The question that catches you off guard
Someone asks you what your churn rate is, or what your customer lifetime value looks like, or how much revenue came from a specific channel last month. You open your spreadsheet and the honest answer is: it depends how you count it, and half the inputs are estimates anyway.
This happens constantly when you run a small operation. You don’t have a data team checking your definitions or a finance function reconciling every number before it leaves the building. You have a handful of tools, a spreadsheet, and a rough sense of what’s true. Most of the time that’s fine. The problem shows up when someone wants a number and you have to decide how to present something you’re not fully sure about.
The instinct is to round it off and say it with confidence, because a clean number sounds more credible than a hedged one. That instinct is usually wrong, and it’s worth understanding why.
Why a clean-sounding number can be worse than no number
A number with no visible uncertainty invites people to treat it as fact. They’ll build a decision on top of it, repeat it to someone else, or use it to compare against a target. If the number is actually a rough estimate dressed up as a measurement, the decision built on it inherits an error nobody can see.
This is different from lying. You’re not making the number up. You’re just presenting an estimate with the same confidence as a count. A count is something like “we had 40 paying customers on the first of the month.” An estimate is something like “our churn rate is probably around 5%, based on a definition of active customer that I set myself and haven’t fully audited.”
Both can be useful. But they are not the same kind of statement, and reporting them identically is where the trouble starts.
Say what you actually know
The first move is to separate the part you’re sure of from the part you’re not. Take churn as an example, since it’s one of the metrics that looks simple and rarely is.
Churn rate is usually calculated as customers lost divided by customers you started with, over some period. That’s straightforward arithmetic once you have the two counts. Where the uncertainty creeps in is upstream of the arithmetic: what counts as a customer who “started” the period, what counts as “lost,” and whether someone who downgraded but didn’t cancel counts at all.
If you’re not fully sure about your churn number, it’s rarely because you can’t divide. It’s because your definition of an active customer has edge cases you’ve been handling inconsistently, or because a chunk of your cancellations happen through a payment processor and you’re inferring them from failed charges rather than an explicit cancel action. Say that part out loud. “Our churn looks like it’s in the 4 to 6% range this month, but that includes some customers I’m counting as churned based on a failed card charge, and a few of those probably just need to update their payment method.”
That sentence is more useful to the person asking than a bare “5%,” because it tells them where the number is solid and where it isn’t.
Give a range instead of a point estimate
When you’re not confident in the precision of a number, a range communicates that honestly without hiding the number entirely. Instead of “MRR grew 8% last month,” you can say “MRR grew somewhere between 5 and 10%, depending on how I treat two annual plans that renewed mid-month and get spread unevenly across periods.”
Monthly recurring revenue looks like it should be an exact figure, since it’s just subscription amounts added up. In practice it gets messy around proration, annual plans converted to a monthly-equivalent number, discounts that apply for a limited number of cycles, and trial conversions that land right at the edge of a reporting period. None of that makes MRR useless. It just means the number is a modeled figure, not a bank balance, and treating it as a modeled figure means giving it the honesty of a range when your inputs are shaky.
A range costs you almost nothing to state and it does real work. It tells the reader the size of your uncertainty, not just its existence.
Name the source and its known gaps
A number is only as trustworthy as where it came from, and in a one-person operation the sources are often a patchwork: one tool for ad spend, another for site analytics, a spreadsheet for manual entries, a payment processor’s dashboard for revenue. When you report a number, naming the source takes ten seconds and changes how much weight the reader should put on it.
This matters most with attribution. If you’re reporting that a channel drove a certain share of signups, the honest version of that claim includes how you attributed it. Last-click attribution gives credit to whatever touchpoint happened right before conversion, which systematically undercounts channels that build awareness early and overcounts whatever’s closest to the signup button. If someone found you through a video, then later clicked a search ad and signed up, last-click attribution hands the whole credit to search. That’s not a flaw you can fix by trying harder. It’s a structural property of the method.
So instead of “video drove 12% of signups,” a more honest version is “last-click attribution credits video with about 12% of signups, but that’s almost certainly an undercount, since video tends to be the thing that starts someone’s interest rather than the last click before they convert.” You’re not being falsely modest. You’re describing what the measurement can and can’t see.
It’s fine to say you don’t know yet
Sometimes the correct answer to “what’s your conversion rate” or “what’s your CLV” is that you don’t have enough data yet to say. Customer lifetime value in particular gets reported far too early in a lot of small businesses, because it’s calculated as a projection of future revenue based on a handful of customers who’ve only been around a short time. If your oldest customer has been with you for two months, any CLV number you report is a guess about behavior you haven’t actually observed.
You can still be useful here. “I don’t have a reliable CLV yet because our earliest customers have only been around eight weeks, and I don’t want to project a lifetime value off two months of data” is a real answer. It tells the person what’s missing and why you’re not filling the gap with a guess dressed up as an estimate.
Update the number when you get better data
Reporting uncertainty once isn’t the end of the job. If you told someone churn was in the 4 to 6% range because some of it was inferred from failed charges, and you later go through and manually confirm which of those were real cancellations versus payment issues, go back and correct the number. Treating an estimate as a draft rather than a final answer is part of what makes the honest version worth doing. It also builds the habit of tightening your definitions over time instead of reporting the same soft number indefinitely.
A rough example
Say you’re looking at your dashboard and see 40 active customers at the start of the month, with 3 marked as canceled. That gives you a churn rate of 7.5% by the plain arithmetic. But maybe one of those three cancellations was actually a duplicate account you created for testing, and another had a failed payment that hasn’t been retried yet. Reported honestly, that’s not “7.5% churn.” It’s “somewhere around 5 to 7.5%, with one of the three flagged cancellations needing a closer look.” That’s a number you can act on without overstating what you actually know.
None of this is about being vague for the sake of it. It’s about matching the confidence in how you state a number to the confidence you actually have in it, so the people relying on that number, including future you, aren’t building on more certainty than exists.
If you want more on how specific metrics get calculated and where they tend to mislead, you can find the rest of our explainers on the home page.
Get new guides and videos first — join the Telegram channel.