The Bureau of Labor Statistics keeps a table that follows every private sector establishment that opened in the United States in a given year, and then checks back once a year, forever. It is called Table 7, and the 1994 cohort in it is now 31 years old. Most people who quote this data quote one column. The useful column is the one next to it.
The famous column is headed “Survival Rates Since Birth.” Read it down and you get the number everyone repeats at dinner parties. Of the 569,387 establishments that opened in the year ended March 1994, 79.6 percent were still open a year later, 45.2 percent were open after six years, and 12.6 percent were still open in March 2025. That is the doom column, and it is the one that makes people stop reading.
The column beside it is headed “Survival Rates of Previous Year’s Survivors,” and it asks a different question: given that you made it this far, what are your odds of making it one more year? Read that column down the 1994 cohort and it goes 79.6, then 85.6, then 89.0, 89.6, 91.4. By the twelfth year it has reached 95.5 percent and it simply stays there. It dipped to 93.7 percent in the year ended March 2009 and to 94.9 percent in the year ended March 2021, which is to say it wobbled slightly during a financial crisis and a pandemic, and it never came back down to anything resembling the first year. The most recent reading, for the year ended March 2025, is 95.7 percent.
The first year is not a hazing ritual, it is the whole problem
The year-one figure has barely moved in three decades. The 1994 cohort lost 20.4 percent in its first year. The cohort that opened in the year ended March 2024 was 988,310 establishments strong and lost 22.1 percent in its first year. Thirty years of new technology, new channels, and new advice, and the first-year number is within a few points of where it started.
What changes is everything after. In its first year, the 1994 cohort closed at 20.4 percent. In its tenth, the survivors closed at 5.9 percent. Nothing in the table explains why the risk falls, so what follows is interpretation rather than measurement, and it is worth saying so before saying it.
Here is the reading that survives contact with how small businesses actually work. In year one you have to find every single customer. There is no list, no reputation, nobody walking in because their neighbor mentioned you. Every dollar of revenue is the result of an act of acquisition you personally performed that month. By year ten some proportion of your revenue arrives without you doing anything that month, because a customer came back, or told someone, or because your name is the one that comes up. The curve is not a picture of businesses getting better at their craft. It is a picture of customer acquisition slowly ceasing to be a monthly emergency.
One more line in the same data set points the same direction, though it needs a caveat. BLS also tracks the average employment of the survivors. The 1994 cohort averaged 7.2 employees per establishment at birth and 26.9 in March 2025. Part of that is survivor bias, since larger establishments survive at higher rates and the average rises as small ones drop out. But the direction is not in doubt, and it is worth saying plainly what the direction is: the establishments that made it through three decades did not get there by shedding people. They got there carrying more of them.
A pattern in the industry breakdown that is suggestive, not proof
BLS notes on its overview of establishment survival that health care and social assistance “consistently ranks among the industries with the highest survival rates over time, while construction ranks among the lowest.” That page was last modified in April 2016, so treat it as a long-run observation rather than this year’s news.
It is still worth sitting with. A medical or dental practice acquires a patient once and then sees that patient on a schedule for years. A general contractor finishes a bathroom and the relationship is, in revenue terms, over. Two businesses, two completely different answers to the question “where does next month’s work come from.” The industries are different in a hundred other ways too, so this is not evidence of cause. It is a coincidence pointed in a direction that matches the curve.
Which means the question is slightly wrong
“How do I get customers” is answered, almost everywhere, with a list of channels. Post here, advertise there, try this platform. But channels are the cheap part of the problem and they change every few years. The expensive part is that most owners cannot answer a much smaller question about the customers they already have: how did the last ten of them find me?
Without that answer, every channel decision is a coin flip dressed as a strategy, and worse, you cannot tell the difference between a channel that produced one good customer and a channel that produced one good customer who then sent two more. That second kind is the only kind that bends the curve. Here is how to start recording it, with four moves that cost nothing.
1. Add one question to the moment money changes hands
The question is “how did you hear about us,” and the point is not asking it, which most owners already do in passing. The point is writing the answer down somewhere you will look again. Put it on the intake or booking form as a short text field rather than a dropdown, because a dropdown only records the answers you already thought of. If you take work over the phone, add the question to whatever you already write down when you take a job, even if that is a paper notebook. If you invoice, add a line to the invoice template.
Do this today: pick the one place every paying customer passes through, and add the field there. One place, not four.
2. Claim the free listing that answers while you are asleep
A Google Business Profile is the thing that shows your hours, your phone number, and your reviews when somebody searches for what you do near where you are. Google states on its own help page that you can add or claim your business at no charge, and the process it describes is to go to business.google.com/add, enter your business details, then pick a verification method and complete it. Verification is a separate step that Google requires after you enter your details, it is the step people abandon, and an unverified profile is close to useless. Plan on finishing it in a second sitting rather than assuming you are done when the form submits.
Do this today: search for your own business name and town in a private browser window and see what comes back. If the listing is missing, wrong, or says “claim this business,” that is the job.
3. Ask the person who just paid you, by name
The highest-yield moment for a referral is the ten minutes after a customer is visibly happy with something you did, and it passes. The ask that works is specific rather than general: not “send people my way” but “who else do you know with this exact problem.” A name is a lead. A vague promise is a pleasantry.
If what you want is a public review rather than an introduction, the ask itself is regulated, and it is regulated more tightly than most owners realize. A sentence on Google and the same sentence on your homepage fall under two different definitions with two different sets of prohibitions, and three perfectly ordinary asks land on the rule. We worked through all of it in our piece on how the FTC rule changed the testimonial ask. Read it before you offer anyone a discount for a five-star review.
Do this today: write the one sentence you will say, and say it to the next satisfied customer without editing it in your head first.
4. Count at ten, not at a hundred
Ten answers to “how did you hear about us” is enough to see something. You are not looking for statistical significance, you are looking for a surprise, and the surprise is usually one of two things: a channel you spend real money on that produced nobody, or a source you were not counting at all that produced three.
Once you have the tally, two follow-on questions become answerable that were not answerable before. What did each of those customers cost you, counting your own hours and not only your ad spend, which we worked through in counting acquisition cost in hours rather than cash. And which single number should you watch every week, which is a smaller decision than it sounds and is covered in picking a growth metric you can already count. If the tally shows that people arrive but do not buy, the problem has moved upstream to what you say about yourself, and the six lists behind the value proposition canvas are the cheapest way to work on that.
Do this today: put a recurring reminder in your calendar for the day you expect to hit ten answers, and title it “read the answers.”
Where a prompt genuinely helps, and one thing to ignore on the page
None of the four moves above needs AI. What AI is good at here is the part that comes after the tally, when you have a messy handful of answers and want to think through what they imply without hiring anyone to think alongside you.
The Win Your First 10 Customers prompt on BusinessPrompter.com is built for that. It is one of the free ones rather than a Pro-locked entry in the library those prompts live in, and it is structured as a five-step chain rather than a single block of text, so you paste step one, read what comes back, then carry the useful parts into step two. Its first step asks the model to analyze the core principle, the problem it solves, who benefits most, the historical context, and the modern relevance, which is a reasonable way to make an AI argue with your assumptions instead of flattering them.
One honest warning about that page, because we checked it this week and it would confuse anyone who read it carefully. The prompt’s title is “Win Your First 10 Customers,” it sits in the Marketing and Sales category, and its summary describes “a systematic approach to customer acquisition that validates your product-market fit and builds momentum.” All three panels underneath, headed What This Prompt Does, Who It’s Best For, and What You’ll Have When You’re Done, describe something else entirely: a pricing prompt, about “establishing pricing structures,” about businesses “seeking to optimize product pricing,” producing “a detailed pricing strategy document outlining optimal price points.” The title, the summary, and the actual five-step chain are about customers. The three panels are about pricing. Trust the chain and the summary, and skip the panels.
The tools, and why the free tier is the correct plan rather than a trial
You do not need software to run the four moves. A notebook works, and for the first ten customers a notebook is genuinely better than software because nothing is faster than a pen. The moment a spreadsheet stops holding it, three real options exist, and their free tiers are sized almost exactly for a business at this stage.
HubSpot’s free tools are listed at $0 per month, free for up to two users and with no credit card required. Its paid Starter Customer Platform is currently shown at $7 per seat per month next to a listed $20 per seat per month, under a banner stating the special pricing is for new customers only and available for a limited time, so read that page carefully before assuming $7 is the standing rate. Zoho CRM publishes a free edition at US$0 with no credit card required and a stated limit of three users, with paid tiers at US$14, US$23, and US$40 per user per month. Capsule CRM lists a free plan at US$0 capped at two users and 250 contacts, with its Starter plan at US$18 per user per month billed annually, which raises the contact limit to 30,000.
Note what those free tiers actually contain. Capsule’s free plan holds 250 contacts. A business working on its first ten customers will not reach 250 contacts for a long time, and the honest advice is that the free tier here is not a trial you are supposed to grow out of quickly, it is the right plan for now. Upgrade when a specific thing you need is missing, not when the page suggests it.
Frequently Asked Questions
Does the BLS survival data cover sole proprietors and one-person businesses?
Business Employment Dynamics counts establishments with employees, so a business with no payroll is not in the cohort. The pattern it shows, that the odds of closing fall steadily with age, is about businesses that have at least one employee.
What if nobody can tell me how they found me?
That is a real and common answer, and it is data. If most people genuinely cannot say, the practical read is that you are being found through something diffuse like local reputation or a search result they do not remember clicking, which is an argument for making sure the Google Business Profile is claimed and correct rather than for buying ads.
Is a free CRM enough, or will I have to migrate later?
For the first ten to a few hundred customers a free tier is enough. Migration between these tools is a real cost, so if you already know you will need something specific, such as a shared team inbox, pick for that. If you do not know, pick the one you will actually open.
How long before the survival odds improve for my business?
On the BLS numbers the biggest single jump is between year one and year two, and the rate keeps climbing through roughly the first decade before flattening out at around 95 percent. There is no threshold you cross. It improves gradually, which is why the fourth move matters more than it looks: acquisition that leaves a record behind is the thing that compounds.
Start with the question. Not the channel, not the platform, not the ad budget. Ask every person who pays you how they found you, write it down where you will see it again, and read the answers when you have ten. That single habit is the difference between doing acquisition again next month and doing it on top of something.
