Everyone bought the same lead list

A viral thread says to swap Apollo for six public databases: the instinct is right, two of the six are already crowded, and the part that actually works is the part it spends the least time on

The diagnosis is better than the fix

A thread went around recently arguing that if you are running outbound off Apollo you are effectively pre-deleted, and that the fix is six public databases nobody scrapes: SEC EDGAR, USAspending, LinkedIn job posts, G2 reviewers, Crunchbase funding rounds, and BuiltWith.

The opening is the strong part. A list product sold to thousands of teams, filtered on the same job titles against the same ideal customer profile, produces the same inboxes. You are not losing on the offer at that point, you are losing on position in a queue you did not know you had joined.

The supporting numbers are worth less than the mechanism. "Eighty cold emails a day" and "24 calls in the first week" have no source attached and should be read as illustration. The argument does not need them: if the filter you typed was available to everyone who bought the same seat, it has been typed before.

A signal is only worth what it costs to find

Here is what the thread does not say about its own list. The value of a public source is not that it is public. It is that reading it takes work most senders will not do. That work is the whole advantage, and it is rented rather than owned, because it disappears as soon as enough people read the thread.

Crunchbase is the proof sitting inside the list. "A company that raised $3.2m last week has $3.2m to deploy, and zero cold emails about it" is the one claim any founder who has closed a round will laugh at. A funding announcement is the most heavily worked trigger in B2B outbound; there are products whose entire job is to email you the morning after the news breaks. It appears on the list because it is famous, and it is famous because it is already farmed.

So sort the sources by how much work they cost to use, not by how obscure they sound.

Which of the six hold up

Federal spending is the best one. USAspending publishes every federal award with amounts and dates, through a free public API, and it is genuinely underused, mostly because most companies cannot sell to a contractor who just won a defence award. If you can, you get something almost nothing else gives you: budget that is confirmed and dated rather than inferred.

SEC filings are useful with the framing corrected. Public companies are not required to disclose their vendors. They file material contracts as exhibits and describe priorities and worries in the annual report, and EDGAR full text search covers filings back to 2001 for free. What you are reading is strategy and risk, not a supplier list. A risk factor is a company writing down, under legal pressure to be honest, what it is afraid of. That is a fine thing to have read before you write.

Hiring is the right signal from the wrong source. LinkedIn fights scraping, and its postings duplicate and go stale. The same jobs sit in applicant tracking systems as documented public JSON: Greenhouse's job board API, Lever's postings endpoint, Ashby's equivalent. No blocks, one request per company. And the spec beats the job title by a mile. A title tells you a role exists; several paragraphs written by the person who will own the problem tells you what is broken and how they think about fixing it.

Tech stack is real, and "a tech stack is a public confession of priorities" is the best line in the thread. Just budget for it. BuiltWith at any useful volume is a paid product, and running stack detection against a domain list you already have is the cheap version of the same idea.

Review sites are the one to change rather than adopt. Harvesting reviewer identities from G2 or Capterra and pairing them with found email addresses runs against those sites' terms, and in Europe it puts you the wrong side of GDPR, since a review is personal data and "they posted it publicly" is not on its own a lawful basis for emailing them. The reviews are still the most valuable thing on the list. They are your competitor's customers writing down which parts hurt, in the vocabulary they use for it. Take the language, not the people.

The part worth stealing

Step three of the thread is the whole thread: "saw you're hiring four SDRs, pipeline must be priority number one." The signal is not really a targeting filter. It is evidence you looked. Every database question is downstream of that, which is why any source producing a sentence true of exactly one company will work, and why no source producing only a job title will, however obscure it is.

That gives you a test that costs nothing. Could the first line of this email be pasted into an email to any other company on the list? If it could, the source did not help you, and you have bought a slower way to be number 71.

If you sell to founders, none of the six apply

Worth saying plainly, because most people reading this are not selling to federal contractors. A solo developer selling a $19 a month tool has no filings to read, no contract awards, and no applicant tracking system to query. At that price you cannot fund one to one selling anyway, which is the same arithmetic that separates agency work from software.

The equivalent public record for a small product is not firmographic. It is people describing the problem in public: the subreddit thread, the GitHub issue filed against the tool they currently use and dislike, the forum question, the "how does everyone handle this" post. Apollo will never carry it, because it is not a company attribute, it is a complaint with a timestamp. It is also the only version of this where a reply is welcome rather than tolerated, because answering a question is not the same act as pitching a stranger. We wrote about where those conversations actually happen.

Running it without becoming the thing you are avoiding

Two sources at most, pulled weekly, in numbers small enough that each message can cite something you personally read. The failure mode is automating the message as well as the collection, which puts you straight back in the queue holding a fresher list.

The thread closes by saying your competitors will keep fighting over the same inboxes. Probably true, and the six databases join that fight the moment the thread gets enough views. What does not get crowded is the willingness to read one company's public record and write one email about it, and the reason it stays uncrowded is that it does not scale. OctoLoops is built around that shape: it does the research and drafts the outreach on a weekly loop, and you keep the judgment about what is worth sending.