The SaaS Cold Outreach Sequence That Actually Earns Links (Reply Rates, Not Open Rates)

Most outreach advice optimises the wrong number. Here is the sequence we actually run, why I stopped caring about open rates years ago, and the honest version of what link outreach looks like inside a SaaS agency.

Let me start with something most agencies will not put in writing: the average cold outreach email gets a reply 8.5% of the time, which means 91.5% of them are ignored (Backlinko, 2019, from a study of 12 million outreach emails). That number has not improved. B2B cold email reply rates actually slipped to 5.8% in 2025, down from 6.8% two years earlier (Belkins, 2025, analysing 16.5 million emails). So if your outreach machine is built to chase opens, you are polishing the one metric that does not predict a single link.

We do not run cold outreach the way the blog posts describe it. We are not firing a thousand “I loved your article” emails a week at journalists, begging for one editorial link. What we actually run is partner outreach and sales outreach, mostly through Instantly, and we outsource a fair chunk of the manual grind so the strategy time goes where it counts. This is the version of outreach I would hand a Series A marketing lead who has been burned by an agency promising “50 links a month” and delivering a folder of DR 12 blog comments.

Why open rate is a vanity metric

Open rate stopped meaning anything in 2021, when Apple shipped Mail Privacy Protection. Around 45.5% of all email opens now come from Apple clients, and the majority of those are preloaded by Apple on delivery, not opened by a person (Litmus, 2026). So a 60% “open rate” can include thousands of opens that never happened.

Here is the mechanism. Apple’s Mail Privacy Protection fetches the tracking pixel in your email the moment the message hits the inbox, whether or not a human ever reads it (Paubox, 2025). Your email platform records that pixel fetch as an open. Multiply that across the roughly 55-60% of opens that Apple’s privacy features touch, and your open-rate column is fiction.

This is not a small calibration error. It is the difference between a number you can manage by and a number that actively lies to you. I have watched marketing leads celebrate a 70% open rate while their campaign secured zero links, because nobody was watching the reply column. Open rate tells you the subject line was not flagged as spam. That is the entire signal. It says nothing about relevance, nothing about whether a relationship started, and nothing about links.

The metric that survives all of this is the reply. A reply requires a human to read your email, decide it is worth answering, and type something back. You cannot fake that with a preloaded pixel. So when we report to clients, the open-rate row is gone. We report reply rate, positive reply rate, links secured, and link quality. Four numbers, all of which a human action sits behind.

Why generic outreach templates fail for SaaS

Templated outreach fails for SaaS because the buyer is technical and over-pitched. Editors at SaaS publications see the same agency templates 50 times a week, and pattern recognition is the enemy. The reply rate on a recognisable template is close to zero, no matter how healthy the open rate looks.

There are three signals that flag your email as a template inside the first three seconds: a subject line that is trying too hard, an opening sentence about how much you “love their content”, and an ask that arrives before any value. A technical buyer clocks all three at a glance. They have seen the move before, probably twice that morning.

SaaS makes this worse than most niches. The pool of relevant publications is small, the editors talk to each other, and the same five agencies are pitching all of them. If your email looks like the other four agencies’ emails, you have lost before you have started. The job is not to write a better template. It is to not look like a template at all.

The data backs the personal angle hard. Personalising the body of an outreach email lifted response rates by 32.7% in Backlinko’s study of 12 million emails (Backlinko, 2019). That is a third more replies from one change, and it is the change every templated campaign skips because it does not scale cleanly. The honest pivot for us came when we stopped treating outreach as a link-per-email transaction and started treating it as relationship-building for guest-post placements that carry multiple partner links. One good guest relationship is worth more than 200 cold pitches.

The four-email sequence we actually use

The sequence is simple: Day 0 a contextual pitch (not an ask), Day 3 a follow-up with a specific reason to reply, Day 7 a genuine value-add, and Day 14 a polite close-out. Four touches, 14 days, then stop. Backlinko found a single follow-up generates 65.8% more replies than a one-and-done send, and a multi-step campaign to multiple contacts beats a single email by 160% (Backlinko, 2019).

That 160% number is the whole argument for running a sequence instead of a blast. People are busy, your first email arrives at a bad moment, and the follow-up catches them on a better one. The mistake is not following up. The mistake is following up badly, or following up forever.

Email 1: The contextual pitch

No ask. You are opening a door, not walking through it. Reference something specific the recipient published or shipped, and float the collaboration in one sentence. The whole email is under 80 words. The point of Email 1 is to earn the right to send Email 2, nothing more.

Email 2: The follow-up with a reason

Sent Day 3. The mistake here is “just bumping this to the top of your inbox”. Give them an actual reason: a specific angle, a data point they would find useful, a named overlap with their work. This is the email that the 65.8% follow-up lift comes from, so it earns its place only if it adds something the first email did not.

Email 3: The genuine value-add

Sent Day 7. This is where you give before you get. A free idea, a relevant stat, a linkable asset to anchor the pitch they could reference. If the first two emails did not land, this one often does, because it is the only one in the set that asks for nothing.

Email 4: The polite close-out

Sent Day 14. “I will stop emailing, but the offer stands if it is ever useful.” Counterintuitive, but this is consistently our highest-reply email. People reply to the close-out two months later, when the timing finally suits them. Following up past 14 days, on the other hand, just trains them to ignore you and risks your sending reputation.

The subject lines that actually get replies

Boring wins. Our best-performing subject lines are “Quick question” and “Quick collaboration”. They read like an email from a person, not a campaign. The clever, curiosity-gap subject lines the outreach blogs recommend consistently underperform for us, because they pattern-match to marketing, and marketing gets archived.

Here is the thing about subject lines: their only job is to not look like outreach. A subject line that is working hard to be opened is announcing itself as a campaign. The unremarkable ones slip through because they look like the hundred other one-line internal emails the recipient gets a day. “Quick question” does not sell anything. That is exactly why it works.

And remember what we established about open rate. You cannot even trust the subject-line A/B test the way you used to, because the open data feeding it is polluted by Apple’s preloading. So judge subject lines on the reply column, not the open column. A subject line that produces a slightly lower “open rate” but a higher reply rate is the winner, every time.

How to pick the right targets

Target selection accounts for most of the outcome. The right pitch to the wrong publication is wasted. We filter on three things: niche relevance at the page level (not just the domain), organic traffic trending up, and topical match to the specific page that would carry the link. DR matters less than traffic on the actual linking page.

The rule I keep repeating to the team: a DR 80 domain with a dead, traffic-less page is worth less than a DR 45 page that ranks and gets read. Links live on pages, not domains. If nobody visits the page your link sits on, you have bought a link that does nothing for rankings or referral traffic. This is the same logic behind high-authority backlinks done properly: relevance and real traffic over a vanity DR score, every time.

For SaaS specifically, we prioritise publications where the audience overlaps the client’s ICP, not just the topic. A link from a general marketing blog is fine. A link from a publication your buyers actually read is worth five of them, because it does double duty: it passes authority and it sends qualified referral traffic. The off-page SEO checklist for SaaS goes deeper on the anchor-text and relevance discipline that turns a placement into a ranking move rather than just a logo on a list.

Targeting is also where the economics live. The average paid guest-post link runs $77.80, and a niche edit (a link inserted into existing content) averages $361.44 (Ahrefs, excluding content cost). Those are real numbers, and they are why one-link-per-email outreach rarely pencils out. You are paying in time what others pay in cash, so the targeting has to earn it back through relationships that produce more than one link.

Personalisation without writing 200 emails by hand

The contradiction in outreach is that personal sells but you cannot personally write hundreds of emails a week. We solve it with an 80/20 split: 80% of the email is a tested structure, 20% is a genuine first sentence about the recipient’s recent work. The 20% is where the 32.7% response lift lives (Backlinko, 2019).

That first sentence is the only part of the email that cannot be a template, and it is the part most campaigns get lazy about. The structure underneath it can absolutely be systematised. What cannot be systematised is the observation that proves you actually looked at the prospect’s recent posts before you hit send. A real observation buys you the next three sentences of attention. A fake one loses you all of them.

One warning that matters more every month: AI-generated “personalisation” that is obviously generated is the new spam signal. “I really enjoyed your insightful piece on [topic]” written by a bot reads worse than no personalisation at all, because it signals you did not even bother to fake it convincingly. The bar is a sentence a human would actually write after two minutes on the prospect’s site, not a mail-merge field dressed up as warmth.

Why we moved from Clay to a Claude workflow

We used to run research and personalisation through Clay. We have since moved to a Claude-based workflow, which does a better job of reading a prospect’s recent posts and drafting a first line that sounds like a human noticed something specific. The research step is where AI genuinely earns its place in outreach. The send step is where it gets you blocked.

The distinction is worth labouring. AI is excellent at the work that happens before the email: scraping a prospect’s last five posts, spotting a genuine overlap with the client, and drafting an opener a human can approve or rewrite in ten seconds. That collapses the research time that used to make personalisation impossible at volume. It does not replace the judgement; it removes the grunt work around it.

What AI is bad at is being trusted with the whole email unsupervised. The moment you let it write and send the full message, you are back to templated spam, just generated faster. So our workflow keeps a human on the opener and the decision to send. The Claude step makes the human faster, not absent. That is the only configuration I have seen actually lift reply rates rather than volume.

Guest-post outreach: multiple links per placement

The reason we lean into guest-post and partner outreach over classic editorial chasing is simple arithmetic. A cold pitch to a journalist, at an 8.5% reply rate, is built to win one link if everything goes right. A genuine guest relationship produces a placement that can carry multiple contextual links, and then produces more placements over the following year.

More than half of SaaS SEO teams already outsource link building (56%, per Editorial.Link’s 2025 survey of 518 teams), and the ones that get value from it are not buying single links. They are buying access to relationships an in-house team does not have the time to build. That is the actual product: not a link, but a standing arrangement with publications your buyers read.

This is also why I am sceptical of the “one editorial link per heroic pitch” model that dominates link-building Twitter. It is a great story and a terrible business. The reply rates do not support it, the cost-per-link does not support it, and it leaves you starting from zero every single month. Relationships compound. Cold pitches reset. We covered the deeper version of this in the digital PR for SaaS guide.

Tracking what actually matters

Open rate is the metric to stop reporting. It is easy to game, easy to misread, and after Apple’s Mail Privacy Protection it is largely fiction. The metrics that matter are reply rate, positive reply rate, link-secured rate, and link quality. Those four tell you whether the campaign is working. Open rate just tells you the subject line was not offensive.

Here is the chain that actually matters: a positive reply rate of even 3-5% on a well-targeted SaaS list is healthy, because a positive reply is a relationship, and relationships compound. From positive replies you get guest-post placements, and a single guest-post relationship can carry multiple partner links over time. That is the maths that makes outreach worth doing. One-link-per-email never pencils out, especially once you price in the $77.80 it would cost to just buy the equivalent placement.

Metric	What it tells you	Worth reporting?
Open rate	Subject line was not spam-flagged (and even that is polluted by Apple)	No (vanity)
Reply rate	Email was relevant enough to answer (8.5% is average)	Yes
Positive reply rate	You started a real conversation (3-5% is healthy)	Yes (the one I watch)
Link-secured rate	Conversations becoming links	Yes
Link quality	Page traffic plus topical fit of secured links	Yes

What does this look like when it compounds over a real engagement? Here is one client where outreach-led links did the heavy lifting on revenue, not just rankings.

Full numbers in the Prospeo case study.

Deliverability is an infrastructure problem, not a copy problem

Most outreach that lands in spam does not have a copy problem. It has a sending-reputation problem. Warm the domain, keep daily volume sane, authenticate properly (SPF, DKIM, DMARC), and send plain text. Fix the infrastructure first, because the best email in the world earns nothing from the spam folder.

This is the unglamorous half of outreach that the templates never mention. You can write a perfect, personal, well-targeted email, and if you are blasting it from a cold domain with no authentication and a 500-a-day volume, it will never see an inbox. The reply rate you measure is downstream of deliverability you may not be measuring at all.

Our rule is to keep sending volume per inbox low and human-looking, run dedicated sending domains separate from the client’s primary domain, and never let a campaign’s bounce rate creep up unmonitored. This is also the main reason we use a platform like Instantly rather than sending from a single mailbox: the infrastructure controls are the point. Once deliverability is solid, the copy and targeting decisions actually get a fair test.

What I would do differently starting from scratch

If I were rebuilding EMGI’s outreach from zero, I would go relationship-first far earlier. I spent too long optimising the cold-email machine when the real leverage was in a smaller set of genuine partner relationships that each produce links for years.

I would also be honest about outsourcing sooner. The manual outreach grind is real work, but it is not strategy, and treating it like strategy slows you down. Build the targeting and the angle in-house, where the judgement lives, and get help with the volume. And I would cap follow-ups at four from day one instead of learning the hard way that email five just annoys people and dents your sending reputation.

The last thing: I would track positive replies from week one and ignore open rate entirely. We would have reached the version of outreach that works a lot faster if I had been watching the number that actually predicts links, rather than the one that just looks good in a screenshot. If you want the bigger picture this outreach work sits inside, start with our SaaS link building pillar.

---

---