Outbound Compass1. Select accounts before contacts
A repeatable B2B prospecting system starts with an account universe, not a pile of email addresses. Define the company characteristics that make the problem more likely to exist: industry, size, geography, business model, technology, growth stage, hiring pattern, or another observable condition. Mark which filters are required and which are merely useful signals.
Then inspect a sample manually. The purpose is to catch false positives in the account logic before you pay to enrich contacts. If 30% of sampled companies are obviously wrong, the problem is not the database—it is the targeting definition.
Account-selection output
- A searchable account definition
- Explicit exclusions
- A reason each included account is plausibly in-market
- A priority tier if some accounts deserve more research or more channels
2. Map the buying roles
Once an account qualifies, identify the people who own, influence, or feel the problem. Role mapping should change with company size. A founder may own outbound tooling at a 15-person SaaS company; a Revenue Operations leader, SDR manager, IT stakeholder, and procurement process may appear at a larger company.
Create a simple hierarchy: primary owner, alternate owner, influencer, and possible blocker. Decide how many contacts you are willing to approach at one company and what happens when one person replies. This prevents accidental over-contacting and makes company-level stop rules easier to configure.
3. Enrich only fields that drive a decision
Enrichment is useful when it changes qualification, routing, personalization, or channel choice. Company size can determine segment. Technology can support a relevant integration angle. A recent job posting can suggest a timing signal. A direct dial can enable a phone step. Collecting fields that never affect a decision creates cost and more data to maintain.
Keep source fields separate from inferred fields. If an enrichment system labels an account “likely expanding outbound,” preserve the evidence that produced that inference so a human can evaluate it later.
4. Normalize, deduplicate, suppress, and verify
Before outreach, normalize company names and domains, deduplicate people and accounts, apply customer and opt-out suppressions, and verify email addresses. These steps are operational controls, not administrative cleanup. They reduce duplicate outreach, bad handoffs, and avoidable bounce risk.
When several sources contribute data, choose a record-of-truth rule. For example, prefer a verified work email from the most recent source, but retain the original source and verification date. The goal is reproducibility: another operator should be able to understand why a lead was considered campaign-ready.
5. Match the outreach motion to account value
Not every prospect deserves the same amount of research. Tier high-value accounts for deeper manual research or multichannel touches. Use lighter, segment-level personalization for broader but still qualified cohorts. The cost of research should be proportional to expected account value and the strength of the available signal.
| Tier | Typical treatment | Best when |
|---|---|---|
| 1 | Manual account research, tailored message, multichannel | High contract value / small account universe |
| 2 | Segment-specific copy plus verified account context | Moderate value / repeatable pattern |
| 3 | Highly standardized message to a narrow ICP | Lower research budget / strong common trigger |
6. Turn replies into structured feedback
The prospecting loop is incomplete until response data changes future targeting. Tag replies by reason: wrong role, wrong account, bad timing, existing solution, no priority, interested, or unclear proposition. Tie those categories back to the source segment and list logic.
If “wrong role” dominates, revise contact mapping. If one industry repeatedly says the problem is not important, remove or isolate it. If a particular trigger produces a higher share of qualified conversations, weight that signal more heavily in the next account-selection pass.
7. Review the system at the funnel level
Track metrics that correspond to each stage. Account acceptance rate evaluates targeting. Verified-contact yield evaluates data sourcing. Bounce patterns evaluate list hygiene and sending readiness. Positive replies evaluate message/offer fit. Meetings and pipeline evaluate whether the process is producing commercial value.
A weak downstream metric should be traced upstream before changing everything at once. This makes prospecting a controlled operating process rather than a series of disconnected tool exports.