Senior and technical
$75 to $120/ hour
Query strategy, technical architecture, structured data, internal link design, measurement into the CRM, and the subject matter writing on technical pages.
SaaS SEO
SaaS SEO agency, for the queries that still convert
AI summaries roughly halved clicks on the queries they appear on, and two independent datasets agree. This page publishes both with their methodology, names the query types that were unaffected, and sets out what a SaaS SEO programme should deliver now that fewer queries are worth chasing and each one is worth more.
What does a SaaS SEO agency do?
A SaaS SEO agency chooses which search queries a software company should own, builds the pages that deserve to rank for them, fixes the technical problems preventing indexing, and reports on the pipeline those pages produce. In 2026 the first of those four decisions matters far more than it used to.
It matters more because the queries no longer behave the same way. Informational queries, the ones the SaaS content industry spent a decade producing at volume, now frequently get answered on the results page without a click. Commercial queries largely do not. An agency that has not changed its query selection in response is selling a strategy whose evidence has expired.
What has not changed: software buyers still research in search, they still read long comparison and pricing pages before they talk to a vendor, and a page that genuinely answers a hard question still earns the visit. The channel is narrower and the surviving queries are worth more each.
The rest of this page is the evidence for that claim, the arithmetic on whether the channel still pays, and the specific query types we would and would not spend your money on.
Has AI search broken SaaS SEO?
AI summaries have roughly halved the click rate on the queries where they appear, and two independent datasets agree on the direction. Pew Research measured clicks on traditional results falling from 15 percent of visits to 8 percent when a summary was present. The damage is concentrated in informational queries.
Two studies are worth knowing in detail, because almost everything written about this subject is a paraphrase of one of them.
Pew Research Center tracked the real browsing of 900 United States adults through 68,879 Google searches in March 2025. An AI summary appeared on 18 percent of them. The published findings:
| Behaviour | No AI summary | AI summary present | Reading |
|---|---|---|---|
| Clicked a traditional result | 15% of visits | 8% of visits | Roughly half the click rate when an AI summary is present |
| Clicked a link inside the AI summary | Not applicable | 1% of visits | Being cited in the summary is not a traffic strategy |
| Ended the browsing session | 16% of visits | 26% of visits | The answer was sufficient, so the journey stopped |
Ahrefs approached it from the other side, comparing 150,000 queries with an AI Overview against 150,000 informational queries without one, using click through rates from aggregated Search Console data. Their December 2025 update reports a 58 percent lower click through rate for the top ranking page on affected queries, up from 34.5 percent in their April 2025 study. The position one numbers:
| Group | Position 1 CTR, Dec 2023 | Dec 2025 | Note |
|---|---|---|---|
| Control: informational queries, no AI Overview | 0.076 | 0.039 | Fell by about half over two years anyway |
| Queries with an AI Overview present | 0.073 | 0.016 | Started level with the control, ended at under half of it |
| Forecast for AI Overview queries without one | 0.073 | 0.037 | What they would have been, modelled on the control's decline |
Read the control column before the headline. Click through rates on informational queries fell by about half over those two years whether or not an AI Overview was present, which means some of what gets blamed on AI is a broader shift in how results pages are built. The AI Overview effect is the gap between 0.037 forecast and 0.016 actual, and it is large. Ahrefs also reports the reduction shrinking down the page, about 51 percent at position two and about 19 percent at position ten.
Both are correlational and both cover mostly informational intent. Neither measures commercial queries specifically, which is the gap that matters most for software companies and the reason the next section exists.
Which search queries still send clicks to SaaS companies?
Pricing, alternatives, comparison, integration and implementation queries still send clicks, because a summary cannot answer them responsibly or because the searcher wants to compare several named options. Definitional, statistical and list queries have lost the most and should no longer carry a content budget.
Our working classification, which we revise as we see new data from client accounts rather than treating it as settled.
| Query type | Still worth it | Why |
|---|---|---|
| Pricing and cost queries | Strong | An AI summary cannot responsibly answer what your software costs. These queries also sit closest to a purchase decision than almost anything else you can rank for. |
| Alternatives and comparison queries | Strong | Searchers want a human judgement between named options, and they click several results to triangulate. One of the few places where ranking fourth still earns the visit. |
| Integration and compatibility queries | Strong | Does it work with our stack is a specific, verifiable question where the searcher needs your documentation rather than a paraphrase of it. |
| Implementation and migration queries | Moderate to strong | Long, specific, and frequently asked by someone who has already decided to buy something and is now assessing effort. |
| Review and vendor evaluation queries | Moderate | Often dominated by review platforms rather than vendors, but the searcher is in market and worth competing for. |
| Definitional and how it works queries | Weak | Exactly what AI summaries answer best. Still worth having for topical coverage and citation, but budgeting for traffic here is a mistake. |
| Statistics and list queries | Weak | The format a summary replaces most completely. These were the backbone of the old SaaS content playbook and they have lost the most. |
The pattern underneath it: AI summaries are good at consensus and poor at judgement. Where the honest answer is it depends on your situation, and the searcher knows that, they still click. Where the honest answer is a paragraph of established fact, they no longer need to.
A consequence worth stating plainly. The surviving queries have lower search volume than the ones that died. A keyword tool will rank your old informational targets above your new commercial ones on volume every time, which is why query selection now needs an argument per query rather than a sort by volume.
Is SaaS SEO still worth funding?
SEO is still worth funding for software companies whose buyers search commercial queries, and it is no longer worth funding as a traffic programme. The arithmetic works on fewer pages against higher intent queries, and it fails on the volume based content strategy that dominated the last decade.
The arithmetic, with the uncomfortable parts included. Suppose a genuinely good deep page against a commercial query reaches a few hundred visits a month, which is realistic for the long tail in B2B software. At a 2 to 4 percent conversion to a qualified enquiry, that page produces a handful of opportunities a year. Whether that is excellent or worthless depends entirely on your contract value, which is why no agency can answer this question for you from a deck.
Where it clearly works: contract values in the thousands or tens of thousands, a category people actively search, and a product specific enough that a real page can out answer the competition. Where it clearly fails: low contract values that need volume the surviving queries cannot supply, or categories where nobody is searching yet.
The honest comparison against paid search is that the cost structures are opposite. Paid search costs money per click forever and starts working this week. Organic costs effort up front and then costs almost nothing per visit, and takes months. For the same cluster of commercial queries, running both and letting the paid data tell you which organic pages to build first is usually better than choosing. That is how we sequence it alongside paid search when a client runs both.
One number for context on why ranking here is valuable at all: the query best saas seo agency carries a cost per click around 42 dollars. Buyers in this market are expensive to reach by advertising, which is exactly the condition that makes an organic position worth the work.
What should a SaaS SEO service actually deliver each month?
A SaaS SEO service should deliver query selection with reasoning, pages deep enough to be the best available answer, the finite technical work that unblocks indexing, deliberate internal linking, and reporting that names which queries produced opportunities. Anything else is activity.
Query research against the pipeline, not the volume tool
The deliverable is a ranked list of queries with an argument for each: who searches it, what they are deciding, and what they would have to see to book a call. Volume is the fourth column, not the first.
Pages deep enough to be the best answer
In this cluster that usually means two to five thousand words that answer a real question completely, with specifics a competitor cannot copy without doing the same work. Thin pages against commercial queries no longer place.
Technical work that actually blocks indexing
Crawl and index coverage, internal linking, page speed, structured data, canonical and redirect correctness. Necessary, finite, and not a monthly retainer in itself once it is done.
Internal linking as a deliberate structure
Hub pages, deep pages beneath them, and links that pass authority toward the pages that convert rather than toward the blog archive.
Measurement against opportunities
Rankings and sessions as diagnostics. Qualified opportunities and pipeline by landing page as the report. If an agency cannot produce the second, you cannot tell whether the first is working.
The word finite in the third item is deliberate. Technical SEO on a normal software site is a project with an end, not a permanent monthly cost. An agency still billing for technical work in month nine should be able to name what broke.
On reporting, the test we would apply to any agency including ourselves: ask for a list of the queries that produced qualified opportunities last quarter, with the landing page for each. An agency that can produce it is measuring the right thing. An agency that responds with a traffic chart has told you what it optimises for.
Why do SaaS SEO programmes fail?
SaaS SEO programmes fail most often because they were built to produce volume against informational queries, which is the exact category of traffic that AI summaries absorbed. The second cause is content produced faster than anyone could check it, leaving pages with nothing specific enough to prefer.
Volume at the expense of intent
Hundreds of posts against informational queries was a defensible strategy when informational queries sent clicks. The datasets above are the obituary for that approach, and plenty of agencies are still selling it.
Content produced faster than it can be checked
Generated content is cheap enough that the binding constraint became review, and most teams removed the review. Pages with no specific, verifiable claims have nothing a search engine or a reader can prefer.
Topical authority as the entire plan
Covering a subject comprehensively helps. Covering it with fifty pages nobody needed is a cost centre with good internal linking.
Reporting on what moved rather than on what matters
Impressions rose, average position improved, traffic grew eleven percent. None of those sentences contain a customer. Ask which queries produced opportunities.
Treating AI citation as the new goal
Being named in an AI answer is worth having and is not worth planning a budget around. Pew measured clicks on links inside AI summaries at one percent of visits.
The uncomfortable version of the first point is that a lot of SaaS content marketing was always weakly connected to revenue and the AI change simply removed its cover. A programme producing twelve posts a month that generated traffic but never appeared in a sales conversation was already failing on the metric that mattered. It is just no longer producing the traffic either.
What does technical SEO involve for a SaaS site?
Technical SEO for a software company is mostly about three things: making sure the marketing pages render without JavaScript, keeping the application out of the index, and deciding where documentation and the blog live. The generic checklist matters far less than those decisions.
Software companies have a specific set of technical problems that a general SEO checklist misses entirely.
Client rendered marketing pages. Plenty of SaaS sites are built in the same framework as the product, which means the marketing pages render in the browser rather than on the server. Search engines do execute JavaScript, and they do it slower, less reliably, and with no guarantee for a page that fetches its content after load. Server rendering the pages you want to rank removes an entire category of problem that is otherwise very hard to diagnose.
The application leaking into the index. Logged out states, shared links, invite pages and password reset screens all get indexed more often than people expect, and they are thin pages on your own domain competing for your own brand terms. The fix is deliberate robots and noindex rules on application routes, decided once.
Where documentation lives. Technical documentation is frequently the most linked and most searched content a software company owns. On a subdomain it builds authority that does not accrue to the marketing site. In a subfolder it does. There are legitimate engineering reasons to choose a subdomain, and if you do, know that you are making that trade deliberately rather than by default.
Faceted and generated pages. Integration directories, template galleries and comparison matrices generate large numbers of near identical URLs. Each one needs to answer a distinct question or it should not be indexable. This is where programmatic page strategies usually go wrong, and the test is simple: if two pages differ only by a swapped variable, they are one page.
Beyond that, the ordinary work applies. Crawl and index coverage, correct canonicals, no redirect chains, reasonable Core Web Vitals, structured data that describes what is actually on the page. Necessary, measurable, and finite.
Agency, consultant or in house?
Choose by what is scarce. An agency suits companies that need both strategy and production capacity. A consultant suits companies with writers who need direction. In house suits companies with subject matter experts and someone to coordinate them, which is the cheapest option and the hardest to staff.
| Option | Suits you when | The real risk |
|---|---|---|
| Agency | You need strategy and the hands to execute it, and nobody internally has the time | Generic output, because the people writing do not use your product. Mitigate by insisting on engineer review time. |
| Independent consultant | You have writers and a publishing path but no clear direction on what to target | Capacity. One person cannot also produce the pages, so the plan outruns the execution. |
| In house hire | You have subject matter experts willing to contribute and want the knowledge retained | Hiring well. A good in house search lead is hard to assess if nobody interviewing them knows the discipline. |
| Nothing yet | The category has no search demand, or you need pipeline this quarter | Spending a year finding out that search was the wrong channel for your stage. |
The honest note on our own position: we are the first row, and the first row has the generic output risk. The way we manage it is to require subject matter review on technical pages and to say so at the start rather than discovering in month three that nobody has an hour to spare.
How do you judge an SEO agency's previous work?
Ask for the specific queries that produced qualified opportunities for a client last quarter with the landing page for each, check whether pages they wrote two years ago still rank today, and ask what they recommended cutting. The third question is the most revealing.
Four checks that are harder to pass than a case study.
Do their old pages still rank? Ask for work from two years ago and look it up. Pages that ranked and then fell away suggest either thin content that aged badly or a strategy aimed at queries that have since been absorbed by AI summaries.
What did they recommend not doing? An agency that has never advised a client to stop producing content, drop a keyword cluster or delay a project is either very lucky with clients or not giving advice. Ask for an example and listen for specifics.
Who writes, and do they understand the subject? Ask who would write your technical pages, whether you can speak to them, and how they get the specifics they do not know. A process that involves interviewing your engineers is a good sign. A process that does not should worry you.
Does their own site rank for anything commercial? Not decisive, because plenty of good practitioners neglect their own marketing, and it is informative. If an SEO agency does not rank for its own service terms, there should be a reason they can explain.
What about link building for SaaS?
Links still matter and most of the ways they are sold do not work. Paid placements, guest post networks and directory packages are a known pattern that search engines discount or penalise. What works for software companies is original data, integration partners and being the source other people cite.
We do not sell link building, which makes this easier to say plainly. The market for links is largely a market for the appearance of links, and the price of a placement correlates with how easy the site is to buy rather than how much the link is worth.
What has worked for software companies in our experience, in order of durability. Original research or data nobody else has, because it gives writers a reason to cite you. Integration and partner directories, which are legitimate, relevant, and frequently forgotten. Customer stories published by the customer rather than by you. Documentation and open source work, which developers link to without being asked. And being specific enough in public that journalists and analysts come to you for a quote.
All of those are slower than buying links and none of them stops working when someone updates an algorithm. If you want a faster option, paid search buys attention honestly and the SaaS PPC page covers what that costs.
How long does SaaS SEO take to pay back?
Expect six months to a useful signal and twelve to a fair verdict. Long tail commercial queries can move in two to four months, head terms in competitive categories take longer, and anybody promising a quarter is either targeting queries nobody searches or has not measured a programme to completion.
Month one
Technical and research Crawl and index audit, query research against your actual buyers, a published plan with the reasoning for each target. Fixes that unblock indexing go live immediately.
Months two and three
The first commercial pages Three to six deep pages against the highest intent queries, usually pricing, alternatives and integration angles. These are the pages that pay back first.
Months four to six
Movement and the first honest read Rankings establish on the long tail, and the first opportunities with a traceable landing page appear. Head terms are usually still moving.
Months seven to twelve
Compounding The pages that work get extended, the ones that do not get cut or rewritten. This is the window where the channel either becomes your cheapest source of pipeline or demonstrably does not.
Two things change this timeline materially. An established domain with existing authority moves faster than a new one, sometimes much faster. And a site with a publishing path that requires engineering involvement for every page moves slower than one where the content team can ship, which is a constraint we check before quoting a timeline rather than after.
Should you optimise for being cited in AI answers?
Being cited in AI answers is worth having and is not worth planning a budget around. Pew measured clicks on links inside AI summaries at 1 percent of visits to those pages. Citation is useful as brand presence, and it is not a traffic channel.
The practical steps that help are the ones that were already good practice. Answer the question directly in the first paragraph under each heading, in language that can be quoted without surrounding context. Publish specific, verifiable, attributable facts rather than consensus paraphrase, because a model summarising five sources that all say the same thing has no reason to name you. Keep the page structure clean enough to parse. Use schema where it genuinely describes the content.
What we would not do is reorganise a content strategy around citation metrics. The measurement is immature, the platforms change the interface frequently, and the one solid number anyone has published puts the click value at roughly nothing. Write pages that deserve to be the source, report on opportunities, and treat citation as a side effect worth watching.
How much should SaaS SEO cost?
We bill hourly against an agreed scope rather than a monthly retainer, so a month of research and three deep pages costs what it costs and a quiet month costs less. Deep commercial pages are the main expense and the main source of return.
For shape: a first ninety days covering the technical audit, query research and three to six deep commercial pages typically lands between 60 and 120 hours depending on how much the site needs unblocking and how much subject matter review your team can absorb.
Production
$24 to $45/ hour
Page build, research support, on page work, internal linking implementation, reporting.
Indicative until we have seen the site and the search landscape. Engagements start at ninety days with written outcomes, then month to month with thirty days notice. We will tell you inside the first month if we think the category cannot support the channel.
When should you not hire a SaaS SEO agency?
Do not hire an SEO agency when your category has no search demand, when you need pipeline inside a quarter, when nobody internally can review technical content, or when shipping a page takes a release cycle. Each of those defeats the channel regardless of who runs it.
Your category has no search demand yet
If nobody is searching for the problem, there is nothing to rank for. Demand creation is a different discipline and SEO is the wrong budget line.
You need pipeline inside a quarter
The honest minimum is six months to a useful signal and twelve to a fair verdict. If the runway is shorter than that, paid search and outbound are the right answers and we will say so.
Nobody internally can approve technical content
The pages that rank in this cluster contain specifics only your team knows. If no engineer or operator can review drafts, the output will be generic and will not place.
The site cannot be changed without a quarter of planning
If every page requires a release cycle, the iteration speed that makes SEO work is not available. Fix the publishing path first.
The third one is the most frequently underestimated. The pages that rank against commercial software queries contain details only people who have built or operated the product know. We can research, structure, interview and write, and we cannot invent your implementation specifics. An hour a week of a competent engineer's attention is usually the difference between pages that place and pages that read like everyone else's.
If what you actually need is the broader marketing function rather than search specifically, the B2B SaaS marketing page covers it, and the marketing audit publishes our full diagnostic framework so you can run it yourself before hiring anyone.
FAQ
Common questions
What does a SaaS SEO agency do?
A SaaS SEO agency selects which search queries a software company should own, builds pages that deserve to rank for them, fixes the technical problems blocking indexing, designs internal linking, and reports on the pipeline those pages produce. Query selection is now the decision that matters most, because informational and commercial queries have diverged sharply.
Has AI search killed SaaS SEO?
AI summaries have roughly halved clicks on affected queries rather than killing the channel. Pew Research found clicks on traditional results fell from 15 percent of visits to 8 percent when an AI summary appeared, across 68,879 searches. The damage is concentrated in informational queries, while pricing, comparison and integration queries still send clicks.
Which queries still send clicks?
Pricing and cost, alternatives and comparisons, integration and compatibility, and implementation or migration queries. These either cannot be answered responsibly by a summary or involve a judgement between named options that searchers want to triangulate themselves. Definitional, statistical and list queries have lost the most traffic.
Is SEO still worth the investment for B2B SaaS?
SEO remains worth funding when your buyers search commercial queries and your contract values are in the thousands or above, and it is no longer worth running as a traffic programme. A good deep page against a commercial query might produce a few hundred visits and a handful of opportunities a year, which is excellent at a high contract value and worthless at a low one.
How long does SaaS SEO take?
Six months to a useful signal and twelve to a fair verdict. Long tail commercial queries can move within two to four months, and head terms in competitive categories take considerably longer. An established domain with existing authority moves faster than a new one.
Should we optimise for AI citations?
AI citation is worth having and not worth budgeting around. Pew measured clicks on links inside AI summaries at 1 percent of visits to pages with a summary. Answer questions directly, publish specific verifiable facts rather than consensus paraphrase, keep structure clean, and treat citation as a side effect rather than a target.
How much does a SaaS SEO service cost?
We bill hourly against an agreed scope rather than a monthly retainer. A first ninety days covering technical audit, query research and three to six deep commercial pages typically runs 60 to 120 hours. Deep commercial pages are both the main expense and the main source of return.
Why do most SaaS SEO programmes fail?
Because they were built to produce volume against informational queries, which is precisely the traffic AI summaries absorbed. The second common cause is content produced faster than anyone could review it, leaving pages with no specific or verifiable claims that a search engine or reader could prefer over the alternatives.
Is programmatic SEO still viable for SaaS?
Programmatic pages still work where each page genuinely answers a distinct question with distinct data, such as integration or comparison pages built from real detail. They fail where the pages are templates with a variable swapped, which describes most of what was published under that name and most of what no longer ranks.
Should we build SEO in house or hire an agency?
Choose by what is scarce. An agency suits companies needing both strategy and production capacity, a consultant suits companies with writers who need direction, and in house suits companies with subject matter experts and someone to coordinate them. In house is the cheapest and the hardest to staff well.
Does link building still work for SaaS?
Links still matter and most of the ways they are sold do not work, because paid placements and guest post networks are a known pattern that gets discounted. What works for software companies is original data, integration and partner directories, customer published stories, documentation and open source work, and being specific enough in public to be quoted.
What technical SEO issues are specific to SaaS sites?
Three: marketing pages that render in the browser rather than on the server, the application itself leaking into the index through logged out and shared link states, and whether documentation sits on a subdomain or in a subfolder. Generated integration or comparison pages are the fourth, where each page must answer a genuinely distinct question.
What should we ask an SEO agency before hiring?
Ask for the queries that produced qualified opportunities for a client last quarter, with the landing page for each. An agency measuring the right thing can produce that list. An agency that answers with a traffic or impressions chart has told you what it optimises for, which is not the same thing as pipeline.
Keep exploring
Related
- SaaS PPC agencyBuy the same queries while the pages mature
- B2B SaaS marketingThe whole function, not one channel
- SEO and programmatic contentThe same work outside software
- B2B marketing auditThe full framework, published
- SDR vs BDRWhat inbound replaces, and what it does not
- Free B2B data enrichmentBuild the list without credits
Let's talk
Tell us what you're building.
Get a free 30-minute strategy call. You'll leave with a clear plan, a realistic estimate and our honest opinion, whether or not you work with us.
- Email usmail@voliom.com