages that lie. Concede real losses.
Done when: Open each alternative page and count the rows of its comparison table, and how many of those rows the competitor wins. Report both numbers for every page. A table you win on every row is not a comparison, and an engine checking it against the competitor's own pages will treat it as one that lies. A table shipped as an image counts as zero rows: an engine cannot read it.
2h
3.1.1.3Add a “when to choose them over us” section
The single highest-trust block on the page, and it disqualifies bad-fit signups for free.
Done when: Quote the passage on one alternative page that tells a reader to choose the competitor, and name the other pages that carry one. It must name a situation in which the competitor is the better pick. General praise of the competitor does not count; neither does a concession buried in a table cell.
1h
3.1.1.4Add FAQ schema to every alternative page
These are the pages you most want quoted in an AI answer. Make them extractable.
Done when: Read the JSON-LD in the source of every alternative page and list each page whose markup holds an FAQPage with at least one Question that carries an answer. An FAQ visible on the page with no FAQPage in the markup does not count.
1h
3.1.1.5Write a “[you] vs [rival]” page for the two rivals buyers name most
A different query from “[rival] alternative”, typed later by a warmer buyer who has already shortlisted you both. One audited site had 73 alternatives pages and no head-to-heads, while comparison pages were its best template. Put the date beside every competitor price you quote; an undated price goes wrong within a year and reads as a lie.
Done when: List each “[product] vs [competitor]” page that is live, and for each say whether every competitor price on it carries the date it was checked. Pages with an undated competitor price do not count.
3h
3.1.1.6When a rival announces it is closing, publish the switch guide that week
Its users now have a deadline and must choose something, so searches for “[rival] alternative” spike and then stop once the doors close. Lead with how to export their data and what maps to what in your product, not with your pitch. It earns traffic only while the deadline is still ahead.
3h
3.1.2.1Ship 3–5 /for/[audience] or /use-cases/[job] pages
Same product, their words, their screenshot, their objection handled.
Done when: Open each audience or use-case page and list it with the audience it names, one phrase from its copy that matches how that audience describes the problem in public, the screenshot it shows, and the objection it answers. A page that swaps the audience name into otherwise identical copy is a template, not a page.
10h
3.1.3Write the best-of list for your own category
Include your competitors honestly. This ranks for category queries and becomes an AI citation source.
Done when: Open your best-of page and list every product it ranks, with its position and one weakness the page names for it. Then say where your own product sits and which published criterion puts it there. A list with you first and no weakness named for you is an ad, and engines read it as one.
3.1.4Ship one free tool or template gallery
Typeform’s template library drove 30k non-branded signups. It is also a permanent link magnet.
Done when: Open each free tool in a browser that is not signed in, change an input, and confirm it returns a result with no signup, no email gate and no account. List every tool URL that passes, with the result you got. A tool behind a form is a lead magnet, not a link magnet, and nobody links to a form.
How do you choose and brief each post?
Pick a pattern a new site can win, read the results page before writing, and add one thing the top three lack.
3.8.1Pick three topic clusters from real search demand
Volume you can plausibly rank for at DR 20 — long-tail, low-difficulty, high-intent. Phrase them the way competitor reviews and threads do; nobody has ever searched the words a marketer would pick.
2h
3.8.13Pick one keyword pattern a new site can win, and work it one row at a time
A pattern is one intent with one changing word: “[rival] alternative”, “[you] vs [rival]”, “[tool] for [audience]”. It turns topic-picking into taking the next row. Start with the least-contested pattern, and give every variant something only it can say, or it is a find-and-replace.
Done when: Name the pattern and list at least five variants, each with its monthly volume and where that number came from. For each variant, say whether any of the top ten Google results has the exact query in its title. A volume with no source does not count.
3.8.14Read the live results page before you brief each post
The top ten tell you the page type Google wants (a comparison, a list, a guide), what the first result also ranks for, and what all of them skip. Search your own site first: a second page on a query you already cover splits the clicks and both lose.
Done when: For the last three posts published, give the target query, the page type of the top three results, and the page type of your post. Then say whether any other page on your site targets the same query. A post whose type differs from all three results fails.
30mhow to do it
3.8.15Give every post one thing the top three results do not have
A screenshot only a user could take, a number from your own product, a test you ran. Google’s spam policies cover unoriginal content made at scale “no matter how it’s created”, and the page that rewords the top three is the one they describe.
Done when: Open your last five posts and name, for each, the one element none of the top three results for its query has: a screenshot, a figure from your own data, a test or a quote from a named person. A post with nothing to name fails.
30mhow to do it
3.8.2Publish ten posts against those clusters
Three or four per cluster beats ten scattered posts. Depth in one place is what ranks.
40h
3.8.4Publish one definitive how-to guide
The page you would send someone who knows nothing. Long, structured, genuinely complete.
12hhow to do it
3.8.5Publish a cost and pricing explainer for your category
“How much does X cost” is bottom-of-funnel traffic that competitors refuse to answer.
3.8.11Harvest People Also Ask under each cluster term for the spoke questions
Click every question open and Google spawns more: that is Google listing the sub-questions it wants answered. One question per page, answered in the first sentence.
45m
3.8.7Link every post to the page that takes money, from inside the prose
Not the nav, not a banner — a sentence in the body where the product is the answer to what the reader just learned. This is the difference between content that ranks and content that pays.
2h
3.8.8Put a real byline on every post, linked to one author page
Do it for the reader and the links, not for rankings — sites with real bylines earn roughly twice the referring domains and show no ranking lift once those links are accounted for.
Done when: Open three posts. For each, say whether the visible byline links to the same author page, whether the Article markup names that author as a Person with that page as its url, and whether the markup’s datePublished matches the date shown on the page. A mismatch on any of the three fails that post.
1h
How do you grow the queries Google already shows you for?
Search Console lists the queries you half-rank for. They are cheaper than any new keyword.
3.8.6.1Pull the query-to-page table for the last 90 days
Query, landing page, impressions, clicks, position. Three months so one freak day cannot set the agenda, and leave off the newest days: Google marks them preliminary, with a dotted line on the chart.
3.8.6.2List the queries whose landing page is about something else
Impressions against a page that only half answers the query. That mismatch is the whole signal.
30m
3.8.6.3Kill the gaps that would cannibalise a page you already have
Check volume and difficulty, read the live SERP for intent, and drop any query an existing page already answers. Splitting one intent across two pages ranks neither.
30m
3.8.6.4Publish one page per surviving query, then re-check its position
Exact query in the title, URL and opening answer. Log position and impressions before the next batch — this loop is only as good as its feedback.
3.8.6.5Push the queries sitting on page two before writing anything new
Position 11–20 with impressions is a page Google already half-trusts. Exact query in the title, H1 and first hundred words, then three internal links to it. An afternoon, and cheaper than any new page.
Done when: For each query you pushed, give its average position with Search Console filtered to that exact query AND that exact page, from before the change and from at least four weeks after. An average across all queries or all pages does not count: it mixes rankings that have nothing to do with each other.
2h
3.8.6.6Rewrite the titles that get seen and not clicked
High impressions and few clicks is a title problem, not a ranking problem. Promise what the query asked for in the title and description, then read the click-through rate again in a month.
1h
3.8.6.7Take your own name out of every Search Console view
On a young site most clicks are people typing the name, which says nothing about whether the content ranks. Google’s branded filter needs a top-level property with enough impressions, so a new site usually filters the query instead: “not containing” the name and its misspellings. Read every step in this group on that view.
Done when: Give the clicks and impressions for the last three months with every query, then again with your name and its misspellings excluded, and say which filter you used. The second view is the one the other steps in this group read.
15m
3.8.6.8When a page-two query will not move, count the links behind the pages above it
Once the title and internal-link push has had four weeks, compare the referring domains to your page with those of each of the top three, page against page, not site against site. A gap you can close in a quarter is a link job for that URL, never your homepage. A gap you cannot close is a reason to pick a smaller query.
Done when: For one query that stayed on page two after the push, give the number of referring domains to your page and to each of the top three results’ pages, with the tool and the date you counted. A count for a whole site rather than the page does not count.
1h
How do you get cited by AI search as well?
AI answers read the same index. Being fetchable and quotable is the part on top of ordinary SEO.
1.3.3.1Publish llms.txt
A plain-text map of what you are and which pages state it. Half an hour, then move on: Google says AI Overviews need no AI text file, so it is cheap insurance for the agents and tools that do ask for one, not a ranking lever.
1.3.3.2Put a plain “what this is” paragraph in the first 100 words
Extraction reads the top of the page. Poetry above the definition costs you the citation.
Done when: Read the first 100 words of the homepage body with no page around them, as an extractor would. Quote them. Do they name the product and say plainly what it is?
30mhow to do it
1.3.3.3Record your AI-citation baseline
Ask ChatGPT, Claude, Perplexity and Google AIO “best [your category] tools”. Log the answer verbatim.
Done when: Ask two AI answer engines a question your category owns, and record whether the product was cited. The dated note of what they said is the evidence.
30m
1.3.3.4Confirm the AI search crawlers can actually fetch you, in the logs
Look for the crawlers that decide citations — OAI-SearchBot, Claude-SearchBot, PerplexityBot, Applebot — not GPTBot or ClaudeBot, which gather training data and say nothing about whether an answer can quote you. Cloudflare blocks AI crawlers by default now, so a site can be invisible without anyone choosing it.
Done when: Count the requests from OAI-SearchBot, Claude-SearchBot and PerplexityBot in the server log that returned 2xx. All three must appear, fetching pages rather than only robots.txt. GPTBot and ClaudeBot do not count: they gather training data, not citations.
20m
1.3.3.5Check that Brave has your key pages, and ask it to re-fetch the missing ones
Brave’s index is what Claude’s web search reads. There is no console and no Brave user agent: it crawls only what your robots.txt lets Googlebot crawl. Its submit form re-fetches a page; it does not promise to index a site it has never seen.
Done when: Search site:yourdomain on search.brave.com and list which of the homepage, the pricing page and your three most important pages appear. For each one missing, give the date you submitted it for re-fetching.
1.3.3.6Serve a Markdown version of each content page to agents that ask for it
An agent that asks for Markdown gets the page without menus, scripts or popups, so it reads what you wrote rather than your layout. The same words as the HTML, never extra ones: a version only machines can see is cloaking.
Done when: Request three content pages with the header Accept: text/markdown. For each, give the Content-Type returned, the Vary header, and the canonical URL in its Link header, and say whether the headings match the HTML page. Anything other than text/markdown with Vary: Accept and a canonical pointing at the HTML page fails.
3.4.1Publish five pages answering real user questions
Pull the questions verbatim from your support inbox and Phase 1 interviews.
3.4.2Retrofit question-shaped headings on your top 10 pages
Headings that match how people ask are what extraction matches against.
Done when: Count the H2s on your highest-traffic pages and how many are shaped as a question a reader would actually type — ending in a question mark, or opening with what/why/how/when/which. Report both numbers. An extractor matches a question to a heading, so a page of noun phrases answers nothing it was not already ranking for.
3.4.3Publish a definition and stat block per core term
Self-contained, 40–60 words, one citable number. Written to be lifted whole.
Done when: Open a core term's page and read the block under the heading. It must answer 'what is X?' in 40-60 words as a passage that stands alone, and be followed immediately by one dated statistic with its source. Quote the block and count its words — a definition an engine cannot lift whole, or a number with no date and no source, does not count.
3.4.4Earn genuine mentions on Reddit and Hacker News
Both are indexed heavily by assistants. Genuine is doing the work; astroturf gets caught.
3.4.5Write the twenty prompts you will re-ask forever
Prompts are keywords. Six shapes carry the decision: category, use case, comparison, alternative, “is X actually any good”, and the one with the buyer’s constraints in it — must do this, must connect to that, under this price. The last is where a shortlist is actually cut.
Done when: List the prompts. Twenty or more, written as a buyer would type them, and recorded somewhere their wording will not drift — a prompt reworded between runs measures the rewording. The list must include queries you expect to LOSE, or the score can only ever go up.
1h
3.4.6Re-ask the whole portfolio monthly and score the answers
Three clean sessions per prompt, with the model version logged: one answer a month is variance, not movement. Then score position, recommendation, accuracy and the sources behind it.
Done when: Report how many of the twenty prompts named the product and how many CITED it with a link, across the engines, for this month's run. Named and cited are different numbers and both are needed — one is training data bleeding through, the other is the retrieval layer finding a page.
2h
3.4.7Read the prompts you lose as a to-do list
Whoever gets recommended instead has something you do not — a category page, a review count, a comparison, a thread. Name that asset, then go and build it. On a prompt with constraints in it, check the cheaper failure first: often the product qualifies and the fact that proves it is written down nowhere a model can reach.
45m
3.4.8Mine Bing’s grounding queries for the prompts you are already cited on
The AI Performance report counts your citations in Copilot and Bing’s AI summaries, page by page, and lists the grounding queries — what the model actually searched when it used you. It is the only first-party record of an AI engine’s own queries; fold them into the portfolio.
Done when: List the grounding queries the AI Performance report in Bing Webmaster Tools shows for the last three months, each beside the page of yours it cited.
3.4.9Read which pages AI Overviews and AI Mode pull in, in Search Console
The generative AI report counts impressions in AI Overviews and AI Mode by page, country and device — no clicks and no queries, so read it for which pages Google’s answers draw on, not for traffic. Low everywhere can mean the site is opted out: check that setting before blaming the content.
Done when: List the five pages with the most impressions in the Search Console generative AI performance report over the last three months, each with its impression count.
3.4.10Cite a source beside every claim, and quote a named person, on the pages you want quoted
In the Princeton GEO study, citing sources and adding quotations raised a page’s visibility in generative answers as much as adding statistics did. A link at the end of the page does not count: the source goes in the sentence that makes the claim.
Done when: Open your five most important pages. For each, count the factual claims, how many link their source in the same sentence or paragraph, and the quotes attributed to a named person. Report the three numbers per page.
3.10.1List every third-party page that keeps appearing
Run your ten highest-intent searches and ten prompts and write down what comes back: threads, listicles, review categories, videos. The repeats are the map.
2h
3.10.2Rank each source by how often it appears and how much it sways
A thread cited across nine of your queries beats a DR 80 mention nobody buying reads. Work only the handful scoring high on both.
45m
3.10.3Earn a place on the top five, one at a time
Find out who owns each one and how it earns before you spend the afternoon: many ranking listicles are affiliate properties, where the price is a commission and no pitch will move them.
8h
3.10.4Make every source describe you the same way
One category sentence, repeated by your site, your listings, your reviewers and your press. Consensus is what turns a claim into something a model will repeat.
Where do the links come from?
Original data, other people’s lists, directories, and the links you already earned but never collected.
3.8.3Publish one original-data study
“We analysed 10,000 X and found Y” becomes the citation everyone else has to use.
16hhow to do it
3.8.12Put the study’s raw data up as a dated file, with the headline numbers ready to quote
Writers link to a source they can check, and quote the line already written for them. On one audited site all five editorial links pointed at a single market report that offered both; nothing else on the site earned one.
Done when: Open the study page. Does it link a downloadable data file (CSV or similar) that states the date it was collected, AND show at least three headline figures written as sentences someone could quote?
1h
3.8.10Publish the “[niche] statistics” page writers have to cite
Fifty-plus figures, each traced to the body that produced it, dated and stated once, with ten quotable lines on top. Writers Google it the night before publishing; the link is the citation of the figure itself.
3.8.9Pitch the study while the number is still new
A statistic earns you referring domains, not positions — and only the ones you go and collect. Send the finding, not the product, to the people who already write about your category.
3.2.7.1Find 20 ranking listicles for your category
Search “best [category] tools 2026” and “best [competitor] alternative”. Note the author.
1h
3.2.7.2Pitch all 20 for inclusion
Short, specific, offer a free account for their readers, and make it trivial to say no.
2h
3.2.7.3Follow up once after five days
Once. A second follow-up costs you the relationship for a link you were not getting.
30m
3.2.8.1Find unlinked brand mentions
Search your brand name in quotes, minus your own domain. Log every mention without a link.
45m
3.2.8.2Email each one asking for the link
Thank them first, then ask. Conversion here is unusually high because they already like you.
1h
3.2.8.3Fix broken links pointing at you
Redirect each dead URL to the page that answers what the reader came for, or restore it. Sending them all to the homepage instead reads as a soft 404 and drops out of the index, which loses the equity you were rescuing.
45m
3.2.8.4Set a standing alert for new mentions
A Google Alert on the brand and your own name. The backlog is finite; what the plan earns later is not.
10m
3.2.12.1Pull the sites linking to two or more competitors and not to you
Link Intersect in Ahrefs, Backlink Gap in Semrush. Two competitors is the threshold that separates a habit from a coincidence.
45mcosts money
3.2.12.2Open each page and write down why the link is there
A tool, a study, a template, a quote — or an acquisition, a sponsorship or a friendship. Mark those last three as not winnable rather than pitching them; most of a competitor’s profile is not an opportunity.
2h
3.2.12.3Find the dead and outdated links on those same pages
A link that 404s, or still loads while describing a product that shut down. Name the exact line, offer the page that finishes the reader’s task, and do not call a resource obsolete just because its owner changed.
3.2.1.1Write a tagline under 10 words
Outcome first. Every form has this field and most have a hard character limit.
20m
3.2.1.2Write the 50-word short description
What it does and for whom, in the words a directory’s own search results show. Forms with a 60-character limit take the tagline instead.
20m
3.2.1.3Write the 250-word description, and vary its opening per form
One text, three ways in: lead with the outcome on one directory, with the alternative it replaces on the next, with the mechanism on the third. Never paste the same first sentence everywhere — duplicates get down-weighted.
1h
3.2.1.5Write the founder story in three sentences
Curated directories reject listings without one. Reused verbatim on launch day.
25m
3.2.1.6Set up the submission tracker
Directory, date, listing URL, status (live, submitted, skipped and why), dofollow verified. A marketinque project has it built in: the directory sprint on the plan page, checked nightly. You will be back here quarterly.
20m
3.2.2.1Submit to AlternativeTo
Nofollow, but it ranks for “[competitor] alternatives” and feeds AI answers heavily.
25mAlternativeToDR 79 · nofollowhow to do it
3.2.2.2Submit to SaaSHub
Dofollow, and ranks well for alternatives queries with genuinely high buying intent.
20mSaaSHubDR 77how to do it
3.2.2.3Create the G2 listingB2B
Listing only for now — it stays invisible until the review drive in 2.6.
40mG2DR 92how to do it
3.2.2.4Create the Capterra listingB2B
Bought by G2 from Gartner in early 2026 (agreement 29 Jan, close reported 5 Feb). Same review mechanics, separate listing, separate review count.
3.2.2.5Claim the GetApp listingB2B
Often auto-populates from Capterra. Claim it rather than creating a duplicate.
20mGetAppDR 78how to do it
3.2.2.6Claim Software AdviceB2B
Gartner property; syncs with Capterra in several categories.
20mSoftware AdviceDR 88how to do it
3.2.2.7Create the TrustRadius listingB2B
Lower volume than G2, higher review quality, and reviewers write real detail.
30mTrustRadiusDR 72how to do it
3.2.2.8Submit to SourceForge
Legacy, unglamorous, DR 92, and takes twenty minutes.
25mSourceForgeDR 92how to do it
3.2.2.9Submit the company profile to Slashdot
Same operator as SourceForge — do them back to back with the same copy.
20mSlashdotDR 88how to do it
3.2.2.10Submit to TheSaaSDirectory
DR 88, clean categorisation, fast approval.
15mTheSaaSDirectoryDR 88how to do it
3.2.2.11Submit to Tech.co
Directory plus media surface — occasionally turns into editorial coverage.
15mTech.coDR 80how to do it
3.2.2.12Add yourself to relevant Slant questions
A “what is the best…” platform. Answer the questions your buyers actually ask.
20mSlantDR 75how to do it
3.2.2.13Create the Indie Hackers product page
Doubles as the home for your build-in-public updates from Phase 3 onward.
25mIndie HackersDR 76how to do it
3.2.2.14Create the F6S profile
Used by accelerators; the profile outranks a lot of things for your brand name.
20mF6SDR 65how to do it
3.2.2.16Work the long tail: 20 more DR 40+ directories
Startup Stash, Crozdesk, SaaSWorthy, Startup Fame, Taalk, StartupBuffer, 10words, Launched.io and friends. One sitting.
3h
3.2.3.1Create the Crunchbase profile
Heavily weighted in AI training corpora and in “who is this company” lookups.
40mCrunchbaseDR 91 · nofollow
3.2.3.2Create the LinkedIn company page
Required before you can run any LinkedIn distribution in Phase 3.
40mLinkedIn PagesDR 99
3.2.3.3Create the Wikidata item
The structured spine most knowledge panels and assistants resolve entities against.
1hWikidataDR 92
3.2.4.1Set up the GitHub org profile and README
DR 98. Every software product should have this whether or not it is open source.
30mGitHubDR 98
3.2.4.2Publish one technical post on Dev.todev tool
Dofollow, and cross-post with a canonical URL back to your own blog.
2hDev.toDR 90how to do it
3.2.4.3Make your domain your Bluesky handle
One _atproto TXT record and the account becomes @yourdomain.com. Free verification, unlike every neighbour.
20mBluesky
3.2.4.4Claim the Substack publication
DR 93, and it becomes the newsletter home in Phase 3 if you want it to.
30mSubstackDR 93
3.2.4.6Claim the quick-win profiles in one sitting
Gravatar, About.me, Hashnode, Disqus, Scoop.it, Diigo, Pearltrees, Dun & Bradstreet.
2h
3.2.5.1Submit to There’s An AI For That
The largest AI directory, task-based search. Its own pricing page lists $437 with a newsletter mention and $49 as the base listing (the layout is ambiguous about what the $49 covers); a third-party review reports a monthly X thread that picks one tool as the only free route.
40mfree routeTAAFTDR 76how to do itevidence: weak
3.2.5.2Submit to Futurepedia
Its listing and pricing pages returned 404 on 2026-11-10, so no price is confirmed; a third-party review puts a verified listing at $497. Check the price before you budget it.
30mcosts moneyFuturepediaDR 70how to do itevidence: weak
3.2.5.3Submit to Toolify.ai
$99 and listed inside 48 hours according to third-party reviews; its own submit page shows no price but advertises a dofollow backlink, and reviews say at least six links. Public traffic trends make it competitive intel as well as a listing.
30mcosts moneyToolifyDR 71how to do itevidence: weak
3.2.5.4Submit to Future Tools
The one that is still free — a plain form, reviewed by hand. Do this before paying anyone. Small, but disproportionately influential in the AI-tool press cycle.
3.2.5.5Work the AI directory long tail
Acid Tools (free, dofollow), aitools.inc, AIStage, TopAI.tools, Supertools, AI Tools Guide, NewTools.site and the rest.
3h
3.2.6.1List on Glama and optimise for the security grade
Grades are public and comparative. An A is a differentiator; an F is worse than absent.
3.2.6.2Submit to the Linux Foundation MCP Registry
The official MCP registry, still in preview; the protocol moved to the Linux Foundation’s Agentic AI Foundation in Dec 2025. Not a PR: a server.json posted with mcp-publisher after proving the namespace (GitHub for io.github.*, a DNS or /.well-known key for your domain). Description ≤ 100 chars; each version publishes once.
3.2.6.3Publish on Smithery
Sign in, paste the server URL, skip session config unless the server needs a key; its inspector reads your tools and scores the listing. Description, homepage and repository lift the score. “Verified” wants a TXT record on the homepage host and a paid plan — optional.
3.2.6.4Submit to the agent directories
AI Agents List, APITracker, AI Agent Store, AgentHunter, AI Agents Directory.
2h
Yours to do, in your own accounts:
- AgentHunter — Sign in with Google (no password to create); the submit form follows. Fill it from the submission kit.
- AI Agent Store — Register or sign in with Google; the new-agent form opens after. Fill it from the submission kit.
- AI Agents Directory — Sign in to continue; the agent form follows. Pick the Marketing category.
- AI Agents List — Paid: a free eligibility check on your URL, then $29 or $49 once, refunded if not approved. Decide before starting.
- APITracker — No submission form: its team adds MCP servers. Use the Contact us link on this page to ask for yours, with the server URL.
3.2.6.5Submit your server to the Claude Connectors Directory and the ChatGPT plugin directory
Both list only remote servers over HTTPS, and both are reviewed. Claude’s needs OAuth 2.0 for any tool that acts on a user’s account, a title and a readOnlyHint or destructiveHint on every tool, a privacy policy and a test account for its reviewers. Fix those before applying, or the review is the first place you hear about them.
Done when: Give the public listing URL for the server in each directory. A submission still in review counts as not listed: say which one is pending and since when.
2hClaude Connectors Directory: developer portalhow to do it
Yours to do, in your own accounts:
- Claude Connectors Directory — Sign in on a paid Claude plan, choose Submit new → MCP connector, and give the server URL, docs and privacy policy URLs, an icon and a populated reviewer test account.
- ChatGPT plugin directory — Verify your OpenAI organisation as an individual or business, then on the Plugins page upload the plugin ZIP, connect the MCP server, pass the domain check and submit for review.
3.2.9.1Submit to NoCodeFinder
Open submissions, quick approval, and the audience is browsing to buy.
20mNoCodeFinderDR 45how to do it
3.2.9.2Submit to the No Code MBA tools directory
Categorised by project type, so pick the category a buyer would search.
20mNo Code MBADR 55how to do it
3.2.9.3Submit to NoCode.Tech
Long-running curated list; editorial standards are real, so write the entry properly.
20mNoCode.TechDR 35how to do it
3.2.9.4Work the no-code long tail
NoCodeList, NoCodeDevs, We Are No Code, NoCodeFounders, MakerPad. One sitting.
1h
3.2.10.1Submit your blog to Feedspot
Blog discovery plus RSS aggregation — a real referral source, not just a link.
20mFeedspotDR 80how to do it
3.2.10.2Submit your blog to Alltop
Long-established aggregator; one submission, permanent listing.
20mAlltopDR 73how to do it
3.2.10.3Distribute one free press release on a real milestone
PRLog, OpenPR, PR.com. Only for something genuinely newsworthy — launch, funding, a big number. Nobody reads these; the links are the point.
1h
3.2.10.4Stop at three release sites
Past a handful, low-quality distribution starts to look like a link scheme rather than a launch. Real PR is 4.4, pitched to humans.
5m
3.2.11.1Verify every live listing is dofollow
Inspect the link; absence of rel="nofollow" or rel="ugc" means it passes equity.
1h
3.2.11.2Log DR and referring domains weekly
A single number per week, and count recovered links apart from newly earned ones — repairing a redirect moves the same number as winning a link and means something completely different. If it stalls for a month, something flipped to nofollow.
10m
3.2.11.3Re-verify the whole tracker quarterly
Do not edit listings weekly. Do check them four times a year.
45m
3.2.13Answer the journalist requests in your field
The one route where the writer is already looking for you. Answer the question in the first sentence, from experience you can substantiate — no original data required, which is what makes it work at DR 20.
How do you scale it once it works?
More pages only after the first ones are indexed and earning, at a pace you can hold.
5.1.1Build the cluster map for 50 target keywords
Pillar plus spokes, mapped to funnel stage. This is the brief for everything you commission.
Done when: List the 50 target keywords, each with the pillar it belongs to. Fifty distinct keywords, every one assigned to a pillar or a spoke.
4hguide
5.1.2Count how much of your own market you appear for
The share of your mapped queries you appear on at all — fixed once and weighted by intent, because a coverage number moves as easily by redrawing the denominator as by doing the work. A direction of travel, never a score.
2h
5.1.3Audit the library by funnel stage before commissioning more
One audited library ran 2,200 awareness posts to 21 near the purchase. Publish nearest revenue first: comparisons, alternatives, categories, use cases.
1h
5.1.4Set and defend a publishing cadence
Two a week forever beats ten in January and nothing until June.
Done when: Quote the cadence as written down: posts per week and the weekday each one ships. A number and a day.
30m
5.1.5Contract a writer who knows the domain
Cheaper than your time and better than your seventh draft. Domain knowledge is non-negotiable.
10hcosts money
5.1.6Write the editorial brief template
Keyword, angle, outline, sources required, internal links, and what a rejection looks like.
2h
5.1.7Refresh the ten decaying posts each quarter
Updating a ranked page beats publishing a new one, almost every time. Save the live copy and its 90-day numbers, then read what outranks it and fix the gap — decay is usually the intent moving, not the writing ageing. Without the baseline a refresh can be neither reverted nor proved.
8h
5.1.8.1Pick ten high-intent queries off the cluster map
The same six shapes that work in text: best X, X for a use case, X vs Y, alternatives to X, how to reach an outcome, fix a problem. Broad category terms convert like entertainment.
1h
5.1.8.2Search each query on YouTube and read who already ranks
The live results tell you the intent and the real competition. Third-party difficulty scores for video are guesses.
5.1.8.3Write the title and mock the thumbnail BEFORE you record
Packaging decides whether the video is seen at all, so it is the thing to judge while it is still free to change. If the mock does not look clickable next to the current results, the video is not worth making.
45m
5.1.8.4Make each video answer a sub-problem your product solves
Teach the thing completely, product visible throughout, one CTA at the end. A great tutorial with a bolted-on plug converts nothing, and neither does a gated one.
30mhow to do it
5.1.8.5Judge each video per query, not on the channel average
Click-through and watch time are tracked per traffic source. A video can win on the query you targeted and lose on the channel dashboard.
30m
5.1.8.6Embed each video on its matching page, with the text under it
VideoObject markup on the page earns the video result; the transcript under the embed is what the page ranks for. Link the description back for the viewer, not for the link — YouTube’s are nofollow.
5.1.9Turn each guide into one visual card and post it to PinterestB2C
Pins are indexed and keep surfacing for months, so the asset compounds instead of scrolling away. Cheap to test: the guide already exists.
5.1.11Pull the keywords your top three competitors rank for and you do not
Keyword Gap in Semrush or Content Gap in Ahrefs, cut to where they sit in the top twenty and you are absent, difficulty under 40, a buying word in the query. Each survivor is a page to fix or a page to write — mark which, per row.
2hcosts money
5.1.12Re-check every competitor claim against their current pages, quarterly
Competitors ship; your comparison and alternative pages describe last year’s version. That is the page a buyer catches you on and an AI answer marks inaccurate. Their pricing and feature pages are the source — show the old line beside the new one.
2h
5.1.13Link each new post from the old ones, the week it ships
A page published today has nothing pointing at it. Search the back catalogue for the sentence that was already about this, and link from there — new pages cannot earn links from pages written before they existed unless someone goes back.
45m
5.1.14Name why a page lost clicks before you touch it
Compare the last three months with the three before, by page. Rule out your own recent edit first. Then: fewer impressions at the same position means the search shrank, so leave it; a lower position means someone passed you, so read page one; the same position and impressions with fewer clicks means the results page changed around you, such as an answer box above the links. Only the second cause is a reason to rewrite.
Done when: For the three pages that lost the most clicks, give impressions, position and click-through rate before and after, whether you edited the page in that window, and which of the four causes fits: the search shrank, someone passed you, the results page changed, or your own edit.
1h
5.2.1Pick the template with real search demand
Verify the pattern has volume before you build a generator for it. Most do not.
2h
5.2.2Generate /integrations/[partner] pageshas integrations
The Zapier playbook: Ahrefs estimated in 2022 that programmatic integration pages drove about 16% of their organic traffic.
5.2.3Extend alternative pages to every competitor
You wrote four by hand in Phase 2. Now cover the long tail of the category.
20h
5.2.4Build the template or use-case gallery to 100+ entries
One indexable page each, with related entries at the bottom. That internal mesh is the engine.
40h
5.2.5Interlink the generated pages
Unlinked generated pages are orphans and never get crawled at this volume.
8h
5.2.6Add a thin-content quality gate
Noindex anything below the bar. Mass-publishing thin pages is how sites get demoted wholesale.
Done when: Quote the gate's rule: the minimum word count or unique-data field, and what happens to a page that fails. A number and an action.
8h
5.2.7Hold the next batch until most of the last one is indexed
Google rejects scaled pages quietly. One audited directory reached 90,639 pages crawled but not indexed while its impressions fell from 4,107 a day to 48. Publish more only once about 85% of the sitemap is in the index.
Done when: From Search Console’s page indexing report and the sitemap, give indexed pages ÷ sitemap URLs. Does it reach 85% before the next generated batch ships?
30m
5.2.8Exempt every page that already earns clicks from the quality gate
A gate tuned on word counts noindexes winners too: one site’s two gates removed 258 pages that had been earning 28% of its clicks. Export the pages with clicks from Search Console first, and keep them indexed whatever the gate says.
30m
6.4.5.4Set hreflang and measure each locale on its own
Without it the versions compete with each other for one slot. An average across locales hides both the win and the failure.
2h
How do you own the category term?
One term buyers already search, the definitive page on it, and the original data others have to cite.
6.6.1Pick the term you intend to own
One term. Something buyers already search, not a phrase you invented in a workshop.
Done when: Quote the term chosen, and name three pages on your site that already use it in a heading. Three URLs.
2h
6.6.2.1Read the ten pages that rank for the term and list what none of them answers
The gaps are the guide. A page that repeats the ten above it adds nothing for a search engine or an assistant to cite.
Done when: List the ten ranking pages by URL, and under each the one question it leaves unanswered. Ten URLs, ten questions.
2h
6.6.2.2Outline the guide around those gaps, one question per heading
Headings that are questions get pulled out as passages. Put the shortest honest answer directly under each.
1h
6.6.2.3Add one original thing: your data, a tool or a worked example
This is what the other ten lack and what earns links. A guide with no original element is a longer version of theirs.
4h
6.6.2.4Write and publish it with a dated byline and cited sources
Name the author, date it, and link every number to its source. Undated, unsourced guides are the ones assistants skip.
Done when: Open the published guide. It shows an author, a date, and at least one cited source per statistic. The URL is the evidence.
6.6.2.5Link to it from every page that already mentions the term
A new page starts with no links pointing at it. Search your own site for the term and add the link where the sentence already exists.
45m
6.6.2.6Put a quarterly refresh date on the calendar
A definitive resource that is a year old is no longer definitive. Ten minutes now saves the decay later.
10m
6.6.3.1Pick the three questions your niche argues about and cannot answer with data
A report earns coverage by settling an argument. Choose questions where buyers already hold opposite beliefs.
Done when: Quote the three questions. Each is a question a buyer in your niche has asked in public, with a link to where.
1h
6.6.3.2Choose the data source: your own users, a survey of 200, or public records
Your own anonymised usage is free and nobody else has it. A survey needs a sample you can state: print the number of answers beside every percentage.
2h
6.6.3.3Collect the data and keep the raw file
Journalists and assistants ask to see the data. A report you cannot reproduce falls apart at the first follow-up.
15h
6.6.3.4Write the method box: sample, dates, what was excluded
Half a page. It is what separates a finding that gets cited from one that gets ignored.
Done when: Quote the method box. It states the sample size, the date range, and at least one thing that was excluded.
1h
6.6.3.5Write the report around the three findings that surprised you
If nothing surprised you, nothing will surprise a reader. Lead with the finding and the number, not the method.
12hhow to do it
6.6.3.6Cut each finding into a chart, a quote and a one-line claim
These are what get posted, embedded and quoted. Make them readable without the report around them.
3h
6.6.3.7Put next year’s edition on the calendar and keep the same questions
The same questions year on year turn one report into a trend line, which is the thing worth citing.
10m
6.6.4.1List twenty journalists who covered your niche in the last 90 days
Recent coverage is the filter. A writer who covered it last year may have moved beat.
Done when: List twenty journalists, each with the URL of a piece they wrote on your niche in the last 90 days. Twenty distinct names.
2h
6.6.4.2Write the pitch as the finding in one sentence, then the number
Journalists want the number, not the product. The product appears once, last.
1h
6.6.4.3Send five a day, each referring to something they wrote
Four days, twenty pitches, none sent as a blast. Mention a specific piece so it reads as written for them.
6.6.4.4Answer every reply within the hour and offer the raw data
A journalist on deadline moves on in an afternoon. Have the file, the method box and a quote ready before you send.
30m
6.6.4.5Log every pitch, reply and link earned
Without the log you cannot tell which angle worked. Three columns are enough: who, what you sent, what came back.
30m
6.6.5.1List forty podcasts your buyers listen to, with the host and booking contact
Ask customers what they listen to. List far more shows than the twelve episodes you want, because most will not reply.
Done when: List forty podcasts, each with the host’s name and a booking email or form URL. Forty distinct shows.
2h
6.6.5.2Prepare one story, one number and one offer to reuse on every show
Twelve interviews with twelve improvised answers teach nothing. A repeated story gets sharper each time.
1h
6.6.5.3Pitch each show with three episode topics, not one
Hosts book a topic, not a company. Three options lets them pick the one that fits their audience.
6.6.5.4Record the twelve episodes within the quarter
Batch them into two days a week for three weeks. One recorded hour keeps returning traffic for years.
12h
6.6.5.5Ask each host to link to a page built for that audience
A page named for the show converts better than the homepage and tells you which show sent the signup.
45m
6.6.6.1Find five events whose attendees are your buyers
Ask your best customers which event they would pay to attend. Ignore events full of other vendors.
1h
6.6.6.2Submit a workshop proposal to each before its call closes
Check each deadline, because calls often close well ahead. Propose a hands-on workshop, which gives the organiser something attendees can use.
6.6.6.3Build the session around one thing attendees can do before they leave
People remember what they did, not what they watched. One outcome, one exercise, one handout.
8h
6.6.6.4Record the delivery and keep the slides
Ask the organiser for the recording, and capture your own audio as a backup.
30m
6.6.6.5Cut the recording into a post, a short video and a lead magnet
One hour on stage becomes the next month of content. Link the handout to a signup page.
4h
More steps on this topic
Industrialise content
- Train a research feed until it shows you your category all day30m
Twenty minutes of deliberately watching and saving the content around your category buys you a permanent, self-updating swipe file.
- Collect a dozen formats, and keep only the ones you can remake in 15 minutes1h
The two-hour video that gets 300 views is what kills the habit. Cheap to remake is a selection criterion, not a compromise.
- Post three different formats a day, not three takes on one5h
Five days of one format tells you almost nothing. Different shapes on the same day is the only way the numbers mean anything.
- Read the failure, not just the view count45m
Walk one ladder: views, then engagement, then clicks, then signups, then people who stayed. The rung it stops on is the thing to fix, and a post that stops at views is not a win.
- Make twenty more of the one that worked before inventing another5h
You have seen every post you ever made; your audience has seen roughly none of them. Boredom is not a signal that a format is spent.
- Hand the winning format to someone else to produce2h
A proven format is a brief anyone can follow. This is the difference between a channel and a hobby you personally sustain.
- Put five small creators on the format you already proved8h
Hire for whether they can stop a scroll, not for follower count, and hand them your winners. A creator inventing your strategy is a creator running your first experiment again.
- Label the synthetic ones, and never let one give a testimonial15m
If the face, voice or scene is generated and meant to read as real, tick the platform’s own AI box — YouTube, TikTok and Meta all require it and detect it anyway. The harder line is separate: a person who does not exist cannot describe using your product. That is a fabricated testimonial, a label does not cure it, and the FTC rule names AI-generated ones explicitly. The same goes for a generated before-and-after: an AI transformation shows a result your product did not produce.
Plays
Reusable patterns that explain how: backed by read sources first, then practitioner advice, then our own ideas.
- List where the AI-curious search: AI directories and registriesBacked · 4 sources
List AI-adjacent products in the directories AI-curious buyers browse — There’s An AI For That and its peers — and in machine-readable registries like MCP server lists where AI agents themselves discover tools. These indexes rank for "[task] AI" searches and increasingly feed the recommendations assistants make.
- Build the page an answer engine can quoteBacked · 3 sources
Write the page so any passage still reads correctly once it is lifted out: a plain definition and one dated number at the top, comparisons as HTML text rather than images, a source linked beside each claim, quotes from named people, and a real author in the markup. Then make sure the search indexes behind the engines can fetch it.
- Give a query Google already ranks you for its own pageBacked · 2 sources
Search Console shows queries where Google gives you impressions but has no dedicated page. Each mismatch is a page you can rank faster than a cold keyword, because the domain-level association already exists. Publish one page per query, matching the exact intent, then re-check position and impressions before the next batch.
- Ship llms.txt: a crib sheet for AI crawlersBacked · 2 sources
Add an llms.txt file and a plain-language facts page stating what the product is, who it serves, pricing, and honest differences from alternatives, written in flat quotable markdown. AI crawlers assemble answers from whatever they can parse; giving them a canonical crib sheet makes the version they repeat the version you wrote.
- Hand agents the Markdown, not the layoutBacked · 2 sources
Serve a Markdown copy of each content page when a request asks for text/markdown, or at the same path plus .md. Keep the words identical to the HTML, send Vary: Accept so caches keep the two apart, and name the HTML page as canonical in the Link header. Agents get a clean read, and nothing competes with the original.
- Claim the review platforms and seed ten honest reviewsBacked · 2 sources
Claim and complete your profiles on the review platforms your buyers check — G2, Capterra, or the niche equivalent — then personally ask your ten happiest customers for honest reviews. Review counts unlock category visibility, the platforms rank for "[category] software" searches you can’t, and B2B buyers verify there before every demo.
- "[Category] alternatives" comparison articleBacked · 1 source
Target high-intent "[category] alternatives" and "best [category]" searches with a balanced comparison article. Use a real feature table, name honest trade-offs, and state who each option suits best — including where you are not the right pick. Comparison content is the single most-cited format in AI answers and converts bottom-funnel searchers.
- One page per "[your product] + [tool]" integrationBacked · 1 source
Create one focused page per integration — "[your product] + [tool]" — covering what the combination unlocks, a three-step setup, and one concrete workflow. Buyers search for the tools they already own, so each page intercepts "[tool] + [category]" queries with near-zero competition and arrives pre-qualified: they use the partner tool today.
- Work one keyword pattern, one row at a timeBacked · 1 source
Pick one keyword pattern, a query shape with one changing word and one intent, and publish a page per variant, two a week, on one template. Choose the pattern your site’s authority can win today, check each variant for a top-ten result missing the exact query in its title, and give every page something only that variant can say.
- Brief every page from the live results pageBacked · 1 source
Read the live top ten before you brief a page: which page type wins, which secondary queries the leader also ranks for, which questions Google suggests, and what every result misses. Check your own site for a page already covering it. Then brief the gap, keyword first in the title, with one proof only you have.
- Write the page for a new segment from the deals you already won in itBacked · 1 source
When a few won deals arrive from a segment you never aimed at, read why they bought and write that segment its own page in their words: the job, what they replaced, the questions they asked first. Send it to the next leads from that segment. You test the segment with a page before committing a price or a roadmap.
- Turn the "crawled, not indexed" list into a merge-or-improve sprintBacked · 1 source
Search Console lists pages Google crawled and chose not to index. Export that list and sort it: improve the pages that matter, merge the ones that repeat another page, and remove the rest. Google says the status is common and not always a problem, so judge by whether the page earns its place, then re-inspect the ones you fixed.
- Build template pages only from data your product already holdsBacked · 1 source
Template pages work when every page carries data only you have. Take an aggregate your product already produces, per integration, city or category, and publish a page for each row with enough data behind it, skipping the thin ones. Google’s spam policy targets pages made mainly to rank, so the data is what makes each page worth publishing.
- An FAQ page built from the questions buyers actually askBacked · 1 source
Collect the real questions from sales calls, support tickets, and community threads, then answer each on an FAQ page in self-contained 40-60 word blocks with FAQPage structured data. AI engines answer questions by extracting passages; pages organized as question-plus-liftable-answer are the easiest content for them to quote — with your name attached.
- Retrofit your best pages into question-and-answer shapeBacked · 1 source
Restructure existing high-traffic pages so each section heading is a real buyer question and the first 40-60 words beneath it answer that question completely. AI engines quote passages, not pages; converting prose into question-plus-standalone-answer units makes content you already own extractable without writing anything new.
- Film the answer to one query people already search on YouTubeBacked · 1 source
Pick a query your buyers already type into YouTube, then publish the video that answers it end to end. Write the title and mock the thumbnail before recording — packaging decides whether it is seen. Teach completely, keep the product visible throughout, and judge the video on that query rather than on channel averages.
- Publish the “[topic] statistics” page writers have to citePractitioner · 2 sources
A “[topic] statistics” page is the query a writer runs when a claim needs a number, so the page that answers it best is cited for years. Aggregate fifty-plus figures traced to primary sources — the filing, the study, the government table — each dated, linked and stated once, with ten quotable one-liners on top. Pick the sub-topic where page one is weak, and stamp the year.
- Answer journalist source requests, 20 minutes a weekPractitioner · 2 sources
Spend twenty minutes weekly answering journalist source requests on platforms like Qwoted and Featured: two quotable sentences, your credentials, sent fast. The reporter already needs an expert — you’re fulfilling demand, not creating it — and each placement returns a citation, a backlink, and a byline that compounds authority.
- Syndicate your best post with a canonical linkPractitioner · 1 source
Republish your strongest blog post on Medium (and similar networks) with a rel=canonical link back to the original. You borrow the platform’s built-in distribution and a fresh audience without splitting SEO equity or risking duplicate-content penalties — the canonical tells search engines your own site is the source of record.
- Brief an original data study from numbers only you havePractitioner · 1 source
Original statistics are the citation other writers need, and one solid study outranks years of opinion posts — but the numbers live in a database only the founder can query. Deliver the brief instead: the question, the cut of data that answers it, the methodology note, and the headline it would earn, every figure an explicit placeholder until the query runs.
- Serialize the build log on Dev.to and HashnodePractitioner · 1 source
Serialize your engineering decisions as a build log on Dev.to and Hashnode with canonical links to your own blog: the tradeoff you faced, the numbers, the code. Developer syndication networks give unknown writers real distribution fast, and a candid build log markets a dev tool better than any feature page.
- Retell your data study as a story for platform readersPractitioner · 1 source
Rewrite your original data study as a narrative for Medium’s reader culture — the question first, the hunt, the charts as the payoff — with a canonical link to the original. The same research earns a second audience on the platform’s distribution, while the canonical protects your site’s search equity.
- Author the honest "best [category] tools" list yourselfPractitioner · 1 source
Publish a genuinely fair "best [category] tools" ranking on your own domain, including real competitors, their strengths, and who each option suits, with yourself placed honestly. AI engines and searchers both favor ranked lists; authoring the definitive one means the comparison happens on your turf, framed by your criteria.
- Definition + statistic block engineered for AI citationPractitioner · 1 source
Lead a page with a 40-60 word definition that answers "what is [category]?" as a standalone passage, then attach one original, dated statistic with a citation. AI engines extract passages, not pages, and citing sources plus statistics are the two highest-leverage moves for getting quoted in AI answers.
- Pitch one original statistic to a journalistPractitioner · 1 source
Run a small original survey or mine your product data for one surprising, defensible statistic, then pitch that single finding to a journalist who covers your space. Reporters need data, not product announcements; an original number gives them a story and earns you an authoritative backlink plus a citation that AI engines reuse.
- Replace the dead links on pages that already rankPractitioner · 1 source
Scan ranking pages in your niche for dead outbound links, then email the owner: point out the broken link and offer your equivalent content as the fix. The opener is a favor, not an ask, which is why broken-link outreach converts when cold pitches don’t — and every placement inherits an aged page’s authority.
- Claim your spot where switchers shop: AlternativeToPractitioner · 1 source
Claim your AlternativeTo listing and position it against the incumbents users leave: accurate description, current screenshots, honest feature flags. The site ranks for "[incumbent] alternatives" queries — searches run exclusively by people ready to switch — and a well-groomed listing converts that borrowed intent into trials.
- Listing on the niche directories your buyers browsePractitioner · 1 source
Submit the product to the few niche directories and launch platforms your buyers actually browse, with a keyword-aware, benefit-led description and a real screenshot. Niche directories convert better than generic ones, often rank for "[category] tools" queries themselves, and supply durable backlinks that strengthen your domain.
- The "how much does it cost" page competitors won’t writeOur idea
Publish an honest "how much does [category] cost" page: real price ranges, the factors that move them, and when the budget option is enough. Buyers searching cost are close to a decision, most vendors dodge the question, and the page that answers it plainly wins the click and the trust.
- A free template page for the task buyers do weeklyOur idea
Build a page of free, genuinely finished templates — the spreadsheet, checklist, or brief your buyers recreate constantly — targeting "[task] template" searches. Template queries carry doing-intent rather than browsing-intent, so visitors arrive mid-task, use your resource, and meet the product exactly where it removes that same work permanently.
- Tear down a famous company’s playbook, step by stepOur idea
Reverse-engineer how a famous company does the thing you know best — its onboarding, pricing, packaging, or funnel — with screenshots and honest critique. The brand’s name draws search and social traffic your own couldn’t yet, while the quality of your analysis converts borrowed attention into readers who trust your expertise.
- One long-tail question, one focused postOur idea
Publish short, single-question posts on syndication platforms — each answering one narrow query your buyers actually search, too specific for your main blog. The platforms’ domain authority ranks for long-tail questions your young site can’t reach yet, and every post links back to your fuller guide.
- A public "your first week with [product]" guideOur idea
Publish a guide to the first week with your product: what to set up on day one, the first result to expect, and what to do on days three and five. New users get a path instead of a blank dashboard, prospects see what using it is like before signing up, and the page ranks for your product’s tutorial searches.
- The "what is [product]" page, answered in one sentence firstOur idea
Publish a page whose first sentence answers "what is [product]?": the category, who it is for, and the one thing it does differently, followed by what it is not and who should pick something else. Branded searches and AI answers pull from this page, and if you do not write it, a directory or a competitor writes it for you.
- Build the tiny tool other sites embedOur idea
Build a small, genuinely useful free tool — a calculator, grader, or generator your buyers reach for mid-task — and make it embeddable with attribution. Interactive utilities earn links passively for years because writers cite tools they use, and the tool doubles as a qualification magnet for the product behind it.
- Write the canonical answer to the sub’s weekly questionOur idea
Identify the question your target subreddit asks every week, then write the definitive answer as a structured text post — thorough, current, and free of links to you. Recurring questions guarantee recurring traffic; the canonical answer gets bookmarked, linked by regulars, and quietly makes its author the recognized expert.
- Say the same sentence everywhere a machine looksOur idea
Write one sentence naming your category and your buyer, then propagate it verbatim across your homepage, listings, review profiles and community answers. Retrieval systems assemble a picture of you from many sources at once; when they all agree, the claim reads as consensus, and when they contradict each other the model hedges or leaves you out.
- Own the annual "State of [your niche]" reportOur idea
Survey your niche annually and publish the results as "The State of [Niche]" — charts, methodology, quotable findings — offering each major outlet one exclusive angle. A recurring report compounds: year one earns coverage, year two earns comparisons against year one, and soon journalists cite you by default.
- Pitch ten niche podcasts one specific story eachOur idea
Pitch yourself to ten niche podcasts your buyers hear, each pitch carrying one specific story only you can tell, with the takeaways promised up front. Podcast audiences are small but parasocial — an hour in their ears builds more trust than a quarter of posting — and episodes keep converting for years.
- Teach a workshop where attendees leave with the thing builtOur idea
Host a recurring free workshop where attendees build something real in 45 minutes — leaving with a finished plan, page, or setup rather than notes. Teaching beats demoing: registrants hand you qualified pipeline, the hands-on format activates them inside the product naturally, and the recording becomes evergreen content.
Founder questions
How long does SEO take for a new SaaS?
Expect months, not weeks. A new domain usually ranks on page five or worse at first, and climbs as pages get indexed and links arrive. Judge a page on its exact query weeks after a change, not days.
What should a new SaaS site publish first?
Pages for buyers who are already choosing: “[competitor] alternative”, “[you] vs [competitor]” and use-case pages. They rank more cheaply than broad topics, and the people who find them are close to buying.
How do I know whether my SaaS SEO is working?
Read Search Console with your own name taken out. On a young site most clicks are people typing the product’s name, which the content did not earn. Then watch the non-brand clicks, and the exact position of each page you changed, weeks after the change.
Does a small site need a full technical SEO audit?
Rarely at the start. Most frameworks already handle the basics. What matters is whether Google can crawl, render and index the pages that make money. Check those, then spend the time on pages and links.
Asked in public, and answered from the thread:
Related topics
Go further
Where it sits in the plan
- 3.8Publish the founding content setin Authority baseline
- 5.1Industrialise contentin Compounding engine
- 6.6.2Publish the definitive resource on itin Scale to $1M ARR
Sources
- Google, Search Central: SEO documentation & quality guidelines
- Google, Spam policies for Google web search (link spam) (Google Search Central)
- Google, Optimizing for generative AI features on Google Search (Google Search Central)
Compiled by The Marketinque Team to our editorial standards.