8:42 AM - Inbox triage
I open the Restaurant AI dashboard and there are already three things waiting. The Slack alert came in at 6:47 AM: "Sign-up completion spike - 12 new trials initiated overnight." That's good. I scan the email backend and pull up the CSV of who signed up. Three came through the partnership with Marco's Food Service in Portland. Two more from a referral chain. The rest organic or from the previous week's local ads.
I flip over to Gmail. Sarah Chen at Chen's Kitchen had replied to the onboarding email I sent yesterday. Her message is short: "Setup took ten minutes. Can the system actually tell me if my pricing is off on the lunch special, or is it just another software thing." That's fair. I've heard some version of that ten times now. I pull up her Stripe account details and see she's on day four of her 60-day trial. I draft a response but I stop. This deserves a real conversation. I send her a calendar link for a 20-minute call Thursday morning.
The third thing is the one I'm dreading. David Okafor at Union Tap has requested a refund. He completed the trial, liked the menu analysis, but the scheduling feature kept suggesting he bring in an extra line cook on Friday nights. His actual traffic data showed he doesn't need it. He says it feels like the system is trying to sell him labor. I flag this in Linear with a question for the product team about how the labor forecasting algo is training on historical density versus actual covers. This is important. I don't refund him immediately. Instead, I Slack the support channel with: "David Okafor at Union Tap - edge case on scheduling recommendations. Says system is over-suggesting headcount. Can someone walk me through the logic here."
10:15 AM - Partner touch base
I have a standing call with Marco, the owner of Marco's Food Service. He's sent three of the new signups. He tells me one of them is moving fast: Jessica Reyes who runs three locations and is already thinking about a team license. He also mentions his next referral round will be in two weeks when he visits his restaurant owners for semi-annual inventory checks. I note this in the shared Notion document we use for deal tracking. It's not in the product. It's just a spreadsheet I maintain manually. That feels inefficient but I don't have a sales pipeline tool yet, so Notion it is.
Before we hang up, Marco asks if we can white-label the onboarding email. Right now his restaurants are getting emails signed "Restaurant AI team." He wants them to see "Marco's Food Service recommends..." I tell him yes and add it to the roadmap. It's a simple ask. Should take the dev team a day. But that's the kind of thing that makes partners stick around.
I check Slack again about David's scheduling issue. No one's responded yet. I leave it alone. We'll surface this in the next product meeting.
12:30 PM - Lunch and the metrics check
I grab a bagel and pull up the Stripe dashboard. Week-to-date: four conversions from trial to paid. Two at 99 dollars a month, two at 149 dollars a month. That's 496 dollars in MRR so far this week and it's only Tuesday. The velocity here is good. Month-to-date we're at 1,847 dollars MRR. That's on pace for the 15k monthly target I set for month two.
I open the dashboard metrics and look at active trials. There are 31 currently in-progress. Average days in trial is 18. Average conversion rate is about 26 percent. If I close 26 percent of those 31, I'm getting eight new customers this month. At an average of 124 dollars a month, that's roughly 992 dollars in MRR from the current active cohort.
I'm not frustrated by the number but I'm not satisfied either. The conversion rate should be higher. I walk back through the onboarding emails to see if there's a messaging problem. I notice the second email in the sequence is generic - it's a "here's how to access advanced features" message with no mention of what those features actually do for a restaurant owner. I rewrite it to lead with a specific win: "Here's how Jessica Reyes cut her weekly scheduling time from 90 minutes to 25." It's real. It's from week one. I send the updated version to all active trial users who are on day three or four of their journey. That's 18 people.
2:08 PM - A customer thank-you and a billing snafu
Melissa Huang at Huang's Dynasty replies to an email I sent two weeks ago. She's now 60 days in, converted to a paid monthly plan at 129 dollars a month, and she's writing to say the menu engineering feature saved her 800 dollars a week by rationalizing her appetizer lineup. She's asking if there's a team plan for her wife who manages a second location. I feel genuinely good about this. This is why I'm doing this. I immediately forward her to the annual plan, two locations at 1,200 dollars a year instead of 129 a month each, and she says yes. That's 1,200 dollars in annual value, and I add her as a case study candidate for next month's outreach.
Then David, the Okafor refund issue, responds on Slack. The product team walked through the labor forecasting with him and it turns out his historical data was incomplete. The system had only seen four weeks of his point-of-sale data before making recommendations. They extended the training window to eight weeks and re-ran it. His scheduling suggestions are now more reasonable. He tells me he's reconsidering the refund. I don't push. I just say I'm glad we could dig into it and that I'm here if he has more questions. He stays on the account.
But then a payment processing error. Another customer, Tom Martinez at Martinez Bros Grill, his card was declined this morning for his monthly renewal. It's sitting in Stripe as "failed payment." Tom hasn't noticed yet but these things expire if I don't catch them. I manually issue him an email with a payment link, re-run the charge through Stripe, and it clears. He gets a courtesy message saying "no action needed, just confirming your payment went through." These are the small moments that keep customers from churning. They're not automated. Someone has to notice.
4:30 PM - Pipeline and reflection
I block out the rest of my afternoon to think about the real constraint. Right now I'm operating this alone. I'm sending emails, managing Slack, responding to support issues, handling billing problems, and tracking pipeline in a spreadsheet. The product itself is working. Customers are seeing value. But the human work is becoming the bottleneck.
I review the pipeline in Notion one more time. 31 active trials, 8 to 10 likely to close by end of month, 4 to 5 on the fence. Jessica Reyes, the three-location owner from Marco, is flagged as "hot." She's 18 days in and has already watched two product tutorials. Her arc is different from the others. She's not cautious. She's sold.
I also notice we've had two churn signals this month. David was one. The other was a small location that upgraded to premium features, didn't use them, and requested a downgrade. That's actually fine data. It tells me the upsell messaging isn't right yet. I add a note: "Premium tier messaging - emphasize team collaboration and reporting, not feature count."
I pull up the activity log from the product itself to see what's actually moving the needle. It's users who activate the menu analysis in the first week and run a report. That's the leading indicator. Every user who doesn't do that by day seven has converted at less than 15 percent. Every user who does runs at 45 percent or higher. This is real. I flag it for Marco and for my next product sync.
6:15 PM - Closing down
I close the laptop around 6:20 PM. Today I reviewed 12 new trial signups, sent three manual customer emails, handled one refund inquiry without losing the customer, closed a team-upgrade opportunity, fixed a payment processing error, identified a leading conversion indicator, and rewrote an onboarding email for 18 users.
The product works. Customers see value. But I'm running this manually. The next month I need to hire someone to take over customer success, or I'll hit a ceiling at 100 to 150 active customers. Right now I have 47. At this growth rate I'll be there in five months.
What worked: treating refund requests as root cause analyses, not churn. Following up on thank-you notes with upsells. Having partners who send qualified leads. Having a product that actually delivers on the promise.
What's fragile: the pipeline is in Notion, not in a real CRM. Billing problems are caught by luck, not by alerts. I'm the only human reading Slack. If I miss something, no one catches it.
I think about Jessica Reyes again. She's probably the 300 to 400 dollar MRR customer that justifies the next three months of work. I make a mental note to call her on Friday, not Wednesday. Let her finish the trial. Let her feel the value sink in.
Then I close the laptop.