I’m a self-funded searcher in the Mountain West. Twenty-plus years in construction, buying through a ROBS, looking for a trades business I can own and run myself. Yesterday I terminated the LOI on the first deal I ever got that far on, after about a year of work on it. I’m writing this while it’s fresh because I learned more from this one than from the forty-some others I looked at and passed on, and most of what I learned I could have learned cheaper. The short version of the deal: a commercial subcontractor, $3M to $4M in revenue, twenty-ish employees, thirty-year owner, listed with a well-established local brokerage. Asking price was built on a broker recast of the seller’s preliminary year-end numbers. I first saw it in February, sent an initial LOI in April, got to a signed LOI in mid-September after five months of back and forth, and terminated five weeks later. In between: an SBA lender, M&A counsel, an ERISA attorney for a pension issue, a QoE firm, a CPA, and a lot of 2:45 a.m. wake-ups. Here’s what I’d tell myself a year ago. 1. If you’re more motivated than the seller, it’s the wrong deal. This is the big one and everything else is a footnote to it. From the first meeting, every piece of effort in this deal flowed one direction. I drafted five versions of the LOI; they marked up one. The seller was late to every call and left every meeting early. The broker said “by Wednesday” and delivered the following month. The initial diligence package, due ten business days after signing, took thirty days. My QoE firm sent a nine-item request list; three weeks later, zero of nine had arrived. A ten-minute payroll report I asked for in writing never came at all. I told myself a lot of stories about this. The seller is old school. The bookkeeper is on vacation. The broker is busy. All of it was probably true. None of it mattered. A seller who wants to sell shows up on time, sends the thing before you ask twice, and calls you back. If you find yourself carrying the deal for both sides, you are not in a negotiation. You’re in a chase, and you will never get a chase to a closing table. In hindsight it looked like a bad relationship, and it ended like one. 2. Anchor on filed tax returns, not the broker’s recast, and ask for the reviewed statements before you sign anything. The price was built on the broker’s recast of preliminary year-end financials from January. When the seller’s CPA produced reviewed statements in September, revenue was about $100K lower and net income was about $150K lower than the preliminary version. Every expense line was identical to the penny; only income had moved. The filed tax return tied to the lower number. The broker’s model, and the asking price, and five months of LOI negotiation, were all built on the version that didn’t survive. The seller’s side knew about the correction months before I did. Nobody mentioned it. I found it myself by laying the two P&Ls side by side the night the package arrived. If I’d had the filed return and the reviewed statements in hand before the first LOI, this deal is over in April instead of October. 3. Don’t spend a dollar on diligence until the documents are in your hands. This is my one actual regret. I engaged counsel, the QoE firm, and ERISA counsel on the strength of a signed LOI, before the seller had delivered the initial package. Reasonable on paper; the LOI said the package was due in ten business days. But the LOI says what the seller agreed to, not what they’ll do. I’m out something like $13K for work that was done against documents that never came. Most of my vendors have been gracious about rolling fees forward to the next one, which I’m grateful for and didn’t expect. But the right move was simpler: sign the LOI, wait for the package, and engage nobody until it’s complete. If the package is late, that’s your answer, and it’s free. 4. The lender’s math is the only math. My lender ran the reviewed statements under the current SBA change-of-ownership rules and the deal didn’t clear the historical DSCR test at the LOI price and structure. Not close. I’d been modeling this deal for months with my own spreadsheets and three different AI tools, and I was mostly right, but “mostly right” doesn’t fund. Get the lender’s actual formula early, with their add-back policy and their coverage threshold, and run the deal through it before you fall in love. If the bank says no, the price has to move, and if the seller won’t move, you’re done. Everything else is noise. 5. Watch what the structure does to your personal balance sheet, not just the business’s. By the time this deal was fully structured it had an SBA loan, an SBA line, a seller note, a lease from the seller, and the seller as a paid consultant for a year. Personal guarantees from both of us on all of it, and a deed of trust on our house. The seller would have been my lender, my landlord, and my employee simultaneously. On the lender’s conservative case, after paying myself and the debt service, the business left about $3K a year (based on the actual numbers, not the recast). That’s a business that can be running normally and still leave you nothing, with your house as the collateral. 6. Get three outside reads before you decide anything big, and tell them not to agree with you. Before I walked, I wrote the whole thing up, anonymized, and gave it to three different AI models with instructions to tell me if I was making a mistake. All three said walk, for different reasons, and one of them caught an error in my own debt service math that I’d been repeating for weeks. It also told me the thing I didn’t want to hear, which was that the next deal I was excited about needed the same discipline. Cheap, fast, and it took the emotion out of a decision that had a lot of emotion in it. 7. Walk clean. When I ended it, I left the broker a voicemail first so he’d hear it in my voice before he saw it in writing, then sent the seller a two-paragraph written termination citing the LOI’s termination clause. No price, no grievances, no list of what they did wrong. Thanked them for their time. The broker’s response was “I understand, very frustrating, I have a couple other companies coming up.” It’s a small market. The brokerage that had this listing will have the next one I want to see. The reputation you keep is worth more than the point you make. 8. The year wasn’t wasted. It was spent. In November of 2025 I’d never read a CIM. I didn’t know what SDE was or that brokers recast it, how a multiemployer pension withdrawal works, what an SBA lender’s coverage test actually requires, or how to write an LOI that a forty-year broker would mark up as “mostly tightening.” I know all of that now. The day after I walked, a broker sent me a different packet. It took me ten minutes to find the balance sheet problem that killed it. A year ago that brochure would have had me much longer. Where I’m at now No regrets about the decision; if anything it was overdue. The sunk costs are manageable and most of them roll forward. The day I walked I sent fourteen new broker emails, started a working capital line while I still have W-2 income, and had coffee with a broker who told me buyers like me are rare in this market, which I’m choosing to believe. I’ve got a couple of smaller, cleaner businesses in front of me now, no bank required, and a hard rule I didn’t have a year ago: if I’m the only one at the table who wants the deal, it’s not my deal. I won’t pretend the week after walking isn’t disheartening. It is. A year is a long time to point at one thing and then set it down. But I know this is my path. In the near future I’m going to own a business that I operate every day, and eventually I’m going to own several. This deal didn’t change that. It just taught me how to get there without betting the house on someone who won’t send a payroll report. Back after it. Happy to answer questions in the comments if any of this is useful to someone earlier in the process.