How PassReady, an online exam-prep startup, lifted activation 29% with behavior-based email segmentation built on Spreeflo. A lifecycle marketing case study.
Spreeflo Team
Author
A free library of past exam questions was sending PassReady, an online exam-prep startup, tens of thousands of visitors a month. Almost none of them were signing up.
PassReady helps candidates prepare for high-stakes professional qualifications: CFA, CIMA, ACCA, the bar exam. Its paid product is a course library with AI-graded mock papers and a personal study planner. But the thing pulling in the most search traffic was the free side of the site: a public Past Paper Question Bank with thousands of real exam questions and worked solutions.
Plenty of people were arriving. Few were staying. Some bounced after a single question. Others came back four or five times, browsed an entire syllabus, and still never created an account.
So the PassReady team stopped treating every visitor the same.
Behavior-based email segmentation is the practice of grouping marketing contacts by what they actually do (pages visited, events fired, content consumed), not by who they are on paper. The pitch is straightforward: someone who's read three pricing pages this week is a different prospect than someone who's read three blog posts. They should get different email.
It's the foundation of most modern lifecycle marketing, and it's what separates a "send one newsletter to everyone" program from one that compounds. PassReady's story is a clean example of what behavior-based segmentation looks like in practice for a content-led business.
The fix started with one move: segment visitors by what they actually did on the site, in real time.
Using the Spreeflo SDK, every page view in the Past Paper Bank fired a custom viewed_question event tied to the visitor, anonymous or identified. That stream fed two live segments. "Low-intent" meant two or fewer question views in the last 30 days. "High-intent" meant repeated returns and multiple syllabi browsed: the pattern of someone actually preparing for a sitting.
In Spreeflo, segments aren't a nightly export. They're evaluated continuously, with nested AND/OR conditions on attributes and behavior. So a visitor coming back for a third time crossed the threshold and entered the high-intent journey on that visit. No batch job, no manual review.
We had thousands of monthly visitors and no real read on which of them were actually preparing for an exam. The moment we started treating two-time visitors and five-time visitors differently, the funnel started behaving.
The team built two parallel journeys in Spreeflo's sequence builder, each triggered by a Join Segment event.
The low-intent journey was educational. Short emails framing how to approach each section of the exam. A worked-example walkthrough. A reminder of which papers tend to be tested most heavily. The goal wasn't to convert anyone immediately. It was to give the visitor a reason to come back, and to start associating PassReady with "useful study companion" rather than "another SEO landing page."
The high-intent journey was conversion-focused. These visitors had already invested time. They got messaging about the mock-exam simulator, candidate success stories from their specific qualification, and a clear, low-friction path to a free trial of the AI-graded papers.
Both journeys ran continuously. A visitor moving from one tier to another transitioned automatically, without being re-imported or re-tagged. That sounds like a footnote, but it's where a lot of behavior-driven programs quietly break: the segments live in one tool and the journeys live in another, and contacts get stranded in the gap.
The biggest unlock wasn't the segmentation itself, it was running the two journeys in parallel. Once they were separate, we could iterate on either one without worrying about breaking the other. Half the work in lifecycle marketing is just keeping campaigns from colliding.
The parallel-journey setup PassReady used lives in Spreeflo's visual sequence builder, with Join Segment triggers, Random Split tests, and Check Email Activity branching out of the box.
With the two journeys live, the team ran two quick experiments using Spreeflo's built-in tooling.
The first was a CTA placement test. The original Past Paper Bank page had a single sign-up prompt at the bottom. A Random Split in the journey routed half of identified visitors through a variant with a second CTA mid-scroll. The variant added a 0.7% click lift. Small in isolation, meaningful at search-traffic scale.
The second was an email-timing experiment in the high-intent journey. After the first email, a Check Email Activity branch split openers from non-openers. Openers got a longer wait before the next send; non-openers received a shorter, more direct follow-up. The result: the second and third emails in the sequence kept performing instead of falling off a cliff.
Neither of these were sweeping rewrites. They were the kind of small, behavior-aware tweaks that compound when you have the infrastructure to run them without burning a week of engineering time.
Over the test period, return traffic from low-intent visitors climbed 29%. The educational journey was doing exactly what it was built to do: pulling people back to the site to study, which is where conversions actually happen. High-intent visitors went a step further, with page views per visitor up 12% and a matching lift in free-trial sign-ups.
The structural change was bigger than either of those numbers, though. A chunk of traffic that had previously been invisible to the funnel was now identified, segmented, and moving through it.
None of that came from sending more email. It came from sending the right email to the right segment at the moment behavior earned it.
What we didn't expect was how much of the lift came from the low-intent journey. We'd assumed the educational emails were a soft consolation prize for visitors who weren't going to convert anyway. Turns out they were pulling people back to the site, which is where everything actually happens.
It's tempting to treat free content as a top-of-funnel cost: necessary for SEO, but not really part of the activation story. PassReady's experience suggests the opposite. The bank wasn't the leak. The lack of segmentation around it was.
Every visit to a free resource is a signal. Two visits is a stronger signal. Five visits over two weeks is a near-certain indicator that someone is preparing for an exam they care about. Treat all three the same and you flatten what could be a precise targeting layer into noise.
Most educational startups in this space are small teams running a content-led growth motion. For them, this is the kind of work that punches above its weight. You don't need a bigger team or a larger paid budget. You need infrastructure that can read behavior, sort it cleanly, and act on it without anyone watching.
If you're sitting on organic traffic that converts at a rate you suspect is too low, the answer is almost never "more traffic." It's usually "more intent-aware messaging on the traffic you already have." That's the work Spreeflo is built for, and a free question bank is a surprisingly good place to start.
Spreeflo gives content-led teams the SDK, segments, and behavior-driven journeys to run lifecycle marketing like PassReady's, at a fraction of what most platforms charge.
PassReady's team let the parallel journeys run for about eight weeks before reading the results. Behavior-based programs need a few full cycles of the underlying behavior to produce stable numbers. For a return-visit segment defined on a 30-day window, two cycles is the practical floor. Shorter and you're reading noise.
There's no hard floor, but the math gets uncomfortable below a few thousand visitors a month. The high-intent segment in a behavior-based setup is usually a small fraction of total traffic, and the lower the volume, the longer you wait for each result to mean something. Below roughly a thousand visitors a month, attribute-based segmentation (plan, role, signup source) is typically a better starting point than behavior-based.
Cleaner experimentation. Two separate journeys mean you can swap, pause, or rewrite either one without touching the other, and your analytics stay tidy. Branching inside a single journey is fine when the two paths share most of their content. Here, the low-intent and high-intent journeys had almost nothing in common, so splitting them up was the obvious call.
The terms overlap and most teams use them interchangeably. The distinction worth drawing: behavior-based segmentation is descriptive ("this visitor viewed three syllabi this week"), intent-based segmentation is the reading you put on top of that behavior ("therefore they're studying for a sitting"). In the PassReady setup, the segments themselves were behavioral, but the journeys were built around an intent reading. You're almost always doing both.
Discover more insights on similar topics
How TripStash moved from 0.4% to 9.8% browser push click-through by climbing from broadcast to behavioral to contextual. A three-tier framework with a worked example.
How CodeMosaic, a global developer community, runs event-triggered email at scale on Spreeflo, hitting a 13% CTR across multi-language lifecycle journeys.
Most small teams say they do event-driven marketing. Few actually do. Four timing failures that masquerade as automation, and how to fix them.