Learn / Lead lists
How to use Apollo.io without wasting your credits
I booked over 3,000 sales calls in the past year, and Apollo was a big part of how. Not because the tool does anything clever on its own, but because a list built by someone who knows which filters are load-bearing looks nothing like one built by someone ticking every box in the panel.
Most people asking how to use Apollo.io are really asking two questions: which filters actually matter, and how do I get the data out without handing Apollo my whole budget. This is the written version of the training below. Every filter with a verdict, a list built live, and the cheap export path into enrichment.
Where Apollo lists go wrong
Apollo holds over 275 million contacts and over 73 million companies. It also ships a sequencer, a CRM, a dialer, and a website de-anonymizer still in beta. Ignore all of it. The people and company tabs are what you pay for.
Two failure modes account for nearly every bad Apollo list. The first is building it by hand, one search at a time, until you run out of patience. The second is stacking filters, hitting a huge number, and emailing people who were never going to care. The second is what most of the market does, which is exactly why a precise list still wins. No amount of clever copy rescues a list built by guessing. If you are not sure there are even enough buyers to filter for, size the market before you touch the filter panel.
The filters, ranked
There are a lot of filters and almost no guidance on which carry weight. Here is the whole panel, with a verdict on each based on the lists we actually run.
| Filter | What it actually does | Verdict |
|---|---|---|
| Keywords | Matches company description, name, and other text fields. | Load-bearing. Start every list here. |
| Employee size | Headcount at the company. | Load-bearing. It is on every single list we build. |
| Location | Account level is where the company sits. Contact level is where the person sits. | Load-bearing, and the one people set backwards. |
| Email status | Verified, unverified, user managed, update required, unavailable. | Load-bearing. Take all but unavailable, unless you plan to waterfall. |
| Company | Upload domains to include, or a do-not-contact list to exclude. | Load-bearing and badly underused. |
| Job titles | Typed titles, exclusions, or a management-level cut. | Load-bearing, but push seniority to the top. |
| Personas | A saved bundle of job titles you reuse in one click. | Situational. A real time-saver once your buyer profiles repeat. |
| Time in current role | How long someone has held the seat, and a proxy for company longevity. | Situational. Useful when you want established companies and revenue data is unreliable. |
| Lists and saved searches | Stores the exact filter set so you can pull net-new later. | Situational. Worth it on any list you will refresh. |
| Departments and job function | A pre-made grouping instead of typed titles. | Optional. Rarely beats typing the titles yourself. |
| Industry | Apollo's own category tags. | Fluff. Too broad, they overlap, and keywords do the job better. |
| Buying intent and email opened | Behavioral signals surfaced inside the platform. | Fluff. Beta-grade accuracy. Not worth a second of your time. |
The setting almost everyone gets backwards
Location has two sides. Account level means the company is based there. Contact level means the person is. With hybrid work, plenty of people work for US companies from other countries, so filter on the account for US-based companies, and on the contact when you send by time zone. Backwards, it quietly poisons a list that otherwise looks fine.
Building a list, live
Here is the build from the video: manufacturers, specifically the kind that would benefit from a B2B lead generation offer. Nine moves, in order.
- Pick a segment for a reason. CNC and precision machine shops, because they sell complex high-value products to a narrow buyer audience and lean on trade shows. That came out of research, not the filter panel.
- Turn the segment into keywords. Ask a model to expand it into comma-separated keywords, paste those into the keyword filter, and widen the fields it searches: company description, company name, and the rest.
- Set location. United States at the account level, and the contact level too, so communication stays simple.
- Skip industries entirely. The keyword set describes the business better than any category tag will.
- Add job titles, then sanity-check them. Read the list back and ask whether each title would care about your offer.
- Set email status. Verified, unverified, user managed, update required. That alone cuts the count.
- Cut the bottom and the top on headcount. The distribution always skews small, so 1 to 10 holds most of the contacts and almost none of the buyers. Under roughly 5 to 15 people you are emailing solo operators who cannot afford you, and past a few hundred you are in enterprise territory. That build landed on 20 to 50.
- Push seniority up. Manager titles are the most numerous, so left alone they eat the majority of your sends. Prioritize VP, owner, president, co-founder, head.
- Save it as a new searchwith a descriptive name, something like "US manufacturers 20 to 50", so you can reopen it later and pull only who is net-new.
Notice what that build never does: chase volume.
Sample both ends before you export
This is the step everyone skips and the one that saves the campaign. Apollo ranks results from most relevant to least, so page one always flatters you. Open a few company websites from the top, confirm they are who you meant to target, then do the same on the very last page. In that build the tail was page 98. If the back 40% is companies you would never email, the filters are wrong and no export fixes that.
The database is the product. The export is the tax.
Why we do not export from Apollo
Run the credit math once and you stop arguing about it. That 7,000-lead campaign costs 7,000 credits. The organization plan at $119 a month gives you 72,000 credits a year, about 10 list pulls. One campaign a month and you are out.
Accuracy is the second reason. Best case, with the most talented Apollo user running advanced keywords, exclusions, and layered filters, you get roughly 70% accuracy on emails straight from the platform. Great place to find contacts, mediocre place to finish them. For the head-to-head on whether Apollo is even the right database, that comparison lives here.
So we scrape instead. Copy the URL of your people search, drop it into a scraper like Ample Leads, choose how many leads you want, and start. That runs roughly $1.40 to $1.90 per 1,000 leads depending on volume, and bigger lists currently sit in a 24 to 48 hour queue. What comes back is a CSV.
Finish the list somewhere else
The CSV goes into Clay, where the Apollo columns auto-populate and the real work starts. Three jobs. Verify every email with a tool like Million Verifier so you are not paying for bounces. Enrich what Apollo could not tell you: locations, industries served, similar companies, hiring status. Then filter for contacts with no email and run a waterfall across providers until one returns it, merging the results at the end.
That waterfall is the reason to include Apollo's "unavailable" email status in your filters. On its own it is dead weight. Feeding a waterfall, it is free volume. The full build-out of that workbook is here, and if you would rather skip the point-and-click, we drive the same pipeline from the terminal in the Claude Code list system.
tl;dr: keywords first, then size, location, email status, and titles. Ignore the industry tags and the intent signals. Sample both ends, save the search, then export somewhere cheaper than Apollo and finish the data in enrichment.
FAQ
Which Apollo filters actually matter?
Five carry almost every list: keywords, employee size, location, email status, and job titles. The company domain upload is the sixth and most underused, since it lets you drop in domains you sourced elsewhere. Personas, time in role, and saved lists are situational. Industry tags and intent signals are the two to ignore.
Why are Apollo emails sometimes wrong?
Apollo is a contact database, not a verifier, and contacts go stale as people change jobs. Even a careful list from an experienced user lands around 70% accuracy on emails straight out of the platform. Verify every export before sending, and run a waterfall of providers for the contacts Apollo has no email for.
How many Apollo credits do you actually get?
On the organization plan at $119 a month, you get 72,000 export credits a year. One 7,000-lead campaign eats 7,000 of them, so that plan covers roughly 10 list pulls. Launch more than a campaign a month and the credit ceiling, not the data, is what pushes you to export another way.
Should I use Apollo's industry filter or keywords?
Keywords, almost always. Apollo's industry tags are broad and they overlap, so one tag pulls companies with completely different buying problems. Keywords match the company description and name, which is closer to how the business describes itself. Start with keywords, then narrow with size, location, and titles. Tags are fine only when you are not picky.
How do you check an Apollo list is any good?
Sample both ends of it. Apollo sorts results from most relevant to least, so page one always looks great and tells you nothing. Open a few company websites there, then jump to the last page and open a few more. If the tail is companies you would never email, tighten the filters before spending a credit.
Can you export Apollo leads without using Apollo credits?
Yes. Third-party scrapers take your Apollo people-search URL and pull the same rows for roughly $1.40 to $1.90 per 1,000 leads, a fraction of what that pull costs in credits. Expect a 24 to 48 hour wait on bigger lists. What comes back is a CSV you push straight into verification and enrichment.
If you would rather not build any of this yourself, that is what we do: we build the list, set up the cold email system, and turn interested replies into calls on your calendar. Or keep learning free. Every training, by topic, sits on the trainings hub, and the full video for this one is on YouTube.
PS - the filter panel is the easy half. Deciding which segment deserves a campaign is the half nobody films, and it is why the build above starts with research, not a filter.