SaaS link building

Most software companies still buy links the way they bought them in 2018. A retainer, fifteen placements a month, a spreadsheet sorted by domain rating, and a line item nobody has ever audited against pipeline. The links arrive on schedule. Rankings on the pages that matter do not move, because the pages that need links are never the pages getting them.

That gap is the whole problem. Fix the source of your links and the routing of them, and the budget question mostly answers itself.

Which link tactics died and which still compound

Three things changed between 2023 and 2026, and each one killed a tactic that used to be a staple. Google’s March 2024 core update, which the company said was designed to cut low-quality and unoriginal content in results by 40 percent, hit the exact kind of thin syndicated content that guest post networks are built from. The site reputation abuse policy that followed made hosting paid third-party content on a strong domain an active liability rather than a revenue line. And Cision shut down Connectively, the service formerly called HARO, in December 2024, which ended the era of a junior marketer firing off forty generic pitches a day.

Worth saying plainly: nobody I know who ran a placement-volume program through 2024 and 2025 can show a ranking chart that justifies it. The programs that survived scrutiny were the ones producing something a stranger would link to without being asked.

The four link sources that repeat for software companies

These four are not a menu. They are a sequence, and most companies can only run two of them well at a time.

Original data studies. You have a dataset nobody else has, sitting in your production database. Gong built an entire content function on aggregated sales call analysis. Vanta publishes trust and compliance survey data. Ramp publishes corporate card spending data that trade press picks up because no other source has it. The link comes from the fact that the number cannot be reproduced elsewhere. This is the single highest-yield source for B2B software and it deserves its own treatment, which is why we wrote digital PR for SaaS as a separate method page.

Free tools. A small, genuinely useful utility earns links every month without anyone pitching it. Ahrefs gives away a backlink checker, a keyword generator and Ahrefs Webmaster Tools, and those pages collect links from blog posts, university resource pages and forum answers indefinitely. HubSpot’s Website Grader did the same job for a decade. The engineering cost is real, usually four to eight weeks of one developer, but it is a one-time cost against a permanent asset.

Integration and partner pages. If your product connects to Stripe, Slack, Snowflake, Salesforce or Zapier, there is a partner directory listing, a co-marketing slot and usually a joint setup guide available to you. These links are relevant by construction and they are the only fast source available to a site with no authority yet. They also map neatly onto the bottom of funnel keyword patterns that convert, so the page earns twice.

Founder-led commentary and podcasts. A founder with a real opinion, appearing on twelve industry podcasts a year and writing two contrarian posts a quarter, generates links that no agency can buy. PostHog’s public company handbook is linked by hundreds of domains because it says things other companies keep private. The cost is founder time, which is the scarcest input in any company, and that is exactly why this source has no competition.

What is worth paying for, and what never is

Split the budget by what money can actually buy. Money buys data you do not have, design you cannot produce, and access to journalists you do not know. Money cannot buy relevance, and every tactic that tries to buy relevance directly is the one that gets penalised.

Worth paying for:

• Data collection. A panel survey of 400 qualified respondents through a research vendor costs roughly 6,000 to 18,000 dollars and gives you a defensible sample size.
• Design and charting. A designer who turns your findings into ten chart images with proper alt text, at 2,000 to 5,000 dollars, roughly doubles pickup because journalists reuse the graphics.
• A PR person with real relationships. Either a freelance ex-journalist at 3,000 to 8,000 dollars a campaign, or an agency with named trade contacts in your category.
• Engineering time for a free tool. Treat it as product work with a product spec, not as a marketing side project.

Never worth paying for:

• Guest post networks, at any price, on any domain rating.
• Per-link packages, which are the same thing with a cleaner invoice.
• Link insertions or niche edits into existing articles.
• Bulk media database subscriptions bought so an intern can send 300 untargeted pitches.

The test I apply before approving any link spend: if the tactic stopped working tomorrow because Google changed a policy, would we still have an asset? A data study leaves you a dataset and a methodology page. A free tool leaves you a tool. Sixty guest posts leave you nothing but a cleanup project.

What a link actually costs by tactic

These ranges assume you count internal staff time at a loaded rate, which almost nobody does, and which is why in-house programs look artificially cheap on paper.

The paid guest post row looks cheapest, which is exactly how the category sold itself for a decade. The durability column is the one that matters. A link from a site with no organic traffic and no audience passes nothing and ages into a cleanup task.

Quarterly data asset versus monthly outreach retainer at 6,000 dollars a month

Same annual budget, 72,000 dollars. This is the decision most heads of growth are actually making, so here it is with numbers attached.

Approach	Annual spend	Referring domains, year one	Where the links point	State at month 18
Monthly outreach retainer	$72,000	90 to 180, mostly low relevance	Blog posts chosen by the agency	Links decaying, program must continue or output stops
Four quarterly data assets	$72,000 ($18k each)	100 to 400, high relevance	Four study pages you control and can update	Assets still earning links with no new spend

Note the honest part of that table: two of the four studies will underperform. That is the base rate. A study that nobody covers still leaves you a page you can cite internally and a method you can rerun, which is more than a dead guest post leaves you, but it is a real cost and you should budget for a 50 percent hit rate rather than pretending otherwise.

My position: for almost every SaaS company below 50M ARR, one genuinely linkable data asset per quarter outperforms any outreach retainer at the same spend. The exception is a company in a category with no proprietary data and no partner ecosystem, which is rare, and usually means the real problem is positioning rather than links. If that is you, the retainer is a stopgap while you find something worth saying, not a strategy.

Anchor text and velocity limits that keep you out of trouble

Most anchor text advice is folklore. Here is what holds up in practice across software sites.

Exact-match commercial anchors should stay under roughly 5 percent of your total anchor profile. Not because 6 percent triggers something, but because a natural profile for a software brand is dominated by the brand name, the bare URL, and descriptive phrases like ‘their research on onboarding’. If a third of your anchors say ‘project management software’, someone chose them, and it looks like it.

Velocity is relative. A site earning 4 new referring domains a month that gains 180 in one month has some explaining to do. A site earning 60 a month can absorb a 200-domain PR spike without a flicker. Compare against your own trailing twelve-month average, and expect a legitimate launch or funding announcement to produce a spike you can point to.

Two practical limits worth writing into your process. Cap any single campaign at no more than 25 percent of your trailing annual referring domain total. And never point more than a handful of external links at a page you plan to consolidate or redirect within six months, because you will spend the equity twice.

Measure link impact by page type, not by domain rating

Domain rating is a third-party model. It moves slowly, it compresses at the top, and reporting it alone tells a board nothing about whether the pages that produce trials got stronger. Replace it with four measures that map to money.

Page type	What links are supposed to do	The metric to report	Realistic target
Comparison and alternatives pages	Close a competitive ranking gap on a term with buying intent	Referring domains to that URL vs the top three results	Within 30% of the median competitor
Integration pages	Get indexed and hold position 1 to 5 on a low-volume term	Indexation rate and average position	90%+ indexed, top 5 within 90 days
Data studies and free tools	Attract links that the rest of the site can inherit	New referring domains per asset, and internal links out	25+ domains per asset, 8+ internal links out
Money pages (pricing, product, demo)	Almost nothing directly	Internal link count from linked assets	Do not build external links here

The fourth row is the one people argue with. External link building to a pricing page is wasted effort in almost every case, because nobody links to pricing pages voluntarily and a manufactured link to one is conspicuous. Route authority there instead, which is a question of site architecture rather than outreach. Our guide on internal linking for SaaS covers the routing patterns that move equity from an earned asset to a converting page.

Three reporting habits that make link work defensible in a board meeting. Report the referring domain gap against named competitors on your top twenty commercial terms, not your site-wide total. Show the internal link path from each linkable asset to the pages that produce trials. And separate links that brought referral sessions from links that did not, because a link from a publication your buyers read is worth ten from a domain with the same rating and no audience.

If you need to forecast what the ranking movement is worth before you spend, the organic traffic forecast calculator will translate a position change into sessions and trials, which is the language finance accepts.

How this fits the rest of the SEO program

Links are the third input, not the first. Get the SaaS keyword research right so you know which twenty pages actually need authority, build those pages, then decide what to point at them. A team that starts with outreach ends up building links to whatever already exists, which is usually the wrong set of pages.

Tooling matters less than people think here. Ahrefs and Semrush both do the job for gap analysis and monitoring, and the choice between them rarely changes an outcome. We compare the practical differences in SaaS SEO tools, but do not let a tool decision delay the first campaign by a quarter.

What to do next

Pick one of two paths this week. If you have partner integrations and no link program, spend the next 30 days shipping integration pages and claiming directory listings, because that is free authority sitting unclaimed. If you already have those and rankings are still stuck on commercial terms, scope one data study for next quarter and hire the pitching capability before you hire the analyst.

Either way, stop reporting domain rating as a standalone number. Swap it for referring domains to your twenty revenue pages against named competitors, and the conversation about link budget gets much easier. The lesson on earning links and AI citations in the SaaS SEO Sprint walks through the pitch templates and the tracking sheet, and the wider SaaS SEO pillar shows where link work sits against technical and content investment. If the goal behind all of this is pipeline rather than rankings, the SaaS lead generation cluster covers what happens after the click.