Outbound playbook

    How to personalize outbound at scale without hours of manual research

    By Eric Sprauve · Published

    Most outbound teams are forced into a bad trade: research every prospect by hand and never reach enough people, or automate the work and send messages that sound like everyone else's.

    That tradeoff disappears when personalization becomes a system instead of a writing trick. The system decides who deserves research, gathers evidence against a repeatable rubric, turns that evidence into a point of view, and carries the point of view through every touch.

    200 free credits · No credit card required

    The research-to-outreach workflow

    1. 01

      Discover

      Find buyers from fit and signals.

    2. 02

      Qualify

      Research accounts against a reusable scorecard.

    3. 03

      Engage

      Use evidence and your style guide to draft a sequence.

    4. 04

      Follow up

      Carry context into replies and the next touch.

    Workflow illustration, not a product screenshot. Follow the product tutorials to inspect the setup, or explore the product stages.

    From evidence to a relevant message

    Evidence
    A fictional company’s careers page lists its first two account executive roles. Its product page describes a move into enterprise accounts.
    Implication to test
    The team may need a repeatable way to research larger accounts as its sales motion changes. The job postings do not prove a pipeline problem.
    Message
    Saw you’re hiring your first two AEs as you expand into enterprise. Is account research still something each rep assembles from scratch? We help small teams turn account evidence into outreach in their own voice. Happy to show what that could look like for one of your target accounts.
    Review before sending
    Check that the postings are current, the enterprise expansion is explicit, and the offer fits the recipient’s role. Keep the operating problem framed as a question.
    Illustrative example with fictional inputs, written for this guide. It is not a customer message or a captured AI output.

    Personalization is not a first line

    Mentioning a funding round or recent LinkedIn post is not enough. If the rest of the message could be sent to the prospect's nearest competitor, it is still a template.

    Useful personalization connects three things:

    1. a fact about the account or person;
    2. what that fact plausibly changes now; and
    3. why your offer is relevant to that change.

    The goal is not to prove that you looked someone up. It is to show that you understand why a conversation could be useful.

    The five-part operating system

    1. Start with a narrow problem, not a broad list

    Define the trigger, account shape, buyer role, and disqualifiers before sourcing. “VP Sales at software companies” is not a motion. “New sales leader at a 50–300 person B2B company building its first repeatable outbound program” is much closer.

    2. Qualify before you personalize

    Use the same scorecard for every account. Research enough to grade fit, urgency, and evidence quality. Deep personalization on a bad-fit account is still wasted work.

    3. Separate evidence from inference

    Store the source fact, then state the implication as a hypothesis. This protects the message from invented certainty and makes review faster.

    4. Teach the system your voice

    A style guide should learn structure, sentence length, directness, vocabulary, and how your team changes tone by audience. It should not merely ban a list of cliché phrases.

    5. Review exceptions, not every word

    Humans should review low-confidence research, sensitive claims, and high-value accounts. They should not rebuild every sequence from a blank page.

    A practical quality rubric

    Score each draft from 0–2 on five dimensions:

    Dimension012
    SpecificityCould go to anyoneNames a factConnects a fact to a unique implication
    AccuracyUnsupportedPlausibleDirectly evidenced
    RelevanceGeneric painRole-level painCurrent account/role tension
    VoiceObvious templateAcceptableSounds like the sender
    RestraintOverclaimsSome hedgingClear hypothesis with honest uncertainty

    Do not launch low-scoring drafts just to hit a volume target. Fix the research rule, prompt, or source data that caused the failure.

    What this changes for a lean team

    TCS Basys used researched personalization to enter the automotive-service market, a vertical the team had not sold into before, and reported a first executive meeting in under 30 days and a $16M qualified pipeline from two target accounts. These are one customer’s reported results, not a prediction for another team. The repeatable lesson is that research quality and speed reinforce each other once the research is systematic instead of manual.

    Frequently asked questions

    What is personalized outbound?

    Personalized outbound uses evidence about a specific account and buyer to make the message's reason, timing, and value relevant to them. It is more than inserting names or surface-level facts into a template.

    How much research should each prospect get?

    Research depth should follow account value and evidence availability. Automate the repeatable collection and grading; route uncertain or strategically important cases to a human.

    Can AI-written outreach sound human?

    Yes, when the system has relevant evidence, a real point of view, and a style model grounded in the sender's writing. Generic inputs still produce generic output.

    How do you measure personalization quality?

    Track qualified replies, interested replies, meetings, edit time, and research time. Open rate alone cannot tell whether the message was relevant.

    Eric Sprauve writes about Strama’s product and outbound workflows. This guide is published by Strama and reflects our perspective; evaluate the fit against your own team’s requirements.

    Try the workflow on your own prospects

    Judge the research, message specificity, voice, and review time with 200 free credits.