Most explanations of prospecting describe activity: calls made, emails sent, sequences launched. Below: the decision that comes first, which is who deserves a message at all. That decision is where a first outbound campaign is usually won or lost, long before the copy is written.
What does sales prospecting actually cover?
Prospecting covers three jobs. First, build a raw list of companies that match basic criteria: country, size, industry. Second, narrow that list to the companies where your product is genuinely relevant. Third, find the right person in each remaining company, by role and seniority, with a working email address or phone number.
The output is a short list of named people at relevant companies, each with a reason why they might care now. Everything after that, the messages, the follow-ups and the calls, is outreach. Prospecting is a recurring job, not a one-off project. Jeb Blount's Fanatical Prospecting calls the result of skipping it the empty pipe: the deals you close this month came from prospecting done months ago, and a pause shows up later as a gap.
Where does prospecting end and outreach begin?
Prospecting ends when you can answer two questions for every name on the list: why this company, and why this person. Outreach begins with the first message to that person. The line matters because the two jobs fail in different ways. A good message sent to the wrong company gets no reply, and no rewrite fixes it. A plain message sent to the right person at the right moment often works.
Leadspicker's guidance puts it in four words: narrow targeting beats good copy. If you sell to sales teams and write to a developer, the product is not relevant to them, and no amount of wording changes that.
How do you decide who is worth a first message?
You decide in stages, each one removing companies for a stated reason. Start from your ideal customer profile. Pull a broad list by firmographic filters. Then use a lookalike seed, a description or a handful of real customers, to rank companies by how closely they resemble the ones that already buy. Enrich what is left with data you cannot see in a filter: technologies used, open job ads, department sizes. Finally, classify each company against your criteria and keep only the clear fits.
| Step | What it removes | Companies left | Work |
|---|---|---|---|
| 1. Pull from a lead database | Wrong country, size or industry | 500 | About 1 hour |
| 2. Rank against a lookalike seed | Companies that match the filters but not your buyers | 250 | About 1 hour |
| 3. Enrich | No relevant technology, no hiring in the target team | 120 | About 2 hours |
| 4. Classify against criteria | Companies where the product has no clear use | 60 | About 1 hour |
Illustrative worked example. The 500 to 60 ratio comes from the Leadspicker knowledge base; the intermediate counts and the hours are the author's estimates, not measurements.
The final step decides relevance, so state its criteria in writing before you start. If you cannot say in one sentence why a company made the cut, it should not be on the list.
Research on how sellers spend their time explains why selection matters. Salesforce's State of Sales, 7th edition (2025) reports that reps spend 40 % of their week selling and 11 % of their time prospecting, so every hour spent on a poor-fit account comes out of a small budget. RAIN Group's prospecting research, based on 488 buyers and 489 sellers, compares what top performers do differently from the rest. Jeb Blount's Fanatical Prospecting adds the 30-Day Rule: prospecting done in one 30-day period pays off over the following 90 days.
What do most first campaigns get wrong?
Most first campaigns start too wide. A founder exports every company in a sector, writes a message that tries to fit all of them, and sends it to thousands of addresses. Three things go wrong at once.
- Relevance drops: the message has to be generic to fit everyone, so it fits no one in particular.
- Complaints rise: people who have no use for the product are the ones who mark it as spam, and that damages the sending domain for everyone else on the list.
- Volume outruns infrastructure: sending thousands of cold emails a month safely needs many mailboxes across several domains, which a first campaign rarely has.
The fix is to send fewer messages to better-chosen people. A few dozen relevant companies give you a readable result and a domain that is still healthy afterwards.
How does prospecting change with signals and AI?
Signals and AI change how fast the selection runs, not what it is. A buying signal, such as a new hire in the target team, a funding round or a management change, adds the answer to "why now" that a static list lacks. AI classification replaces hours of reading company websites with a written rule applied to every row.
What does not change is the need for criteria. An AI column given a vague prompt returns a vague list. Write the rule a careful person would apply by hand, test it on twenty companies you already know, then let it run on the rest.
Leadspicker's own doctrine is that narrow targeting beats good copy, and it frames enrichment as the difference between a team processing 500 leads and a team processing the 60 worth processing.
Frequently asked questions
What is the difference between prospecting and lead generation?
Prospecting is outbound: you choose the companies and people and contact them first. Lead generation usually means attracting people who come to you, through content, ads or events. Both feed the same pipeline from opposite ends.
How many prospects should a first campaign target?
Fewer than most people expect. A few dozen genuinely relevant companies give a readable result and keep the sending domain healthy. Widen the list only once the message is getting replies.
Can AI do sales prospecting on its own?
AI can apply your criteria to thousands of rows quickly, but it cannot invent the criteria. Write the rule first, test it on companies you already know, and check a sample of the output by hand.
