This is a running collection of thoughts, notes, and half-formed ideas about work, life, and the systems I use to keep both moving forward. Some posts are practical, some are reflective, and a few exist simply because they wouldn’t fit neatly into a spreadsheet. If you’re looking for the bigger picture, you can head back to the main page.
I've been reporting Monthly Recurring Revenue for years now. I've built dashboards, explained variance to the board, and argued with engineering about data quality. And yet, at least once a week, I find myself staring at a spreadsheet asking: "Wait, is this actually MRR?"Here's what I thought I knew:MRR = Monthly Recurring Revenue
The predictable, recurring revenue you can count on next month. Subscription revenue. The stuff that shows up automatically.Simple, right?Then reality happens.
The Edge Cases That Keep Me Up At NightA customer pays $100/mo, churns mid-month after we charge them. Is their MRR $100 or $0?We collected the cash. It's on the P&L. But they're gone. We won't see them next month. Do we prioritize "what actually happened" (active, $100) or "current state of the business" (churned, $0)?After weeks of back and forth, we went with $0. Ending MRR should reflect reality going forward, not lag behind what we've already been paid for. But I'm still not 100% sure that's right.A customer has a frozen account. Their $200/mo subscription is paused but not canceled. What's their MRR?They're not paying us. But they might come back. Is that $0 MRR (we won't collect next month) or $200 MRR (they're still technically a customer)?We settled on $0 for frozen accounts. If we're not collecting, it's not recurring. But then someone asks "what about our reactivation rate?" and suddenly frozen accounts need their own category.We give a customer 2 months free ($200 value) as a promo. Their plan costs $100/mo. Is their MRR $100 or $0?I think it's $100. The credit is a cost of acquisition, not a reduction in their commitment. But we've had this argument at least three times and I still get questions about why MRR doesn't match what we're actually collecting.A customer gets charged $100 but the payment fails. Are they still active MRR?For how long? One failed payment? Two? When does "payment failure" become "churned"?We track them as active MRR until we stop attempting to collect, but the exact definition of "stopped attempting" is fuzzy.
What MRR Is Actually Trying To MeasureAfter all these edge cases, here's what I keep coming back to:MRR is trying to answer: "What does our recurring revenue look like going into next month?"Not what we collected this month. Not what we recognized on the books. What we can expect next month based on our current customer base.That's why:
• Churned customers show $0 MRR (even if we collected from them this month)
• Frozen accounts show $0 MRR (even if they might reactivate)
• Promotional credits don't reduce MRR (they're a cost, not a revenue reduction)
• Failed payments stay in MRR briefly (we're still trying to collect)We're measuring the forward-looking state of the business, not the backward-looking cash that moved.
Why This Isn't An Accounting MetricHere's the thing that confuses everyone unfamiliar with the metric:MRR doesn't exist in GAAP.It's not on your P&L. Your accountant doesn't calculate it. The IRS doesn't care about it.MRR is an operational metric, a business health indicator. It's completely separate from:
• Cash (what we collected)
• Revenue (what we recognized under GAAP rules)
• Billings (what we invoiced)Example: When All Three Numbers Are DifferentCustomer on $50/mo plan upgrades to $200/mo on March 15thCash collected in March: $50
(They already paid for March at the old rate on March 1st. The $200 charge hits April 1st)
GAAP Revenue recognized in March: $125
(Pro-rated: $50 for the first 14 days, $200 for the last 16 days, with some accounting magic that makes it roughly $125)
Ending MRR in March: $200
(Their current plan is $200/mo. That's what we expect in April)All three numbers are correct. They're just measuring different things.
The Walk (And Why It Matters)The MRR walk shows how MRR changed:Starting MRR: $100K
(+) New MRR: +$10K (new customers)
(+) Expansion: +$3K (upgrades)
(-) Contraction: -$2K (downgrades)
(-) Churn: -$5K (cancellations)
= Ending MRR: $106KThis tells a story:
• We're growing (+6% net)
• New customer acquisition is strong (+10K)
• But we're losing almost as much to churn (-5K)
• We should probably focus on retentionYou can't get this story from the P&L. Revenue recognition smooths everything out. Cash is lumpy. MRR shows you the underlying subscription dynamics.
The Questions Engineering Always Asks"Why doesn't MRR match revenue on the P&L?"Because MRR isn't revenue. It's a forward-looking health metric. A customer who paid annually in January shows $100 MRR every month, but we only collected cash once."Why does this customer show $0 MRR but we're still recognizing revenue from them?"Annual customer who churned. We collected $1,200 upfront, we'll recognize it over 12 months, but their MRR went to $0 the moment they canceled. We won't see them next year."Why don't we include usage in MRR?"Because usage isn't recurring. It's consumption. A customer might use $5K of compute this month and $2K next month. That's not "contraction," they just used the product less. Usage gets its own line."So how do I know if we're actually making money?"
Look at the P&L. MRR tells you if the subscription business is healthy. Revenue tells you if we're making money. Different questions.
Where I'm Still Figuring Things OutApp errors that create credits
Sometimes our app messes up and we issue a credit. Does that reduce MRR? I don't think so (it's an expense/refund, not a pricing change), but it definitely affects what we collect. Should we track it separately? Probably. Do we? Not consistently.Customers who pay but don't use the product
We have customers paying $50/mo who haven't logged in for 6 months. They're active MRR. But are they? They're going to churn eventually when they notice the charge. Is that $50 "real" recurring revenue or are we just borrowing it?I don't have a good answer. For now, paying = active MRR. But it feels optimistic.The difference between "paused" and "churned"
Some customers pause their accounts with intent to return. Some just ghost us. Both stop paying. Where's the line? We track paused accounts as $0 MRR, but should they be in a separate "hibernating" category for reactivation tracking?Still working on this one.
What I've Learned After Years Of ThisMRR isn't a number. It's a philosophy about how you want to measure your business.Do you want to prioritize:
• Cash accuracy? Then mid-month churns stay active for the month they paid.
• Forward-looking truth? Then churned is churned, even if you collected their last payment.Do you want to prioritize:
• Current commitments? Then frozen accounts are $0 MRR.
• Potential reactivation? Then frozen accounts get their own category.There's no GAAP for MRR because it's not a GAAP metric. Every company makes these calls slightly differently.The only wrong answer is being inconsistent or not knowing which philosophy you've chosen.
My Working Definition (For Now)
After all this, here's what I tell people:MRR = Forward-looking subscription commitment revenue at period-endThat means:
• Churned customers show $0 (even if they paid this month)
• Frozen accounts show $0 (not collecting = not recurring)
• Promotional credits don't reduce MRR (cost of acquisition, not pricing)
• Failed payments stay in MRR until we stop trying (brief grace period)
• Usage revenue stays separate (consumption, not commitment)But I'm prepared to revisit this definition next month when someone finds a new edge case I haven't thought of.Because that's the thing about MRR. You never stop defining it.
I spent way too long designing my daily wellness tracker for 2026. AI bots suggested 1-5 for everything. Here's why I didn't listen to them:Energy/Wellbeing: 1-5
This needs nuance. There's a meaningful difference between "okay" and "good" that a 4-point scale misses.Diet Adherence: 0-4
Having a zero feels psychologically accurate. When you completely fall off track, calling it a "1" feels like grade inflation.Movement: 0-3
This one's simpler. You either didn't move, took a walk, had a nice workout, or pushed yourself to the max. Five options would be overkill.The insight isn't the specific numbers; it's that different metrics deserve different scales based on how granular they actually are in practice. Movement is relatively binary. Energy exists on a spectrum.Plus, having different scales makes each metric feel more distinct when you're tracking daily. It's a small thing, but it keeps the whole system from feeling like you're just filling out the same survey question three times.Now if I could just remember to actually track it every day.