At 4:28 a.m. on a late June morning, a founder answered a query I had posted on Peter Shankman’s Source of Sources (SOS). I was looking for case studies about marketers using AI in ways nobody else was, and his pitch was among the best of the 35 I received. He had pointed a computer-vision tool at 4,500 product photos on his review site, fixed the bad ones, and watched clicks climb 24 percent.
I’ll call him Jack. That’s not his name, and by the end you’ll see why I’m leaving it out.
Over the next 111 days, I sent Jack nearly two dozen emails. He sent back three replies, and one of them held nothing but my own note quoted back to me. On Oct. 15, my editor’s deadline passed, the story died one photo short of done, and I didn’t get paid, because freelance writers get paid on publication.
I’m not writing this to settle a score. Jack did most of it right, which is exactly why founders, and the people who run their PR, should study how it went wrong.
He did the hard part right
Most AI pitches I see are chatbots writing LinkedIn posts with the serial numbers filed off. Jack’s wasn’t. His site reviews expensive gear for sports enthusiasts, the kind of buyer about to drop $600 on something they have never held. The reviews were careful, but many of the photos next to them were a mess (out of focus, too dark, or shot from an angle that had nothing to do with the feature described two paragraphs above). Time on page had fallen 11 percent.
Call it the dirty bathroom problem. The kitchen can be spotless, but a filthy restroom makes the diner wonder about the kitchen anyway. On a review site, the blurry photo is that bathroom.
So Jack defined what a good photo looks like on his site and ran all 4,500 through an off-the-shelf vision tool for about $250. The tool flagged 650 images (14 percent), sorted worst first. He looked at each one, and only 350 actually needed a reshoot or replacement, which means nearly half the flags were false alarms. He told me that without being asked, and founders rarely volunteer the mistakes
The AI didn’t decide anything. It shrank a pile too big to look at into a shorter list Jack could work through, and the final call stayed with Jack.
Click-through on the tables where he earns affiliate commissions went from 3.4 percent to 4.2 percent over 60 days, which came to roughly $1,000 in extra earnings. He called that “relatively modest” before I could. He also wouldn’t claim a controlled experiment, because traffic from his target audience swings with the season.
His first pitch didn’t sound like any of this. It promised “deep learning computer vision to isolate structural anomalies” and “technical imagery validation.” It took 12 questions to pull the plain-English version out of him (a shorter list, worst first, and a person makes the call), and that version was the story. A modest number told plainly beats a big one wrapped in jargon, every time a reporter must explain it to an editor.
Then he stopped at the 1-yard line
Let’s backtrack a bit. On June 29, I told Jack he’d made the cut and asked for a half hour on the phone. Four follow-ups later, it was Aug. 31 and I still had radio silence.
So I sent “my fifth and last note.” I told him a yes or a no would close the file with no hard feelings, and I ended it with two lines I’d tape to every founder’s monitor if I could: “Either answer works for me. Silence is the only one I cannot use.”
It worked. He wrote back within four hours, apologized, said some issues had taken up his attention, and offered to answer by email but not by phone. I sent 12 questions on Sept. 1. He answered all 12 on Sept. 7, and the answers were really good.
I wrote the story, and my editor liked it. On Sept. 16, I sent Jack the draft with a few follow-up questions and a deal-breaker request: a single before-and-after photo, the image the whole piece needed. I even offered to dig the old photo out of the Wayback Machine myself if he would just name a product.
Then nothing.
I followed up on Sept. 21, Sept. 24, Sept. 28, Oct. 5, and Oct. 11. On Oct. 13, I tried his personal address, then had to send a second note because I had attached the wrong version of the story. (Four decades of telling executives to check their attachments, and I’m the one sending a “whoops” email.) On Oct. 14, I told him plainly that if the story died, I would probably publish it as a case study on my own site.
This is that case study. Nobody wrote back. Not Jack. Not the two family members I copied on the email.
I lost the fee and 20+ hours of my life I’ll never get back. Jack lost the opportunity to raise his visibility with tens of thousands of marketers that would have included a link to his site heading into the holiday buying season. And he’d have a writer and brand strategist who would have taken future calls without a second thought. He spent $250 to fix his photos, then left the free coverage sitting on the table.
The room founders forget
I write a lot about the rooms where your words get read by people who can hurt your business: the board room, the investor call, the prospect’s inbox or conference room, even your own all-hands. But there’s one other that founders forget: the reporter’s inbox.
I spent 20 years on that side of the desk, and I can tell you how it works. Reporters don’t remember who said no. They remember who ghosted them. A clean no costs a reporter an afternoon, while silence costs a week of follow-ups, an editor asking where the story is, and sometimes a paycheck. When your company’s name comes up in a story meeting a year from now, that’s the part they will remember.
I’ve been on the other side of this. I had a client who got a pitch from a reporter for a story that could have been great for us if we handled it right. We asked for more details and his initial questions. I gave my client a page of talking points that would have helped the writer and probably set us up for a positive story in the future. But my client said he didn’t trust the media and told me to stop responding and focus on other things. Their biggest competitor got the quote.
Jack’s version of this is common at small companies. In this case, the founder pitched the story himself at 4:28 a.m., which meant the founder was the only one who could finish it. When his attention went somewhere else, nobody else owned the follow-through, and that gap is what a PR team (or even one assistant with a calendar) is there to close. Plenty of the 35 responses came from PR people. I picked the founder because his story was better, and I took the risk that came with it..
A pitch is a promise. When you answer a reporter’s query, you’re telling them you have the goods and will deliver them on their deadline. Jack had the goods. He just never delivered the last 10 percent.
Before you answer the next query
If you or your team answers reporter requests through Source of Sources, Qwoted, or anything like them, do four things before anyone hits send.
1. Put the receipts in a folder first. The before-and-after, the baseline number, the dates, and a screenshot of the dashboard. If you can’t find them in 30 minutes, you aren’t ready to pitch.
2. Name one owner. One person, by name, who answers the reporter within 48 hours, even when the answer is “next Tuesday.”
3. Know your four numbers. Where you started, where you finished, what it earned, and what else changed in the same window. Those are the questions that decide whether a story runs.
4. Write your exit line now. “We need to pull out, and I’m sorry for the trouble.” One sentence keeps the door open for the next story.
Does this apply to you? Open the last pitch you or your team sent a reporter (or the last reporter query you answered) and check whether there’s a folder of receipts and a name next to it. If there isn’t, start the folder today.
Jack may have had a good reason for going quiet. Something distracted him over the summer, and maybe it came back. If so, I hope he’s OK, and the offer to talk still stands.
But the lesson doesn’t depend on his reason. Every founder who pitches a story is one busy month away from becoming Jack.
If your team is fielding reporter requests and you’d like an extra set of eyes on what you’re promising before it goes out, set up a call with me here.
