$80 on duct tape, 1.5 million views, 500 new subscribers

Matt Diehl, who runs the Eating Baltimore newsletter, had a video take off in a way he did not expect. A friend stuck about $80 to a post near the Baltimore water with duct tape and filmed it. The clip is about 10 seconds long, and across platforms it reached close to 1.5 million views.

He came back on the show only a few weeks after his first appearance because the results were too good to wait on. This conversation covers how the video was made, how each platform treated it, and how he turned the attention into email subscribers. It also gets into the restaurant directory he is building.

The idea was to copy accounts that were already working

Cash drop accounts hide money in public places and post videos of people finding it. TJ compares the appeal to MrBeast giveaways: people like winning, and they like watching other people win. Matt had seen the accounts doing well, and he sent his friend Find the Cash ATX as an example, along with a couple of others.

Matt did not do the drop himself. He Venmoed his friend about $80 or $100, and his friend taped the cash to the back of a piece of duct tape on a post in a touristy spot. The instructions were short: go slow, then pan out so people can tell where it is, like a scavenger hunt. Matt says it did well partly because it was so raw and unedited, and he has seen other accounts do even less.

Facebook took off, and Instagram barely moved

Matt's own count was Facebook at 1.2 to 1.3 million views, about 200,000 on TikTok and about 5,000 on Instagram. He also posted a follow-up video using the clip of the person who found the cash, which got more than 60,000 views.

TikTok moved first, hitting about 100,000 views within 24 hours. At around 48 hours Facebook had only about 500 views, and two days later it was at 50,000 and still climbing. Matt thinks his one mistake was letting Facebook cross-post to Instagram instead of posting to each app on its own. He is not sure why that mattered, and TJ says he has heard the same about auto-posting tools.

Hundreds of comments about duct tape

Matt says that as of the day he recorded, the video had about 500 comments and roughly 90% of them were people doubting that you can get money off duct tape. Some suggested hot water. He has not heard from the finder complaining, so he assumes it came off fine.

TJ's take is that a small obvious flaw in a post gets people to argue, and arguing is engagement. If you fix the problem, you probably get fewer comments. Matt agrees that even comments questioning whether the drop was real kept the video moving.

Turning the views into subscribers

Matt says the numbers were the point. His Facebook page went from about 4,500 followers to 14,000, and he had 500 new email subscribers as of that morning. On TikTok he went from about 10,000 to 14,000, but he had bought that account earlier, and says about 2,000 of that is new. Instagram added maybe 37.

What he would do differently is have a templated subscribe comment ready. On Facebook he could not pin a comment as the author, so he dropped his link at around 500,000 views, and that one comment brought in about 250 subscribers. After that he replied to comments, including trolls, with the link. His subscribe comment got over 100 likes, more than the top duct tape comment, which had about 83 to 85.

The finder showed up 53 minutes in

Matt copied one thing from Find the Cash ATX: the caption asked whoever found it to tag him. Someone found the cash 53 minutes after he posted and filmed it, when the video had under 5,000 views. That person was not even following the page. Matt says he expected it to take longer.

He is already thinking about the next one. He wants to put the cash in an envelope with branded stickers, and he lined up a restaurant owner near the harbor for a drop in August with cash plus a gift card. He leans toward monthly drops to keep them special, while TJ would try weekly with a bigger one monthly. TJ's own concern is how a restaurant keeps staff from grabbing the money, and Matt has not solved that yet.

What he edited and what he left alone

The only edit was trimming the first second, where his friend's girlfriend said "go." Matt did it on his phone, then posted from a draft with a ready-made caption. Setting the location to Baltimore also helped, both for who sees it and because the word is on screen at the start.

TJ points out that the cash drop accounts he looked at do not film or edit well, yet they have huge followings, and they have no newsletter or other business behind them. A local media company can monetize the same attention in many more ways.

The directory behind the newsletter

Matt's restaurant directory gets about 1,300 visits a month from organic Google search and one or two new subscribers a day. He first required an email to get in, felt that turned people off, and switched to a popup after about five seconds of scrolling. He also built an ad dashboard in Replit so sponsors can see impressions and clicks on their banners, and he used it as a bonus for the four or five sponsors he landed in early July.

He built it by pulling restaurant data by zip code from Outscraper, having ChatGPT convert the spreadsheet to JSON because Replit handled that far better than a CSV, and then handing it to Replit. His Baltimore directory has about 2,300 restaurants and the pull cost about $25. A smaller directory for his second newsletter has 165 restaurants and cost about $10 to pull. That one took him an hour and a half or less to build.

One restaurant claimed its listing after he posted about it, so he swapped the claim button for a verified icon on that page.

Ask the AI what it missed

Late in the episode they trade notes on AI tools. Matt had a ChatGPT-written SEO blog post improved by Claude, then sent back to ChatGPT, and says it now ranks on the first page for a Timonium restaurant search. TJ's simpler version is to ask the same chat to review its own work and say what it missed, then fix it. He says it works with any model and that it will usually list what it left out.