How to Find B2B Leads for a Small Business: Define, Search, Verify

Most small businesses do not fail at outreach because their emails are badly written. They fail because the list was wrong before the first email was drafted: the wrong industry, the wrong size of company, the wrong person inside it, or an address that no longer exists. No subject line rescues a message that lands in front of someone who was never going to buy.

This guide is about the work that happens before outreach: deciding who you are looking for, finding them, checking that they are real and reachable, and keeping the list clean. It is deliberately about prospecting rather than about what you write once you have a list. A small, well-chosen list you can personalise will usually beat a large one you cannot.

Key takeaways

Start with the customer, not the database

Filters can only narrow a search as well as you can describe the person you want, and most owners have never written that down. The document that does this is an ideal customer profile, or ICP. It describes the kind of company that gets the most value from what you sell. If you have customers, list the five you would clone and two who were a poor fit; the patterns are most of your ICP.

The five parts of a usable ICP

• Industry. Be specific. “Healthcare” is a sector; “independent physiotherapy clinics” is a market you can search and write to.
• Size. Employee count, and where relevant revenue or number of locations. Size usually decides whether the person you email has budget, and whether they are the buyer or an assistant to the buyer.
• Role. Who owns the problem you solve and who signs off? In a ten-person company that is often the owner. In a two-hundred-person company it may be a head of operations. Pick one or two titles, not twelve.
• Trigger events. Something that makes the problem urgent now: hiring for a role you could replace, opening a new location, launching a product, raising funding, switching tools. A trigger is the difference between a cold message and a timely one.
• Exclusions. Who looks like a fit on paper but never works out. Competitors, agencies that resell, companies below a minimum size, regions you cannot serve, past customers who churned badly. Exclusions save more wasted effort than any other part of the profile.

A worked example

Imagine a three-person agency that builds booking websites and automated reminders for appointment-based businesses. A first attempt at an ICP might be “small businesses that take appointments.” That is a description of a million companies and no way to choose between them. Here is the same idea made searchable.

Every line of the usable version corresponds to a reason the offer is relevant: a clinic with a second location has an immediate scheduling headache, while a franchise with central marketing cannot buy even if it wants to.

Where leads actually come from

With a profile in hand, the question is where to find people who match. Most small businesses combine two or three sources.

• Lookalikes of existing customers. Companies similar to your best customers. Closest to proven demand, and the easiest message to explain.
• Referrals. Ask your best customers who else has the same problem. A warm introduction beats any cold list, though it is slow to scale.
• Directories and associations. Good for niche or local markets that databases cover poorly, but manual, and often only a general inbox.
• Contact databases. Fast and filterable, with named people and emails, but data decays and needs verification.
• Communities and events. People self-select into a topic, but a public post is not an invitation to be mass-messaged.
• Intent and trigger signals. Hiring posts, new locations, funding or tool changes tell you why to write now. Use them to prioritise a list that already fits the profile, not to replace it.

Searching a contact database: filters and why narrow wins

A contact database lets you describe your ICP as filters and returns people who match. The filters you will use most are job title or seniority, company size, industry, location and sometimes technology or keywords. The quality of the result depends almost entirely on how you set them.

Narrow beats broad

A search for “marketing managers at companies with 11 to 200 employees” returns a huge, mixed list you cannot personalise. A search for “owners of independent physiotherapy clinics with two to twenty staff in one region” returns something you can read in an afternoon. Narrow searches also make errors cheaper, let you compare segments, and spend provider credits only on people who matter.

Use filters and exclusions together

Most tools support excluding as well as including: remove certain titles, skip companies above a size, drop a list of domains you already work with. Apply your exclusions from the ICP here, and keep a running list of domains to suppress: existing customers, active deals, people who have opted out. Searching without exclusions is how you end up emailing your own client with an introductory pitch.

A contact record is not a qualified lead

This distinction is worth being strict about, because mixing them up inflates your numbers and wastes your time. A database returns contact records: a name, a title, a company and probably an email. Those are raw material. A qualified lead is a record that has passed some checks.

• Fit. Does the company match industry, size and region? If not, drop it.
• Role. Is this person the owner of the problem or the buyer? If not, find the right person.
• Reachable. Is the email verified and the data recent?
• Reason. Is there a trigger that makes the message relevant now?
• Allowed. Are we permitted to contact them, and are they absent from the suppression list?

You do not need an elaborate scoring model to start. A simple three-bucket judgement works: fits and has a reason now, fits but no reason yet, and does not fit. Only the first bucket gets personalised outreach straight away. The second is for a lighter, slower touch. The third is deleted.

Verify emails before you send anything

Email addresses in any database are a mixture of current, outdated and guessed. People change jobs, companies change domains, and some addresses are inferred from a pattern rather than confirmed. Sending to all of them and hoping is the single most damaging habit in cold outreach.

Why bounces matter

Mailbox providers watch how many of your messages bounce and use it to decide where your future email goes. A high bounce rate looks careless, and legitimate messages, including ordinary ones to customers, start landing in spam too.

What the statuses mean

Verification services and data providers label addresses with a status. The names vary between tools, but they generally fall into three groups.

No verification is perfect. It lowers the chance of a bounce and says nothing about whether the person wants your email, so keep watching your actual bounce rate.

Enrichment: add what helps, keep what you need

Enrichment adds data to a record: company size, phone number, a profile link. Enrich for a purpose, and ask what you will do with each field.

• Keep: name, title, company, verified work email, company size and industry, the reason you chose them, and the source and date you found them.
• Keep if you will use it: phone for a call step, a public profile link, location for time zones.
• Skip: anything unrelated to a business conversation or collected only because a tool offered it.

Recording source and date lets you judge how old the data is and explain to a recipient how you found them.

Start with fifty you can personalise

A long list feels like progress, but fifty well-chosen contacts let you read every record, write one specific sentence about why this company, and learn from the result. When a segment produces conversations, widen it by relaxing one filter at a time so you can tell what caused any change.

Compliance: consent, opt-outs and keeping less

Rules about unsolicited commercial email differ a great deal by country, and sometimes by state or by type of recipient. Some regimes require prior consent for marketing email; others allow business-to-business outreach with conditions such as accurate sender identification and a working way to opt out. Data protection law adds a separate layer on how you collect, store and use personal data, including work email addresses, which are often personal data too.

This is general information, not legal advice. If you sell across borders or into regulated sectors, take advice from someone qualified in the countries you target.

Principles that hold up almost everywhere

• Know your basis. Under GDPR-style regimes, “legitimate interest” is sometimes used for business outreach, but it is not a blanket permission. It generally depends on the message being relevant to the person’s role, on weighing their interests against yours, and on making objection easy.
• Be relevant and honest. Identify yourself clearly and do not disguise the purpose.
• Honour opt-outs promptly and permanently. Keep a suppression list and check every new search against it.
• Keep only what you need, for as long as you need it. Delete records that did not fit.
• Be ready to say where you got it. Record the source.

Measure the list, not just the campaign

When outreach disappoints, the cause is usually either the list or the message, and the two look the same from the outside. Measuring them separately is how you tell which one to fix. You do not need industry benchmarks, which vary too much by market and offer to be useful to a small sender. Compare against your own earlier results instead.

• Bounce rate: bounced divided by sent. It reflects data and verification quality; if it rises, stop and re-verify.
• Reply rate: replies divided by delivered, with positive and negative counted separately.
• Meetings per 100 contacts: meetings booked divided by contacts in the segment, times 100.
• Opt-outs and complaints: a sign you are writing to people who should not be on the list.

Meetings per hundred contacts is the most useful single number, because it lets you compare a list of fifty with a list of five hundred and see which segment is genuinely better. Keep results by segment and source, so that after a few rounds you know which part of the ICP is producing conversations and which is only producing rows.

Mistakes that quietly ruin lead lists

• Buying a giant list. Usually broad, old, resold many times and tied to no profile. Building a small list yourself is slower and almost always better.
• No exclusions. You end up writing to existing customers, competitors and people who have opted out.
• Never cleaning. People move on and addresses expire. Re-verify before reuse and remove bounces and opt-outs after every send.
• Treating enrichment as qualification. A complete row is not a fit. Fit and reason come from your ICP.
• Skipping the reason. Note why each contact was chosen while the answer is fresh.

How Scalry handles this

Scalry has a lead finder that runs searches against data providers; Apollo and Prospeo are supported. You set up a search with saved filters, and each search is stored as a run with a name, its filters, a status (draft, running, completed or failed) and a result count. That means you can see what you searched for, repeat it, and compare runs rather than relying on a spreadsheet from last quarter.

Each result is a lead record with a name, title, email, an email status and phone. Results begin as a preview, so you can look through them and choose which to keep instead of importing everything. That fits the advice above: read the records, keep the ones that fit, and leave the rest. Providers use credits, so a narrow search is also the economical one.

For ongoing prospecting there are lead sourcing plans that can use a saved ideal customer profile instead of manual filters, and a plan can be attached to an agent. Leads you keep can feed the CRM and sequences, so the record you qualified is the one that gets followed up, and you can see its history in one place. The lead finder is available on selected plans.

The tool supports the process; writing the ICP and checking the rules in your market are still yours to do.

Questions to ask before you commit

Whichever way you are leaning, these are the questions that change the answer. Work through them against your own numbers rather than anyone's feature matrix — including this one.

• Can we describe our ideal customer in five lines? If not, searching is premature.
• Do we have exclusions written down? Competitors, existing customers, regions and past opt-outs.
• How do we verify emails before sending? If the answer is that we do not, start there.
• Do we record the source and date for each lead?
• What is our bounce rate on the last send? If we do not know, we cannot manage it.
• Where is our suppression list, and does every new search check against it?
• Which rules apply to the countries we are writing to?

If you cannot answer the first three, you are not ready to choose between Scalry and your current prospecting process yet — you are still deciding what you need, which is a cheaper problem to solve first.

A two-week plan to build your first list

• Days 1 to 3: write the one-page ICP and the exclusions.
• Days 4 to 6: turn it into a narrow search, read the results by hand and adjust the filters.
• Days 7 to 9: verify emails, hold back risky and unknown rows, check the suppression list and add a one-line reason per contact.
• Days 10 to 12: send to the first fifty in small batches and watch bounces daily.
• Days 13 to 14: calculate your three numbers, change one thing, then widen by one filter.

Frequently asked questions

How do I find B2B leads for a small business?

Start by writing an ideal customer profile covering industry, company size, the role you sell to, trigger events and exclusions. Then use sources that match it: lookalikes of your best customers, referrals, directories, contact databases, communities and buying signals. Search narrowly, verify emails and start with a list small enough to personalise.

What is an ideal customer profile?

It is a written description of the kind of company that gets the most value from what you sell. It usually covers industry, company size, the buying role, trigger events that make the problem urgent, and exclusions for companies that look like a fit but never work out. It turns a vague market into something a database can filter.

What is the difference between a contact and a qualified lead?

A contact is a record with a name, title, company and email. A qualified lead is a contact that fits your profile, is the right person for the problem, has a verified email, has a reason to hear from you now and is allowed to be contacted. Only the second kind should get personalised outreach.

Why should I verify emails before sending?

Addresses in any database go out of date. Sending to ones that do not exist produces bounces, and a high bounce rate harms your sending reputation, which can push even your legitimate email into spam. Verification removes the obvious bad addresses, though it is not perfect, so keep watching your real bounce rate.

What do verified, risky and unknown email statuses mean?

Tools word them differently, but verified generally means the address appears to exist and accept mail. Risky means it may work but carries uncertainty, for example a catch-all domain or a role-based address. Unknown means the check could not reach a conclusion. A cautious approach is to send to verified addresses and hold back the rest.

How big should my first lead list be?

Around fifty is a sensible start. It is small enough that you can read every record, write something specific to each company and learn from the results. If a segment produces conversations, widen it by relaxing one filter at a time so you can see what changed.

Is cold outreach to business email addresses legal?

It depends on the country and sometimes the type of recipient. Some places require consent for marketing email, while others allow business outreach with conditions such as clear sender identification and an easy way to opt out. Data protection law may also apply, including GDPR-style rules on legitimate interest. This is general information, not legal advice, so check the rules for the countries you target.