Your SaaS Has 1 Click From Google. Here Is the 90-Day Plan I Would Actually Run.
One click from Google feels insulting when you have already done the work.
You bought the domain. You built the thing. You set up Search Console, installed analytics, wrote a few articles, added a sitemap, and waited. Three weeks later, the dashboard says 614 impressions, one click, and an average position somewhere in the forties. You start wondering whether you chose the wrong product or whether SEO is just a polite name for yelling into a forest.
Neither conclusion helps.
For a new SaaS, low traffic is normal. More importantly, it is not one problem. "We have no SEO traffic" can mean Google has not indexed the important pages, that you are ranking for words nobody uses to buy, that you are competing with companies a thousand times your size, or that visitors arrive and do not understand what you sell.
The fix depends on which one is true. Writing 30 articles before you know is how founders turn a solvable distribution problem into a content graveyard.
Here is the 90-day plan I would use if I were starting from almost zero today.
First, Read the One Click Correctly
An impression is Google saying, "I know this page exists and I tried it in front of someone." It is not Google saying the page is good. A few hundred impressions with an average position of 45 means it has probably started testing your pages for long-tail queries. That is better than zero. It is too early to judge the result.
The mistake is treating every weak number as a content problem. At this stage you need a diagnosis, not output.
Open Search Console and make four small lists:
- Pages that are indexed and getting impressions.
- Important pages that are excluded, crawled but not indexed, or not found.
- Queries with impressions where your average position is 11 to 30.
- Queries with impressions that clearly do not match what you sell.
The third group is your earliest opportunity. The fourth tells you whether your wording is attracting the wrong audience. If your AI invoice tool gets impressions for "free invoice template," do not congratulate yourself. You may be getting very good at drawing people who do not want your product.
Days 1 to 14: Fix the Boring Baseline
Before keyword research, make sure Google can visit, understand, and index the pages that deserve to rank.
Check the pages, not the score
Do a site:yourdomain.com search. Compare the rough count with your sitemap. In Search Console, inspect the homepage, pricing page, primary product page, and your three best articles. Check the last crawl, canonical URL, index status, and whether Google selected a different canonical.
Then load those pages in an incognito window with JavaScript disabled if your framework makes that useful. Can you still see an actual H1, a concrete description of the product, and internal links to the next relevant page? A beautiful empty shell is not a marketing page.
Also make sure you have one clean answer on each commercial page:
- who it is for
- what job it does
- what it replaces or improves
- how someone tries it or buys it
That does not mean every page needs to look like a 2014 landing page. It means the useful sentence should not be trapped under an animation and three metaphors.
Fix conversion leaks before buying traffic
If you are getting fewer than ten visitors a day, a conversion-rate experiment is mostly theatre. But obvious friction still matters. Ask a person who does not know the product to spend sixty seconds on the homepage. Then ask what the product does and what they would do next.
If their answer is fuzzy, rewrite the page. Do not add more features. Do not buy another analytics subscription. Fix the sentence.
Days 15 to 30: Find a Narrow, Buyable Problem
Most early SaaS content goes after broad terms because broad terms feel like a big market. "AI SEO tools," "project management software," and "best CRM" have demand. They also have entrenched competitors, comparison sites with decades of backlinks, and search intent that may be impossible for a small product to satisfy.
Start from the task your product handles at its most specific.
If you make a tool that tracks AI citations, "AI visibility platform" is a category term. "How do I know which ChatGPT prompts name my competitor" is a painful question asked by a buyer who is much closer to action. It will have lower volume. That is fine. One hundred people with the problem is more useful than 10,000 people browsing a category.
Build a small keyword sheet with five columns: the query, intent, the page that should answer it, the current top result, and the evidence you can add that the top result cannot.
That last column filters out most bad ideas. If the best page already gives a complete generic explanation and you only have a slightly different adjective, skip it. If you can include a calculator, a template, a real workflow, a comparison based on your product's actual capability, or data you collected, you may have something.
This is the distinction between useful programmatic SEO and the thin version that got crushed this year. Pages that survive at scale contain a data layer or a job the reader can actually do. They do not exist just because a keyword can fit in an H1.
Days 31 to 60: Build Three Pages That Can Win
Not 30. Three.
I would make one high-intent product page, one comparison or alternative page, and one useful free artifact.
The high-intent page
This should be for a clear use case, not your brand name. A page called "AI visibility tracking for B2B SaaS" is better than a fifth homepage variation if that is a specific problem your product solves.
Explain the starting situation, the method, the result, limits, price, and who should not buy it. The "who should not" part is underrated. It makes the page more credible and stops you from trying to sell everyone.
The comparison page
Comparison pages work because someone searching "X alternative" has already accepted that the category exists and is evaluating options. Do not make a fake comparison where every row conveniently favors you. Name the cases where the incumbent is better. A reader can smell a rigged table from the first screen.
The practical rule: only compare products you have used enough to describe accurately. If you cannot do that, write the problem page instead.
The free artifact
For an early product, a no-login checker, calculator, template, or short audit can outperform another article because someone can get value before trusting you. It also gives a person a reason to link or share something other than your sales page.
The best version is adjacent to the paid product, not a random free toy. A visibility tool can offer a short citation check. An invoicing product can offer a rate calculator. The point is to let people experience the problem becoming smaller.
Days 61 to 75: Make the Site Legible Beyond Google
Search does not begin and end with a blue link anymore, but this is where people make the opposite mistake and start chasing every AI-search checkbox.
Do the basic technical work. Use clean titles and descriptions. Publish useful structured data. Make pages fast enough and crawlable. Add llms.txt if you have deep documentation and can maintain it. But do not mistake it for the plan. It is documentation hygiene, not an AI citation lever.
The thing that makes an assistant more likely to name you is the same thing that gives Google confidence: a clear answer on your own site plus evidence that the product exists somewhere else.
Pick five places where the exact buyer spends time. Reply to real questions there without forcing a link. Submit to the relevant directories, not a bulk list of 400. Ask a friendly early customer for a specific review. Write a short case study only if you can name a real outcome.
It feels slower than a backlink blast because it is slower. It also leaves a footprint that a hundred directory submissions never will.
Days 76 to 90: Decide What the Evidence Says
At the end of 90 days, do not ask "did SEO work?" Ask narrower questions.
Are important pages indexed? If not, stay on technical diagnosis. No content cadence fixes an indexing problem.
Did impressions rise for the three pages you chose? If yes but positions remain low, improve the evidence and internal links before publishing more.
Are you getting the right queries? If not, change the page framing. You may have a positioning problem wearing an SEO costume.
Did any visitor take the next step? If traffic arrived and no one clicked a demo, tried the free artifact, or signed up, watch a few sessions and talk to a few users. The bottleneck moved from discovery to message or product.
What did people say in the places you showed up? Community replies are qualitative research. A person saying "I would use this if it did X" is more actionable than another position-tracking chart.
The output of 90 days should be a decision. Double down on a page cluster that has traction. Fix an invisible technical problem. Reframe the product. Or accept that search is a slow background channel and spend the next month on direct conversations instead.
The Part Founders Do Not Want To Hear
SEO is a compounding channel. It is not a launch plan for a SaaS that needs its first ten customers now.
Your first users will often come from the places where you can have a conversation: a community, a customer introduction, a small audience you have built, an integration partner, or a direct message that is specific enough to earn a reply. I sent 2,159 cold messages and got four replies, so I am not going to pretend those channels are pleasant either. They just give you feedback before Google does.
Run SEO in the background. Make a few pages that deserve to rank. Watch what happens. Then keep building distribution you can touch with your own hands while the search engine takes its time.
That combination is a lot less glamorous than publishing daily, and much more likely to get you past one click.
Frequently Asked
How long does SEO take for a new SaaS?
Expect the first useful signal to take months, not days. Early impressions can appear in weeks, but sustained rankings depend on the page, the competition, technical health, and evidence outside your own site.
Should I publish lots of AI-written articles to get traffic faster?
No. Publish a smaller number of pages that resolve a real decision, include original evidence or a useful tool, and fit your product. Thin volume is a cost, not an asset.
What should a SaaS founder do before SEO starts working?
Talk to likely users directly, participate in the communities where they ask questions, build one useful free artifact, and use those conversations to sharpen the product page. SEO should compound alongside those activities, not replace them.