SaaS outbound has a different objective from most cold email: the ask is usually a demo, not a conversation. That changes both what you write and what you measure — the leading indicator is a click on your demo link, not a reply, and the fastest win is testing a new positioning without rebuilding a campaign from scratch.
A reply starts a conversation; a click on your demo link shows intent before anyone types anything. In a self-serve or low-touch motion the click is the step that predicts pipeline, and Mailor tracks it per lead — so a campaign that produces clicks and no replies is working, just on a different signal.
A new ICP in Mailor starts from a one-sentence product description — or your product URL. Change the sentence and the inferred profile changes with it, including industry, company size, buyer roles and technographic signals. That makes positioning tests cheap enough to run several times a month instead of once a quarter.
| Test dimension | What you change | What it tells you |
|---|---|---|
| Segment | The product description | Which industry or company size responds at all |
| Tone | Direct, value-first, curiosity, social-proof, story or auto | Which framing survives contact with the audience |
| Language | Any of 18 languages | Whether the bottleneck is message or market |
| CTA | Demo link, docs link, or a question | Whether the audience wants to see it or discuss it |
Short, specific, and aimed at one click. The structure that holds up: one line on their situation, one line on what changes for them, one clear next step. Word-count guidance in the editor keeps drafts inside the range that cold-email practice consistently favours, and a spam pre-check runs before the send.
Volume should follow fit, not precede it. Start by sending to your top-scoring leads and read the click rate before expanding — a list that produces clicks at scale is worth adding to, and one that only produces sends is not. Sending from a domain you control, with SPF, DKIM and DMARC in place, keeps delivering while you scale.
Describe the product, review the inferred ICP, generate leads scored against it, draft per prospect, and send from your own domain. Link clicks and replies are tracked per lead, so the loop from positioning test to signal to iteration is measured rather than guessed. Credits are consumed per action, so a test that produces no engagement is also a test that cost almost nothing.
It works when the targeting is specific and the ask is small. Broad, generic outreach in a saturated category is what fails. Inferring a narrow ICP and asking for one click on a demo is what makes the numbers work in crowded markets.
Change the product description used for ICP inference and regenerate. In Mailor the target profile is derived from that description, so a repositioning means editing a sentence and re-running lead discovery, not rebuilding a campaign.
For self-serve products, a demo link usually outperforms — it lets an interested prospect act without committing to a conversation, and the click is measurable. Open-ended questions suit enterprise deals where a conversation is the goal.
Judge on clicks rather than volume. Once a meaningful share of your top-scoring leads has been contacted, the click rate tells you whether the positioning lands. Expanding a list that produces no clicks multiplies cost without information.