You Sell Technology. I Run a Business. You Never Understood Mine

Hey, it's Tony. Over the course of my B2B media career, I sold hundreds upon hundreds of millions of dollars in advertising, marketing, exhibition, and sponsorship programs to technology companies. As my career progressed and the P&Ls became my responsibility, I became the customer, and this gave me a completely different view of these technology companies. The switch was eye-opening. I saw vendor claims, sales tactics, negotiating tactics, and the whole customer experience from the other side of the desk. And more times than I can count, I was struck by how out of sync a vendor's brand and marketing messages were with how their sales teams actually sold. It was jarring enough that it turned into a forcing function inside my own company. Sales and marketing had to start from a clear, ongoing understanding of the customer, their needs, their concerns, how they saw our products, where we were falling short. Here's the problem. In markets where demand stays high, and B2B technology has been one of those markets for decades, that kind of customer understanding is hard to sustain. The skills that get rewarded instead are competitive skills, sharper positioning, faster follow-up, tighter deal execution skills that have nothing to do with understanding the customer in front of you. For 50 years, strong demand covered for that gap. In 1975, the entire U.S. stock market—every public company, every industry—was worth about $700 billion. Today, 5 technology companies alone are worth almost $21 trillion. This is about to change as agentic AI moves into procurement and vendor evaluation. Not understanding the customer stops being a blind spot. And starts becoming a liability. That's what this episode is about. Okay, let's get to work. This episode is called You Sell Technology, I Run a Business, You Never Understood Mine. 50 years of growing demand has convinced B2B technology companies they win competitive bake-offs because they understand the customer. They don't, and Agentic AI is about to expose them. Every technology vendor I've ever sat across the table from was good at exactly one thing, beating the other vendors on the shortlist. Smarter packaging and pricing, sharper deck, faster follow-up, a reference call with the right customer logo, the right relationship at the board level. By the time the deal closed, I knew more about how well they competed than I knew about whether they understood my business at all. I usually found out the difference during implementation, and I found out for certain at renewal. That gap is not a sales training problem. It's the whole story of B2B technology for the last 50 years, and almost nobody selling technology today has had a reason to notice it until now. Customer demand did the selling. Nobody really learned to. Here's the thesis: 50+ years of rising corporate technology spending has not materially changed how technology companies position, sell, implement, and renew their products with businesses. The GTM muscle memory built in the mainframe and ERP eras is largely still the GTM muscle memory in use today. It's not because technology companies never had the incentive to get better at it. It is because they never had to. Once the personal computer put a connected terminal on every desk, demand for business technology became so persistent, so self-renewing, that closing a customer stopped being the hard part of the job. Beating the 3 other vendors chasing the same budget was the hard part. Technology companies got very good at that competition and very good at telling themselves that winning it was the same thing as understanding the customer. It never was. It just didn't matter because there was always another wave of demand behind the last one to paper over the difference. We were billed for modules we never used. I've seen this move more times than I can count, a competitive pitch dressed up as customer focus. It shows up most often at renewal, the moment the relationship has the least leverage and the vendor has the most. One renewal stands out. A multimillion-dollar contract came up and the increase landed north of 10%. The justification? New fees for modules that were apparently in the contract that we had never used. Fees that could be negotiated away, but only if we signed the renewal. From there, it was scorched earth. The account team escalated from our Chief Technology Officer straight to our CFO with a thinly veiled threat about what would happen if we looked at a competitor. The rep running the account had never once set foot in our offices. That is competitive skill working exactly as designed. Nobody on that account understood our business. They understood how to extract a signature under pressure. And for a long time, in a market where demand did most of the selling for them, that worked really well. I've had the good fortune of selling in bull markets and the good fortune of leading businesses through bear markets where the rules changed and customer focus was the competitive differentiator. Which market I was in never decided that. What showed up the moment the leverage ran out did. 2 scales get confused is one. There are 2 distinct capabilities buried inside every technology company's win rate, and the industry has spent 50 years measuring only one of them. The first is competitive skill, the ability to outposition, outpresent, and outnegotiate the other vendors on the shortlist. The second is customer understanding, knowing with precision the actual job the buyer's business needs done and building a product and relationship around it. A high win rate can be the evidence of either one. Most B2B technology companies have never built a way to tell which. Plot those 2 capabilities against each other and you get 4 quadrants, not 2. Almost every established technology company believes it's the category leader because its win rate has always been strong. Most of them actually been the borrowed win, protected the entire time by demand strong enough to make the difference invisible. What was hiding the difference? The evidence that the borrowed win is more crowded than anyone wants to admit is already sitting in the data. Gartner has found that 60% of technology buyers involved in a renewal decision regret nearly every purchase they've made. McKinsey's own research found that only 16% of digital transformations both improve performance and sustain that improvement long-term. Those are not implementation statistics. They are customer understanding statistics showing up 2 steps downstream for where the deal was actually won. A vendor that understood the job would not produce those numbers at that scale for that long across an entire industry. A vendor that was simply good at winning competitive bake-offs would produce exactly this. The industry has had the data in front of it for years and read it as a services problem instead of what it actually is. Why the COVID is coming off now. The demand cushion depended on one structural fact: the buyer's evaluation ran through a human, and a human can be outpositioned, flattered, and outlasted. Agentic AI removes that fact, not gradually, but as fast as agents get adopted inside procurement functions. I've written before about what happens when the buyer's shortlist gets built by an agent before a rep is ever involved in the buyer left the building. An agent evaluating vendors on behalf of a buyer does not care how the deal was packaged or priced. It does not respond to a sharper deck. It does not remember a good reference call. It checks whether the product demonstrably does the job it was asked to evaluate, and it moves on if the answer is unclear. Competitive skill, the entire capability the industry spent 50 years perfecting, does not register on that test. Customer understanding is the only thing that does. That shift lands hardest on the people who can least afford it. I've also written about the asymmetry between the rep who walks away from a bad decision and the buyer who staked their internal credibility on it in The Decision Was Made Without You. That buyer is exactly who is now handing evaluation work to an agent built to catch the gap between a vendor's pitch and a vendor's actual fit because they were the ones left holding the outcome the last time the pitch won and the fit didn't. The borrowed win quadrant only worked because the buyer couldn't see the difference before signing. The agent can. What this actually requires— if you sell technology, stop measuring win-loss against the competition and start measuring it against the company's job actually got done. 12 months out, not 30 days after the signature. Treat a strong win rate with no customer understanding evidence behind it as a warning. Not a victory lap. Rebuild the account team around the person inside the buying business who is personally exposed by the decision, not the procurement function that approves the budget, because that is who is about to be doing the evaluating with an agent at their side. If you run a business, stop running evaluations that reward the vendor who is best at being evaluated. Do the work of defining the job you actually need done with the same rigor You'd apply to hire before a single vendor is in the room. Hold every vendor accountable to the outcome past the signature, not just the promise before it. And treat renewal, not the initial close, as the real test of whether the fit was ever real. 50 years of demand— what strong demand always does— it led a lot of companies to mistake being hard to say no to for being right for the job. Agentic AI does not care how hard you are to say no to. It only checks the job. You've spent 50 years selling technology. Whether you ever understood your customer business is what Agentic AI is about to find out. Hey, that's it for this episode. If you found this useful, please do one thing. Follow the Uphoff on Media podcast on Apple or Spotify and leave a rating. It's the single best way to help another operator find this. I'm Tony Uphoff, back at it next week.