How Many Accredited Investor Verifications Does a Round Actually Need?

A sponsor emails: we are raising eight million at a fifty thousand dollar minimum, how many verifications do we buy?

The instinct is to divide. Eight million over fifty thousand is a hundred and sixty investors, so a hundred and sixty verifications. It is a clean calculation and it is wrong in both halves. The minimum is not the average check, so the investor count is too high. And the investor count is not the verification count, so whatever number you land on is too low.

Those two errors do not cancel. They are independent, they pull in opposite directions, and the only way to get a usable number is to do them separately. This is how to do that — including where to find a real benchmark for your market, which turns out to be free, public, and almost entirely unused.

The arithmetic most sponsors start with

Target raise divided by minimum check. It is the first thing everyone reaches for, and it fails twice.

The minimum is a floor, not a forecast. You set a minimum for administrative reasons — below some number the paperwork costs more than the money is worth. It tells you what you will accept, not what people will write. Actual average checks sit above the minimum, often well above, because the investors who take the whole allocation seriously are not the ones scraping in at the bottom. Size on the minimum and you will plan for far more investors than you get.

An investor is not a verification. One subscription can be four verifications, or two people can be one. Some of the people you verify never fund. On a long raise some of them need verifying twice. The relationship between the two numbers is a multiplier specific to your investor mix, and for most syndications it runs well above one.

So the sizing problem is really two problems. First, how many investors will this raise actually have? Second, how many verifications does that number of investors require? Sections three and four take them in that order.

The number nobody publishes

Search for the average number of investors in a private placement and you will find figures. Before you use one, ask where it came from, because the obvious source does not have it.

The SEC publishes a substantial Regulation D statistics series. It covers the number of offerings and initial filings, amended filings, total filings, the aggregate amount sold, the mean amount sold and the median amount sold — all of it broken out by claimed exemption, by fund versus non-fund issuer, by issuer type, and by state, running from 2009 to the current quarter. It is a genuinely good dataset.

Number of investors per offering is not one of the variables. Not in the statistics series, not in the SEC's guide to the dataset, and not in the DERA market-statistics report on exempt offerings published in March 2025, which gives filing counts and aggregate capital raised for 2009 through 2024 without disaggregating investor counts at all.

That matters more than it sounds. The organization that receives every one of these filings, that has the investor count on each of them, and that already publishes eight other statistics derived from the same forms, does not publish this one. So when a vendor page or an industry blog tells you the typical syndication has some specific number of investors, there is no aggregate public dataset standing behind it. It may be a reasonable guess from somebody's own deal flow. It is not a market figure.

We are not going to offer you a substitute average, for the same reason. What we can do is better than an average anyway, because you do not actually want the industry's number. You want the number for deals that look like yours.

Where the number actually lives

Every issuer relying on Regulation D files a Form D, and Form D asks for exactly the things you need.

• Item 6, Federal Exemptions and Exclusions Claimed. Rule 506(b) and Rule 506(c) are separate checkboxes, so you can tell a generally-solicited deal from a private one at the filing level even though the published statistics do not split them.
• Item 11, Minimum Investment. "The minimum dollar amount of investment that will be accepted from any outside investor."
• Item 14, Investors. The one that matters: "Regardless of whether securities in the offering have been or may be sold to persons who do not qualify as accredited investors, enter the total number of investors who already have invested."

Two ways to get at it. One deal at a time, look the sponsor up on EDGAR by company name and read their Form D filings. In bulk, the SEC publishes the Form D Data Sets — structured quarterly extracts covering September 2009 through June 2026, as of the 2026 Q2 posting. Six files; the one you want is OFFERING, which carries these fields:

Divide the first by the second and you have a comparable sponsor's realized average check — not their minimum, not their pitch deck, what actually landed. Do it for ten sponsors raising deals the size and shape of yours and you have a benchmark for your market that, judging by how rarely anyone mentions this data exists, none of your competitors has built.

From investor count to verification count

Now the second calculation. Four things move the verification count away from the investor count, and only one of them moves it down.

Entities are counted by owner, not by subscription. This is the big one in real estate and the one that surprises people. A special purpose vehicle formed so four friends can take one allocation cannot use the balance-sheet route to accreditation. An SPV is almost always an LLC, and Rule 501(a)(3) accredits a limited liability company "not formed for the specific purpose of acquiring the securities offered, with total assets in excess of $5,000,000." A deal SPV fails the first clause by definition, and usually the second as well. So the SPV is accredited only if all of its equity owners are. That is four verifications arriving as one line on your cap table. We covered the full set of entity cases in the syndication runbook; for sizing purposes the rule of thumb is that your entity investors need counting by head, and you will not know how many heads until you ask.

Married couples are usually one, not two. The only multiplier that runs in your favour, and it is worth knowing because overbuying on this one is a common mistake. Rule 501(a)(6) accredits a natural person with individual income over $200,000 in each of the two most recent years, "or joint income with that person's spouse or spousal equivalent in excess of $300,000 in each of those years." Rule 501(a)(5) does the same for net worth: individual net worth "or joint net worth with that person's spouse or spousal equivalent" exceeding $1,000,000. A couple subscribing jointly is normally a single verification against the joint thresholds.

Some of the people you verify will never fund. If you verify at commitment, you pay for everyone who soft-circled and drifted. If you verify at close, you avoid that cost and compress the entire verification workload into the final weeks, against a close date set by a purchase agreement rather than by you. Most sponsors land somewhere in between and absorb some waste on purpose, because a scramble in the last ten days is more expensive than a few unused seats.

A long raise verifies some investors twice. Third-party confirmation under Rule 506(c)(2)(ii)(C) has to have happened "within the prior three months" of the sale. On a raise that runs five or six months, the investors who verified in week two are stale by the time they fund. That is not a legal trap, it is a calendar fact, and it means a fraction of your investors need a second letter. The longer the raise and the earlier the enthusiasm, the larger that fraction.

Put together, the shape is:

A worked example, with the assumptions on the table

Back to the eight million dollar raise at a fifty thousand dollar minimum. Every input below is an assumption, and they are ours for illustration, not benchmarks — the whole point of section three is that you should be replacing them with numbers pulled from comparable Form D filings and from your own last deal.

Two things are worth noticing. The first correction cut the count by more than a third, and the next three added two thirds back. A sponsor who did only the first correction would have bought a hundred seats and run out in month four. A sponsor who did none of it would have bought a hundred and sixty and been roughly right for entirely the wrong reasons, which works exactly once.

The second thing: the swing between the naive figure and the worked figure is not noise. It is the difference between the 100-seat conversation and the 200-seat conversation, and the inputs driving it are all knowable in advance.

The two numbers only you have

Comparable Form D filings give you the average check. The other two inputs are not in any dataset and never will be, because they are facts about your investor base.

Your entity mix. Pull the subscription documents from your last deal and count the subscribers that were not natural persons. For each one, count the equity owners. That ratio — verification subjects over subscriptions — is your entity factor, and for real estate sponsors it is routinely the largest single correction on this page. Twenty minutes with a closed deal's paperwork will give you a number you can use for years.

Your commitment-to-close rate. Same exercise, different column. How many people said yes, and how many wired? If you have tracked soft circles at all, you already have this. If you have not, start on this raise, because it is the input with the widest range across sponsors and the one where a borrowed number is most likely to be wrong.

We are deliberately not giving you industry defaults for either. There are no credible aggregate figures for them, for the same reason there is no published investor count, and inventing a plausible-looking percentage would make this page feel more useful while making your budget less accurate.

Buying seats: the asymmetry that decides it

Once you have a number, the purchasing decision is unusually lopsided, and it is worth spelling out why.

Buying short costs you a scramble. You run out mid-raise, at exactly the point where your attention is on the close, and the marginal verification is the most expensive one on the rate card. Buying long costs you float and nothing else, because seats do not expire and are not tied to a single deal — a pack that outlasts this raise carries into the next one.

The arithmetic between pack sizes makes the same point. A 25-seat pack is $1,725 and a 50-seat pack is $3,250. So the twenty-five seats between them cost $1,525, or $61 a seat — below every published pack rate, including the 50-seat rate itself. Topping up a 25-pack one verification at a time instead would run $2,475 for the same twenty-five seats. If your worked number is anywhere near forty, the larger pack is already the cheaper answer, and the seats you do not use this time are not waste, they are next deal's budget.

The exception is a sponsor doing one deal and then stopping. If this is genuinely your only raise, size tight, because the carry-forward argument does not apply to you.

Five ways the count goes wrong

Sizing on the minimum check. The single largest error, and it is an overestimate, which is why it survives — nobody investigates a budget that turned out to be too big.

Counting subscriptions instead of owners. The single largest underestimate. One line on the cap table can be four people who each need a letter.

Counting a married couple twice. The joint income and joint net worth tests exist precisely so you do not have to. Verifying spouses separately is work nobody needed.

Reading initial Form D filings instead of amendments. Do this across ten comparables and you will conclude that everybody in your asset class raises small rounds from three investors. They do not; you read the fifteen-day notice.

Buying seats after the first investor asks. Verification is on the critical path to a close date somebody else set. The sponsors who find this painless are the ones who worked out the number before the raise opened.

None of this is complicated arithmetic. It is just arithmetic that has to be done in the right order, on inputs most sponsors have never gone and collected — which is the whole reason the free public dataset in section three is worth an afternoon.