Learn / Lead lists
Size your market before you send anything
The high-volume approach booked 42 calls in 30 days for one client. The low-volume approach booked 21 calls in 21 days for another. Opposite strategies, both correct, because neither one was a preference. Each was picked from a number: the total addressable market. Every outbound decision you are about to make, how many emails a month, how much infrastructure, how personalized the copy, is downstream of that one figure, and almost nobody calculates it before they start sending.
Skip it and you end up in one of two ditches. You blast a small market at agency volume and have nobody left to contact by campaign two. Or you sit on a huge market hand-writing messages to 400 people a month while 40,000 go untouched. This page is the written version of the training below: sizing your whole market in Apollo, converting that size into a monthly send volume, then pre-qualifying and scoring companies so the data spend lands on people worth reaching.
Count people, not companies
Build the company filter first. For this run I pulled accountants: United States and Canada, 5 to 100 employees, and three industries, accounting, financial services, and management consulting. A handful of fintech companies slipped through the industry filter, so I added keywords to include on the accounting side and keywords to exclude on the fintech side. Writing those search criteria used to be a full-time job on its own. Now you hand the industry to ChatGPT, ask for Apollo keywords, and get a workable set in a minute.
Then move to the people tab, because the point of sizing a market is knowing how many humans you can contact, not how many companies exist. Add the job titles your offer actually speaks to. Titles are the lever here. If the market comes back small, reach further down the org chart to pull the number up. If it comes back large, concentrate on decision makers (founder, co-founder, managing partner, partner) and drop the lower-level managerial roles entirely.
With every filter in, that search returned around 41,000 leads, which is a solid place to be. When we take on a client we want to see 30,000 to 50,000 as the minimum. More is a bonus. Less is not automatically a no, but it forces a strictly account-based approach, and account-based is a different job with different economics. Which Apollo fields are load-bearing and which are marketing fluff is the whole subject of the Apollo guide.
Divide by three and you have your monthly volume
The formula is one line. Take the market and divide it by 3 months. So 41,000 becomes roughly 14,000 prospects a month. Run a two-step sequence, which is what we do, and that is roughly 28,000 emails a month. Now you know exactly how much sending infrastructure to build, derived from your own market instead of a package tier someone sold you. The domain and inbox math that turns that figure into a bill is in the cheap infrastructure breakdown, and the daily send ceiling per inbox is in the volume post.
Why three months and not one? Because you can reach every contact once a quarter without becoming the person who emails them constantly. By the 90 day mark, someone who never replied has no memory that you wrote at all. The slot is free again.
The number also tells you when email alone is the wrong plan. Under 10,000 prospects a month, there is not enough volume to find the winning offer by sending. Two options: widen the job titles to grow the market, or add touch points so each contact is worth more, which means LinkedIn and phone alongside the email.
That was how many you CAN reach, not how many you should
Here is where most people go straight to pulling emails and phone numbers, and it is why they spend their entire data budget on contact information for companies they were never going to close. Go back to the company tab and export company records only. Apollo caps a selection at 10,000, so segment to get around it: employee size split this market into clean exports of 10,000, 10,000, and 6,200.
That CSV goes into Clay, which pulls data from a lot of providers into one table. No emails yet. No mobile numbers yet. Company rows and nothing else, because the next step decides which of these companies deserve the expensive part.
Pre-qualify the company before you pay for the person
Three checks ran against all 10,000 rows:
- How many locations the firm has.
- Whether private equity owns them, and if so, which firm.
- Whether they publish case studies on their website.
The third one matters to our process specifically, so choose your own third signal, the thing that separates a client you close from a company that merely fits the filters. While that column runs, take the bonus: for every company that does have case studies, write the personalization line in the same pass, so the copy is finished before the first email address is bought.
Score every company 0 to 3, then spend accordingly
Three signals, three yes-or-no answers, one score per company. Zero means none of the qualities. Three means all of them. That single digit now decides how much money each prospect is allowed to cost you.
| Score | What it means | What you buy for them |
|---|---|---|
| 3 | All three signals | Everything: extra emails, mobile numbers, catch-all verification on top, and manual follow-up from the setter and the caller. |
| 2 | Two of the three | Mobile enrichment plus the verified email, and a place near the front of the send order. |
| 1 | One of the three | Verified email only, straight into the standard two-step sequence. |
| 0 | None of them | The cheapest path there is, or cut them and move that money onto the 2s and 3s. |
None of the three checks costs meaningful money. Everything downstream of them does. That is the entire argument for doing the qualification on the company side, before a single contact record gets enriched.
Buy the contact data last
Now feed the qualified domains back into Apollo with the job titles attached. That returned a list of 6,500 people here, a sample rather than the full market. Run the Apollo scraper on Apify, which costs around $1.20 per 1,000 leads, and it produced roughly 3,400 leads before I stopped it.
Pull every prospect regardless of whether Apollo already had an email, because Clay can fill the gaps later and cheaper. Filter the table for rows with no email, run them through an email waterfall across several providers, and you recover a good share of them, close to 50% in this case. Then filter for the 2s and 3s only and run those through a mobile waterfall. Out of 3,400 leads, only around 317 scored a 2 or a 3, so the phone budget stayed small and every number in it was worth calling.
Last step before anything is loaded into the sequencer: verification. Merge the Apollo emails and the Clay emails into one column and run the lot through a verifier like MillionVerifier to split them into okay, catch-all, and invalid. Send to the okay addresses first. Catch-all verification costs more, so run it only for the high scores, which is the same principle as the phone numbers. Spend where the score says the return is.
The market picks the strategy. Everything after it is execution.
The same number decides your personalization
Above roughly 30,000 contacts, use one or two tasteful personalizations and let the offer carry the email. Below that, the copy has to work considerably harder, because you get fewer swings and every one of them counts. This market was large, and the offer was strong, so it got exactly one touch: a line reading "prospects like {{case_study}}", pulled from the case study column written back in the qualification pass. Companies with no case study got a fallback built from their service, along the lines of "prospects looking for digital forensic services."
That is the honest ceiling on personalization. It exists to show you did your homework, not to carry a weak offer. List precision plus a real offer beats hyper-personalization every time, and the way you prove which segment of your market actually converts is by testing it, which is the segmentation and testing process.
tl;dr: count contacts not companies, divide the market by 3 to get your monthly volume, qualify companies before you buy a single email, and score every one of them so the expensive data only touches the prospects who earned it.
FAQ
How do I calculate my total addressable market for cold email?
Build your company filters in Apollo first (geography, employee size, industry, plus keywords to include and exclude), then switch to the people tab and add the job titles your offer speaks to. The number you want is contacts, not companies. In the walkthrough on this page, accountants in the US and Canada at 5 to 100 employees returned around 41,000 leads.
How many cold emails should I send per month?
Take your market size and divide it by 3, because you can reach every contact roughly once a quarter without wearing them out. A 41,000 person market means about 14,000 prospects a month. Run a two-step sequence and that is roughly 28,000 emails a month, which is the exact size of the sending infrastructure you need to build.
What is a big enough market to run cold email?
We want to see 30,000 to 50,000 contacts as the floor before starting a client, and anything above that is a bonus. Under 10,000 reachable prospects a month, email alone will not carry the campaign. Widen the titles to pull the number up, or add LinkedIn and phone so each individual contact is worth more effort.
Should I pre-qualify companies before buying contact data?
Yes, and it is the step most people skip. Export company records from Apollo and qualify them in Clay before you pay for a single email or mobile number. Checks like location count, private equity ownership, and whether the site publishes case studies cost almost nothing to run and decide where the expensive data budget actually goes.
How does lead scoring work for outbound?
Pick three qualifying signals, check every company against them, and score each one 0 to 3. Threes get everything: extra emails, mobile numbers, catch-all verification, manual follow-up. Zeros get the cheapest path or get cut. In the run on this page, only around 317 of 3,400 scraped leads scored a 2 or 3, so the phone budget stayed small.
If you would rather not spend weeks learning Apollo and Clay, that is what we do: we map your whole market, enrich it, and build the outbound infrastructure that can actually reach it on a consistent basis. 3,000+ sales calls booked for clients in the past year. Or keep learning free, every training grouped by topic sits on the trainings hub, and the full video for this one is on YouTube.
PS - the 21 calls in 21 days came from the low-volume approach. A small market is not a problem. Treating a small market like a large one is.