This is a step-by-step guide to finding every company in your market that is hiring people to do outbound right now.
Outbound means contacting people who haven’t asked to hear from you: cold calls, cold email and LinkedIn messages. Companies that are building an outbound team say so in their job ads, in the title and again in the duties.
We’ll go through one example: Callwick, a call-coaching tool for new outbound teams. It sells to US B2B companies with 20 to 300 employees.
Say the search returns 6,400 job ads. They lead to 1,150 companies, and you find 1,520 people with a buying title at 860 of them.
This guide is for founders and anyone who sells to sales teams. If you already have job ads with their full text in a sheet, skip to step 3.
How a hiring list works
A job ad tells you two things: the company is hiring, and what the new person will do. The title alone often isn’t enough. “Account Executive” can mean calling strangers or answering leads who asked for a demo. So this list comes from the full text of each ad.
Steps 1 and 2 pull the ads. Step 3 reads every ad and decides whether the role includes outbound. Steps 4 to 6 turn the ads into companies and people. Step 7 is what you write to them, and step 8 keeps the list current.
1. Pick the job titles, the filters and the time window
Search the titles where outbound is the whole job first, then the titles where it’s sometimes part of the job. The first group gives you matches you can almost trust from the title alone. The second group gives you more ads, and the full-text check in step 3 sorts them.
- Outbound by name: SDR (sales development representative), BDR (business development representative), Outbound Account Executive, Founding Account Executive.
- Same job, other names: ADR (account development representative) and MDR (market development representative). Many US companies use these, and a search for “SDR” misses them.
- Sometimes outbound: Account Executive, Inside Sales Representative. Many of these handle inbound leads, people who contacted the company first, or existing customers. Search them only together with an outbound word in the ad text, like “cold call” or “prospecting”.
Let the source filter by country and company size only. Those two are safe to filter early. Industry labels on job boards are often wrong, so check the industry yourself later, on company data (step 6).
Pick a time window you can act on. Pull the past month once to build the first list. Then pull the past week every Monday.
For Callwick, the search plan:
titles_outbound: [SDR, BDR, Outbound Account Executive, Founding Account Executive,
Account Development Representative, Market Development Representative]
titles_sometimes: [Account Executive, Inside Sales Representative]
text_words: [outbound, cold call, cold email, prospecting] # only for titles_sometimes
country: [United States]
employees: [20, 300]
window: past month once, then past week every Monday
That’s 8 titles, so 8 searches.
2. Choose a job data source that returns the full ad text
Two questions decide the source. Does each row carry the full text of the ad? And does the provider sell the data under its own terms, instead of scraping a site that forbids it? Without the text you can only judge the title, and step 3 is the step that makes this list accurate.
| Source | How it charges | Full ad text | Good for |
|---|---|---|---|
| BlitzAPI | Flat rate: one fixed price a month, job ads included | A short summary per ad. Its keyword search covers the full text | Wide searches by title and country, plus company and people data from the same API |
| TheirStack | Credits per month, one credit per ad | Yes, with a filter on words in the text | Full text from many job boards in one place |
| Greenhouse, Lever, Ashby | Free | Yes | One company you already know, or checking an ad again |
| LinkedIn job scrapers on Apify | Per ad, from $0.0004 on Apify’s free plan | Yes | Cheap wide pulls. They scrape LinkedIn, and LinkedIn’s User Agreement prohibits scraping |
A few things the table doesn’t show:
- Greenhouse, Lever and Ashby are the hiring tools many tech companies use. Each publishes every company’s job board through a free public API, a way for a script to ask for the data, with the full text. They don’t help you find new companies, because you need to know the company first.
- BlitzAPI also covers step 6, because company size and people come from the same API. If you start there, get the full text of the ads you keep from the company’s own job board or careers page.
- LeadMagic sells job ad data through an API too, next to its email tools.
Test each source on 50 ads before you commit. Run the same search with a limit of 50 and check four things: does the text in the row match the live ad, how many rows carry the company’s LinkedIn page or website, is the post date right, and how many rows are the same ad twice.
Then let Claude Code, an AI assistant that writes and runs scripts on your computer, build the pull. Put your API key, the password your script uses, in a file called .env, open the Terminal in that folder, type claude and paste:
Read search_plan.yaml and the job search API docs at <link to the docs>.
Build tools/pull_ads.py. One search per title, always filtered by
country, employee range and the time window. Titles under
titles_sometimes also need one of the text_words in the ad text.
Save every response unchanged to raw/<title>/page_<n>.json and skip
pages that already exist, so a crash never means pulling again.
Before the run, print the expected number of ads and credits and
wait for my yes. Then write ads.csv with job_id, title, company_name,
company_url, website, location, posted_at, ad_url and the full text.
For Callwick, say you pick TheirStack, because it returns the full text and filters on words in it. The 8 searches return 6,400 ads, which use 6,400 credits. That fits the $169 plan with 10,000 credits a month.
Two title searches can return the same ad, so remove duplicates by job ID first. That leaves 5,300 unique ads.
3. Read the full text, because a keyword isn’t a need
A keyword match tells you a word appears in the ad. It doesn’t tell you the new hire will do outbound. Here are six ad lines that all match “outbound” or “cold”:
- “You’ll open new accounts through cold calls and LinkedIn outreach.” The new hire does outbound. That’s the ad you want.
- “No cold calling, all leads come from our marketing team.” Matches “cold calling” and says the opposite.
- “Join a sales floor with an outbound team of 14 SDRs.” The keyword describes other people at the company. Whether that’s a lead depends on what you sell.
- “Experience with outbound prospecting is a plus.” A requirement. Weaker than a duty, but it counts.
- “We’re recruiting for our client, a fast-growing software company in Austin.” A staffing agency, so the advertiser isn’t the employer.
- “Outbound call center agent, collections.” A different job. Call centers and warehouses use the word too, so expect plenty of these.
Read the words in front of the keyword. A heading like “About our client”, or a “no” or “without” right before the match, decides the label more often than anything after it.
Here’s the prompt the model gets for every ad:
You label job ads for one question: will the person hired into THIS role do outbound prospecting?
Outbound means cold calls, cold email or cold LinkedIn outreach to people who haven't asked to hear from us.
For every place the ad mentions outbound, cold outreach, prospecting, pipeline generation or new business,
give one label:
OWN_DUTY the hire does it
REQUIREMENT the hire must have done it before
NEGATED the ad says this role does NOT do it
TEAM_CONTEXT someone else at the company does it
AGENCY_CLIENT a recruiter advertises for another employer
OTHER_MEANING the word means something else (outbound logistics, call center collections, support calls)
BOILERPLATE company description, benefits, legal text
A title that names the job (SDR, BDR, ADR, Outbound ...) counts as OWN_DUTY, unless the text says
the role works inbound only. Read the 300 characters before each match as well as the match itself.
Verdict:
OUTBOUND_IN_ROLE at least one OWN_DUTY or REQUIREMENT, and no AGENCY_CLIENT
OUTBOUND_ELSEWHERE the only real matches are TEAM_CONTEXT
NO everything else
Also return:
stage: FIRST_SELLER if the ad says first sales hire, founding role or build the playbook from scratch, else UNKNOWN
reports_to: the title this hire reports to, if the ad names it
Quote the sentence behind every label. No quote, no label.
Output one JSON line per ad: {job_id, company_url, verdict, stage, reports_to, labels: [{label, quote}]}
Test the prompt on ads you labelled yourself. Before the full run, label 30 ads by hand, with a few of each kind from the list above. Run the prompt on them and fix it until it agrees with you. Then read 10 answers from the real batch next to their ads, fix the prompt when it’s wrong, and read 10 new ones. Run the whole batch once two rounds of 10 in a row need no fix.
For Callwick, say 3,100 of the 5,300 ads come back as OUTBOUND_IN_ROLE and 500 as OUTBOUND_ELSEWHERE. The other 1,700 are agencies, inbound roles and other meanings. Keep the OUTBOUND_ELSEWHERE ads in their own tab, because a team that already does outbound might still want coaching, with a different email.
4. Remove reposts and competitors before you group
Two filters cost nothing and run on the ads themselves:
- Reposts. Same company, same title, almost the same text, new job ID. Keep one, with the earliest post date.
- Competitors. For Callwick that’s sales trainers, outbound agencies and other coaching tools. They hire SDRs too, and they won’t buy.
Keep the ad location and the company’s headquarters as two columns. The ad location comes from the search. Where the company itself sits comes from company data in step 6.
For Callwick, say that takes the 3,100 ads down to 2,600.
5. Group the ads into one row per company
You sell to companies, so you need one row per company. A company hiring three SDRs shows up three times in the ads and should get one email sequence.
- Group on the slug of the company’s LinkedIn URL, the part after /company/. LinkedIn URLs can start with a country code, so the same company can arrive as ca.linkedin.com/company/x and www.linkedin.com/company/x. Group on the full URL and one company becomes two rows.
- Keep the proof. Per company: the number of outbound ads, the titles, the first and last post date, the best quote and every ad link.
- Sort by stage. Companies hiring their first salesperson go on top, because they’re building from nothing. Then companies with two or more outbound roles, then single roles.
The prompt for this step:
Group classified_ads.jsonl into one row per company.
Key: the slug after /company/ in company_url, lowercased. Ignore the country part of the host.
Keep only verdict OUTBOUND_IN_ROLE.
Columns: company_name, slug, outbound_ads, titles (unique, joined with " | "), first_posted,
last_posted, stage (FIRST_SELLER if any ad has it), reports_to (most frequent),
best_quote (from an OWN_DUTY label), ad_urls.
Tier: A = FIRST_SELLER, B = 2 or more outbound ads, C = the rest.
Print the count per tier and the 10 companies with the most ads, so I can check them by hand.
For Callwick, say the 2,600 ads come from 1,150 companies: 130 in tier A, 310 in tier B and 710 in tier C.
6. Check each company, then find everyone with a buying title
The ad gives you the company’s LinkedIn page. Size, industry, headquarters and the people come from a company and people database.
Start with GetLeads and BlitzAPI. Both are flat rate: one fixed price a month instead of a price per contact. Use Prospeo last, because you pay per contact there. Look each company up by the LinkedIn page from its ad.
Check the company first. Keep it when it sells to businesses, has 20 to 300 employees and has its headquarters in the US. Watch for subsidiaries: a US page can show 60 employees for a group with thousands abroad. Read the About text for “part of” before you trust the size.
Remove the companies you shouldn’t contact. Before you buy a single email, take out current customers, open deals and everyone who asked not to be contacted. Match them on the website domain against your CRM.
Then pull every person with a buying title, not one person per company. For Callwick that’s whoever will coach the new reps: founder and CEO at the smaller companies, head of sales, VP of Sales and sales manager at the larger ones, plus the title in the reports_to field from step 3. The ad often names the new hire’s manager, and that’s the person who’ll listen in on the first calls. Two things to check:
- Title filters are loose. A search for the CEO can return the CEO’s executive assistant. Read the exact title afterwards and keep an exclude list.
- People move. Databases can be months behind. Check that the person still lists the company on their profile.
Emails come last. Remove duplicate people first, so you never pay twice. Then build the usual formats (first.last@, first@, flast@) and check them with an email verifier like MillionVerifier, which asks the mail server whether the mailbox exists without sending anything. Pay a finder like Prospeo or LeadMagic only for the rest. Keep “valid” results and drop every “unknown”. Catch-all addresses, on mail servers that accept any address, go out only after a tool like LeadMagic or BounceBan confirms them.
For Callwick, say 960 of the 1,150 companies pass the company check and 940 are left after removing customers and open deals. You find 1,520 people with a buying title at 860 of them.
7. Write about the role, never about how you found it
The ad is public, and mentioning it is normal. How you found it stays in your sheet.
You can say which role they posted and when, written as their situation: “You posted an SDR role in January.” You can also say what the role means for them, like new reps or a first sales playbook. Leave “saw” and “noticed” out of it.
Never say:
- “Our system flagged your company”, “I track hiring signals” or “you have 3 open SDR roles according to our data”.
- A tier, a score, a label from the model, or a long quote from the ad held up as proof.
First lines for Callwick, one per tier:
- A, first seller: “You posted your first sales role in January. In the first months someone has to listen to those calls. Will that be you?”
- B, a team: “You posted SDR roles in December and January. When two or three reps start at the same time, who listens to their calls?”
- C, one role: “You posted a BDR role in January, with cold outreach as part of the job.”
All three are dated and in the past tense, so they stay true after the ad closes. After the line, add one offer and one soft question. If LinkedIn is your channel, there’s a separate guide that turns this list into a LinkedIn campaign.
8. Pull again every Monday and check the ads before each follow-up
A hiring list goes out of date within weeks. Four mistakes come up again and again:
- Lists built from titles only. Without the full text, step 3 can’t run, and every “Account Executive” counts as outbound.
- Counting ads instead of companies. Budget and emails run on companies. 6,400 ads sounds like a market. 860 companies with a buyer is the list.
- Ads that closed. Check before every follow-up. A closed job disappears from Greenhouse, Lever and Ashby boards, and TheirStack can notify you when a job it tracks closes. A closed LinkedIn ad forwards you to a job search page that loads normally, and its address contains “expired_jd_redirect”. So check the address you end up on, not whether the page opens.
- One pull and done. Hiring changes every week. The Monday pull keeps the list fresh. A company that’s already on it gets a new date instead of a second row.
Where to start today
You can build your first hiring list today in three steps:
- Write the search plan: titles, country, company size and the past month.
- Pull the ads with their full text and let the prompt from step 3 label every ad.
- Group the ads into one row per company and pull everyone with a buying title.


