SEO for early-stage SaaS: what to do before you have a domain worth anything

We turn down most seed-stage companies that ask us to work with them, which is an odd thing for an agency to admit on its own site. The reason is that an agency retainer is a bad use of $6,000 a month when you have eleven customers and no idea yet which of them represents the market.

That does not mean do nothing. It means do four specific things yourself, cheaply, and wait.

Why the standard advice does not apply

Almost all SEO writing assumes a domain with some authority, a team that can ship pages, and a category with established search demand. At seed stage you usually have none of those.

Your domain is two years old with fourteen referring domains. Competing for “[category] software” against companies with a decade of links is not a plan, it is a wish. And you may be creating the category, in which case nobody is searching for it yet — a genuinely difficult position that no amount of keyword research resolves.

The right frame at this stage is not “how do we rank” but “what are we building now that will be worth something in eighteen months.”

The four things

Get the technical basics right once, cheaply. Server-rendered or static marketing pages, one URL per piece of content, clean internal links, a sitemap, correct canonicals, and page titles that describe the page. This is a two-day job if you do it while the site is small and a two-month job at Series B. Do it now. Astro, Next with static output, or a well-configured Webflow will all get you there.

Write the pages nobody else will write about you. Your comparison pages against the two incumbents. Your pricing page with actual numbers. A migration guide from whatever your customers are leaving. Your security and compliance page. These rank fast because almost nothing competes with them, they convert, and your AEs will send them in emails regardless of whether Google ever notices.

Publish the things only you know. You are close to a specific problem and you have opinions about it that the incumbents are too large to state. That is the only content advantage you have at this stage, and it is a real one. A founder writing 800 honest words about why the standard approach in their category is wrong will outperform 4,000 words of researched explainer, both for links and for the readers who matter.

Start collecting the raw material. Record sales calls from day one. Tag closed-lost reasons properly. Keep a document of every question a prospect asks that surprises you. In eighteen months this is the highest-value asset in your keyword research, and it cannot be reconstructed retroactively.

The category creation problem

If you are genuinely creating a category, search demand for your category name is zero and will stay zero for a while. Chasing it is a waste.

What works instead is ranking for the problem rather than the solution. Before “marketing automation” was a search term, people searched for “how to send different emails to different customer segments.” Before “revenue intelligence,” people searched for “how to know which deals will close.”

Find the workaround your prospects use today and target that. “How to do X in a spreadsheet” is a query with real volume in almost every category, and the people searching it are exactly the people whose problem has outgrown the spreadsheet.

The pages will convert at a lower rate than true commercial pages, and they will be your only viable search entry point for a year or two.

What not to bother with

A publishing schedule. Publish when you have something. Four excellent posts in a year from a founder who knows the domain beats forty commissioned ones.

Link building campaigns. At this stage links come from doing things — launching, being on podcasts, publishing something opinionated, appearing in your investors’ portfolios. Paying for outreach before you have anything worth pitching is spending money to learn that you have nothing worth pitching.

Broad top-of-funnel content. You cannot win those SERPs yet and they would not convert if you did.

Rank tracking, mostly. You will not move for months and watching it daily is corrosive. Check Search Console impressions monthly and get on with building.

When to actually hire someone

The signals we look for: at least fifty customers so there is a pattern to learn from, an ACV that makes the arithmetic work, someone in the company who can approve and ship a page within a week, and enough runway that a nine-month payback period is not existential.

Before that, the highest-return marketing hire is usually not an SEO at all. It is a person who can do product marketing and write, and who will produce the raw material that makes an SEO program possible later.

We have told about a dozen companies this in the last two years. Four came back eighteen months later with a much better business to work with.