"Seo for startups" gets 210 searches a month in the UK. "Saas seo" gets the same, but with a cost per click of £25.46, and "saas seo agency" reaches £37.69: the highest CPCs in the set of SEO terms we measured (DataForSEO, Google Ads UK, September 2026; CPC converted from USD at £1 = $1.3242). On the first page of Google UK for "seo for startups", below the AI Overview, the first organic result is a downloadable PDF guide, and two of the results further down were published in 2021.
SEO for a startup or SaaS is the work of ranking for the problems the product solves, not for the name of the category the product sits in. Someone searching "accounting software" is already at the bottom of the funnel and finds ten competitors with ten years of domain history behind them. Someone searching "government gateway" has a problem, not a budget: and that is where the UK SaaS companies that win in organic search have built their traffic.
Below are the numbers for three UK SaaS and fintech companies (FreeAgent, Crunch, Tide), what separates a site with 7,940 ranking keywords from one with 13,604 that gets about a third of the traffic, and the method we apply when the domain is new.
The sector's keywords have tiny volumes and huge CPCs
The vocabulary a startup uses to look for help is almost empty: "seo for startups" 210 searches a month, "startup marketing" 110, "content marketing saas" 50, "saas seo" 210. The only term with real volume is "growth hacking", 390 searches a month with a peak of 720 in September 2025.
The figure that matters is not the volume, it is the price. A click on "saas seo" costs £25.46 against £5.13 for "seo course". When Google values a click at twenty-five pounds, it means someone on the other side is closing annual contracts. For a SaaS this flips the selection criterion: a keyword with 200 searches a month at that CPC is worth more than one with 1,000 that brings in students.
The practical consequence is that a SaaS's traffic cannot come from the keywords that describe the SaaS. It has to come from somewhere else.
What FreeAgent does to earn 224,000 organic visits a month
FreeAgent has 7,940 keywords in the top 100 of Google UK, 125 of them in first position and 2,904 in the top ten. The value of its organic traffic, calculated as the equivalent cost in Google Ads, is about £1.04 million a month (DataForSEO, September 2026).
The surprise is where that traffic comes from. The page that works hardest is /glossary/government-gateway/, fourth on "government gateway" (550,000 searches a month) and third on the variant "gateway government". A glossary entry. It explains a government login service, not the software.
The pattern repeats across the whole site:
- /glossary/self-assessment/ is third on "self assessment", 90,500 searches a month, and fifth on "hmrc self assessment", 110,000
- /glossary/tax-return/ is seventh on "tax return", 60,500 searches
- /glossary/personal-allowance/ is fifth on "personal allowance", 60,500 searches
- /glossary/utr-number/ is sixth on "utr number", 27,100 searches
- /guides/vat/registering/ is second on "registering vat", 18,100 searches
- /guides/self-assessment/how-to-register-for-self-assessment/ is third on "register for self assessment", 18,100 searches
None of these pages is about the product. All of them are about a problem faced by the people who later buy it: a sole trader or a small company director who needs to understand how an obligation works. The software comes in afterwards, when the reader has already had the answer for free.
The comparison page with the competitor
There is a second, more aggressive pattern. FreeAgent has a page, /features/competitors/, that ranks first on "freeagent vs xero" (320 searches a month, CPC £19.00) and on "quickbooks vs freeagent" (140). It catches people who are weighing up a competitor's product and shows them a comparison.
It is the tactic a small startup can most easily copy, because the volume depends on the competitor's brand as much as on yours: if your competitor is well known, the demand already exists. It works on two conditions. The comparison must be honest and verifiable, with prices and features dated, otherwise it is the page the competitor will use against you. And it must say who wrote it: nobody cites an anonymous comparison table.
Support content works too: support.freeagent.com ranks fourth on "file self assessment" (1,900 searches a month) with a help article on filing the return, and fourth on "p&l accounts" (1,600) with the one on the Profit and Loss report. Help pages are the organic asset that almost every SaaS keeps on noindex out of habit.
Why Crunch has more keywords and a third of the traffic
Comparing three UK companies in the same market explains where the difference lies better than any theory.
| Domain | Keywords in top 100 | In position 1 | In positions 11-20 | Estimated organic traffic |
|---|---|---|---|---|
| freeagent.com | 7,940 | 125 | 1,883 | 223,868 |
| crunch.co.uk | 13,604 | 100 | 2,657 | 81,025 |
| tide.co | 5,224 | 211 | 935 | 54,835 |
Crunch has 71% more keywords than FreeAgent and brings in just over a third of the traffic. The difference is not how many keywords you have, it is where they sit: Crunch has 2,657 keywords on page two against FreeAgent's 1,883, and 11,285 of its 13,604 keywords (83%) are outside the top ten, against 63% for FreeAgent. Most of its work is parked where nobody clicks.
For a startup this is the most useful figure on the page. Publishing more moves the keyword count; it does not move traffic. What moves traffic is getting the twenty keywords that already sit between positions eleven and twenty onto page one.
Where a startup with no domain authority starts
FreeAgent's pages ranking for "government gateway" or "self assessment" are there partly because the domain has years of links behind it. A domain registered six months ago will not get in on those keywords, however well the article is written.
The realistic starting point is the long tail of the problem, not its head. Instead of "self assessment", "self assessment for a sole trader with income from abroad". Instead of "registering vat", "vat registration for a UK company selling to EU customers". These are searches with a few dozen a month, with SERPs filled by forums and partial answers, and they are the only ones a new domain can get into within a few months.
The second place to look is what the product already knows. A SaaS has data nobody else has: how many of its users do a certain thing, how long it takes them, how often they get it wrong. A study built on those numbers, published with its methodology, is the one page competitors cannot copy and other sites cite. It is worth more than ten guides rewritten from the ones already on page one.
When SEO is not the right channel for a startup
There are three cases where it is worth saying so before signing.
The first is a category that does not exist yet. If the product solves a problem nobody knows they have, there are no searches to capture: demand has to be created elsewhere and SEO comes later, when people start searching for the name of the category.
The second is a short runway. On the projects we manage, stable top positions on commercial keywords arrive between the sixth and twelfth month. A startup that has to show traction at its next funding round in four months is buying the wrong channel: Google Ads gives an answer in two weeks, even if you pay per click.
The third is a product that changes every month. If the commercial positioning is not stable, the pages have to be rewritten constantly and the work done is lost. Better to wait for the message to settle.
Recommended reading: how long it takes to see SEO results.
What to measure in the first twelve months
Total organic sessions are the worst metric for a startup, because they also rise when branded traffic rises, that is, when another channel is working.
These are the three we watch on new projects. The number of non-branded keywords on page one, month on month: it is the only one that tells you whether SEO is working. The number of pages that received at least one click in Search Console in the last 28 days, which measures how much of the site is alive rather than just published. And demo or trial requests attributed to non-branded organic, which is the point where the channel stops being a chart and becomes a line in the CRM.
Recommended reading: the SEO KPIs to monitor and how to write SEO goals that hold up.
Frequently asked questions
How much does SEO cost for a startup?
It depends on how much content has to be produced, and when the site is new the content work weighs more than the technical side. A very small monthly budget buys a few articles, not a project: we have set out the figures in how much SEO costs and in how much work a low monthly budget actually buys.
Do I need a blog, or are product pages enough?
Product pages rank for category keywords, which have low volumes and competitors with ten years of domain history. FreeAgent's traffic comes from its glossary and guides, not from its feature pages. With a new domain the blog is not an extra: it is the part that brings in the visits.
Is SEO for a B2B SaaS different from SEO for a B2C one?
The ratio between volume and value changes. In B2B the keywords that matter have volumes in the tens or low hundreds and double-digit CPCs, and the buying cycle goes through several people. We have described the method in B2B SEO.
Should I add other languages if the product is international?
Only if you have the structure to run more than one market. A second language version without correct hreflang and without content written for that market takes resources away from the market you actually sell in. The rules are in international SEO.
Do AI tools change anything for a SaaS?
They change who shows the answer. On the UK SERP for "seo for startups" the AI Overview sits above all the organic results, so part of the informational searches no longer produce a click. For a SaaS the countermeasure is having pages assistants can cite as a source: your own data, clear definitions, dated numbers. We cover it in answer engine optimisation.
Recommended reading: the comparison across sectors.
Where to start
Open Search Console and filter non-branded queries in positions 11 to 20. Those are the pages that are nearly there, and over the next three months they are worth more than any new article. If the list is empty because the site has only just launched, then the first piece to write is the exact question your users ask support most often: it is a search someone else is running on Google right now, and nobody has answered it well yet. If you want an opinion on which keywords make sense to target with your domain, get in touch.
