The Manufacturing Automation Podcast

For the 100th episode of the Manufacturing Automation Podcast, Matt and Michael get candid about growth, focus, and the trade-offs that come with both. Matt shares how Develop doubled headcount in six months while ending up in a stronger cash position than when the growth curve started, and what it feels like to put together three nearly identical proposals where the only real difference is the customer’s product. He explains Develop’s pilot-to-fleet approach for paper converters: a more aggressively priced first cell to build trust, followed by a multi-year rollout where the real ROI shows up.Michael admits Gimbel can feel like two businesses that depend on each other: high-value turnkey integration and e-commerce products like CoolantClear and TumbleBlast. The two dig into why getting a first system into a facility is easy for Gimbel but expanding past five machines is hard (the exact opposite of Develop’s problem), why defense and medical customers will not share machine data, and why Gimbel sees its recurring revenue opportunity in simulation software instead of machine monitoring. They also talk about rising Q4 ad costs, whether product lines should live on their own websites, why considered purchases break ad attribution, and Matt’s test of video mailers sent to top leads. Matt’s advice to Michael: the problem may not be either business, but splitting focus across too many channels at once.On engineering, Matt walks through power distribution and air compressor sizing for a facility where the machines are never the same, and Michael suggests renting capacity for peak demand. Michael shares the final RotoGrip design and the AxisMount Duo, built to bring two-vise automation to compact machines. The two discuss why stockouts rarely cost B2B orders, and why both hosts now favor equipping people with AI over replacing them. In continuous improvement, Matt covers lender meetings and sales SOPs, Michael raises the question of when a growing company needs a controller, and Gimbel shares its move to HubSpot payment links, machine simulation across roughly 300 machines, and a new proposal tracking module.

Matt is back from two weeks on the road with big news: Develop has signed its first dedicated salesperson. He shares why the last three trade shows were record shows for lead quality, and why he credits better processes more than the shows themselves. The two also talk about why small business data is often unreliable for the first few years, how proposal tracking can turn into noise if no one acts on it, and what separates a valuable industry panel from a scripted one.Michael opens up about the saying that between a service business and a product business lies death, and how Gimbel is working in that space by alternating focus weeks between B2B turnkey sales and e-commerce. He explains the vibe-coded mini CRM he built for IMTS follow-up and the difference between working warm trade show leads and hot inbound ones. The highlight is his new approach to ad attribution for considered purchases: measuring marketing efficiency ratio by product line. That analysis showed the CoolantClear ad he turned off, which looked like a 2x return on ad spend, was actually part of a 30x product-line MER, meaning Gimbel had been underspending. He walks through rebuilding the CoolantClear page around real customer objections and setting up separate prospecting and retargeting campaigns. Matt shares the hiring sequence that worked for Develop: a top-of-funnel marketing hire first, then a salesperson once the process was documented.On engineering, Matt covers facility planning, including retention pond design, core drilling for the bridge crane, power distribution for machines that each need different power, and how to size air compressors without overbuilding. Michael shares his plan for small backup compressors at the machine, and the two compare compressor brands that have held up. Michael also previews the bar-fed mill workflow and the new AxisMount Duo, with two opposing AutoVises on a fourth axis. In continuous improvement, Michael explains Gimbel’s full kanban audit and why system compliance matters as much as the system itself.

Matt is in a holding pattern on several large machine builds that are now waiting on board approvals, some after scoping since Automate 2025. He explains when Develop charges for proof-of-concept engineering, and why quoting nearly everything for a couple of years, even with a terrible conversion rate, is what eventually lifted their close rate about 4.5 times. Michael, meanwhile, is on pace for roughly 40 proposals a month when 15 to 20 would cover his goals, and the two dig into what both have noticed lately: founders love novelty, but profit comes from doing the same thing with flawless execution over and over.In sales and marketing, Matt previews Pack Expo and TLMI, and shares how an intense internal interview process produced Develop’s best e-book yet, written for label converters. He tells the story of a site visit where the real automation opportunity was hiding in the back corner of the plant. Michael breaks down what hurt his top of funnel this summer, rebuilding Meta lookalike audiences, building searchable answer documents for CoolantClear and TumbleBlast, and why moving upmarket means giving internal champions the case studies and materials they need to sell automation to their own leadership. The two also discuss why machine tool dealers may not be the right distribution partners for lower-ticket automation, CRM choices, website heat maps, and a friendly disagreement over how far UTM tracking should go.On engineering, Michael shares a dual AutoVise trunnion setup finally getting a real customer application, and walks through vertical bar feeding on a fourth axis, including how it compares to a Swiss. Matt explains why he pushed back on a customer who wanted a cobot with area scanners instead of a fenced industrial robot. In continuous improvement, Matt covers his crane quote, SOPs down to edge cases, and his strategy for NIST and CMMC: complete the audit now, but wait to get listed until a customer requires it.

Matt is finalizing a 13-page lending story and three-year forecast ahead of meetings with lenders, while Michael is deep in post-IMTS follow-up and caught in the pinch of being big enough to have big-company problems without the headcount to fix them. Michael shares what diving into the data revealed: a 29-day average time to close, a floor of four turnkeys a month, and the fear of overselling. Matt’s advice: sell like you will oversubscribe, put first come, first served in your contracts, and have the conversation if it happens.In sales and marketing, Matt is down to two sales candidates, preparing an e-book for the printing and converting space, and heading into a run of Pack Expo, TLMI, and a speaking panel. He explains why his network built over the last decade still drives more revenue than every trade show combined. Michael talks through e-commerce volatility after a Google update, sending 46 hardware quotes in a single week, and the attribution problem of not knowing whether LinkedIn, Google, or something else earned the sale. Inspired by the book Oversubscribed, he admits founders will try anything but sales, and lays out a shift toward enterprise customers, distributors, and complete solutions, with Shopify treated as upside instead of the foundation.On engineering, Matt shares a bridge crane lesson that saved real money: one 5-ton bridge with two 2.5-ton hoists instead of a 10-ton system. He also has senior and junior controls roles open. Michael admits to feature creep in his proposal software, compares Loom video walkthroughs against formal proposals, and previews fourth and fifth axis bar feeding.In continuous improvement, Michael explains how logging B2B orders as $0 Shopify draft orders streamlined fulfillment but fed bad conversion data to Google and Meta, contributing to a roughly 30% drop in return on ad spend, and why unwinding a process the whole team loves has to be done carefully. Matt covers ClickUp, SOPs that are finally being used, an aftermarket recurring revenue model, a decision authority table to keep routine questions off his desk, and why he will rent a 3D scanner before buying one.

Q4 quoting season is here, and both hosts are running on empty. Michael is only 10% through IMTS follow-up and already booked solid for two weeks, while Matt had just over two hours of open calendar all week. They open with price increases (Develop is on its sixth in two years, and Gimbel adjusts pricing on the 15th) and why your products and processes getting better justifies charging more.The core of the episode is a question Michael has been wrestling with: should a product-focused automation company take on projects that are a 75% fit? Matt shares how Develop handles high-risk, bespoke work, borrowing from large construction: listing assumptions, assigning risk categories, using allowances instead of fixed bids, and pricing every off-standard request as its own line item. The takeaway is simple: price for asymmetric upside, or do not take the project.In sales and marketing, Matt narrows his sales hire to three strong candidates and explains Develop's push into the label printing market after joining TLMI. Michael gives a practical walkthrough of UTM links, why untracked Instagram, LinkedIn, and PDF links all show up as "direct" traffic, and why Facebook and Instagram ads are underused in B2B manufacturing. He also shares early Bing ads setup.On engineering, the two talk about posting innovations before patents are finished, playing offense versus defense with IP, Gimbel's upcoming mini pallet tending system, and a new automated 16C collet with an external rotary union for bar pulling. In continuous improvement, Matt's ClickUp rollout hits 80% and he plans a mobile integration van, while Michael explains why Gimbel is consolidating out of its second building and moving from waterjet to laser.

One week after IMTS 2026, Michael breaks down what it actually cost to bring a Robodrill to Chicago: $40K to $50K before labor, from rigging and freight to hotels and ads. He shares how a 10x10 booth pulled in 350 leads, how 22 of 25 CoolantClears sold right off the floor, and how a geo-targeted Instagram ad reached a buyer in his hotel room who sent a colleague to the booth.Matt explains why moving his vacuum fixture product onto its own domain took it to the top spot for its main keywords, and how he built a Claude project that drafts technical customer emails with near-perfect accuracy. The two dig into hiring a senior salesperson, why Matt went senior-only and is open to hybrid candidates, the risk of a single point of failure in sales, and whether agencies or a dedicated growth manager overseeing freelancers get better results.On the marketing side, Michael shares what happened to Google performance after turning off Meta ads, why blended ROAS and MER may be the better metrics, and why he is betting on Bing ads for a 40 to 60-year-old buyer. He also admits that 30 phone calls drove more revenue than 3,000 emails.In operations, the two cover booth lead capture rules, kanban compliance ("technically solved but not culturally solved"), taking IntraLoad orders from existing customers to fund better unit economics, and why Gimbel is raising implementation pricing to bring average deals from around $30K to $35K to $40K. Plus Matt's team taking over weekly resource planning without him, and a friendly SOP showdown between OneDrive with Copilot and Notion with Claude.

Michael joins fresh off six days at IMTS 2026 in Chicago, and Matt is running on four hours of sleep after his own label printing show in Rosemont. Michael breaks down what made this IMTS night and day compared to two years ago: a loaned FANUC Robodrill in a 10x10 booth, the last-minute bailment insurance scramble to get it there, and geo-targeted Instagram ads that pulled CNC machinists within a four-mile radius straight to the booth. CoolantClear 3 officially launched and is finally back in stock.Matt shares how his team scores and enriches trade show leads before they hit HubSpot, and why they make visitors wait so every conversation gets complete notes. The two compare proposal workflows (HTML versus PDF), talk through how to prioritize 100+ post-show proposals by buyer persona, and dig into ad fatigue, attribution blind spots, and what happened when Michael turned off $10K a month in Meta spend.Also covered: Matt's 22,000 sq ft facility addition and ClickUp go-live, Michael's hard lesson about letting one product's demand surge pull the whole business off balance, and an honest conversation about how AI can leave you with seven projects 80% done instead of one finished. Plus the story of Teddy's 19-hour escape through Chicago.

Michael opens with one of his rougher weeks of the year. A silent Shopify bug — only visible to brand-new users in a completely cookie-free browser — killed e-commerce sales for six to seven days and took 80 hours to track down. Combined with ad fatigue on Meta, a 43-day deal close lag hitting turnkeys right now, and IMTS prep reshaping the entire booth layout at the last minute, it's been a week of everything converging at once. Matt's week is a different kind of pressure — two full-day site visits, contracts in multi-month red-line negotiations, and a sales bottleneck that's no longer a pipeline problem but a pure capacity one: his team is quoting more than ever in company history and still can't keep up.The sales and marketing section is one of the most tactical the show has had. Michael makes the case for conquest ad campaigns — buying competitor keywords, running direct AutoVise versus Air Vice comparisons, targeting CoolantClear against Rotoclear — and walks through how his Meta ads quietly poured all budget into one winning creative while 60% of his ad library never saw a single dollar spent. Matt admits he's never run a competitor keyword campaign and gets talked through why it's probably already happening to him. They go long on the AI content saturation problem — everyone's marketing is better than ever, nobody's converting because everyone else's is too — and why the only differentiator left is the content that's intentionally hard to make. Matt gets a LinkedIn posting mandate from his marketing team. Michael's multi-AI debugging workflow using Claude, Manus, and Groq in parallel is what finally cracked the Shopify bug.Engineering covers Develop's largest machine ever — 30-plus feet long, split for transport, loaded by forklift because it won't fit through a door in one piece — losing two full weeks to one supplier on large milled plates, and a new HMI architecture debuting at Lupe for the packaging market. Michael is finalizing IMTS demo assembly, filing provisional patents he can't discuss, and stashing 30 CoolantClear units in a hotel room to sell in ways that may or may not comply with show floor rules. The CI section closes with the most substantive supply chain conversation the show has had — Matt making the case for deep blanket POs with fewer, growth-minded vendors as a strategic relationship rather than a procurement tactic, and drawing the analogy to a building lease: nobody hesitates to sign a five-year lease on the space that generates their revenue, so why hesitate to commit to the product that does the same?

Matt opens with decision fatigue at a peak — seven hours of back-to-back meetings on Thursday, four reports on his desk, and a team that's been waiting for him to put authority in writing that he assumed they already had. The fix isn't more hierarchy — it's a decision authority matrix that makes explicit who can approve what, up to what dollar threshold, without Matt. Michael opens on his first full week with an EA and a quieter inbox, and lands on an uncomfortable insight: he'd been using email and EA-type tasks as a refuge from the harder, higher-stakes decisions he was avoiding. Both are figuring out what their jobs actually are now that the scaffolding is coming off.The sales and marketing section covers a lot of ground. Matt is watching machines his team bid on three to four years ago — machines that went to the low bidder — getting decommissioned in the corner, and customers spending 50 to 100% more to start over with Develop. Two competitors released marketing content this week copying the scoping theory Develop has been building for four years. He takes it as confirmation, not a threat. They go long on margin philosophy, risk scalers on new-to-world machine builds, when to strategically compress margin to enter a market, and why B2B conviction during the quiet years is the only way the flywheel eventually starts. Michael is running 20–30X return on Meta ads — until it drops off a cliff, every time — while Google holds steady at 10X without being touched. The CoolantClear price increase campaign was their most successful email campaign yet, and his marketing hire Nate's idea.Engineering covers Develop's five-robot Project 29 machine now in full build with all robots landed, the critical path getting squeezed by one supplier on large milled plates who will not be supplying again, and the IMTS booth being completely redesigned around a last-minute Fanuc RoboDrill that nobody wanted to say no to. Michael is finalizing the in-house CoolantClear 3 production supply chain, narrowing down a product launch priority order across four items, and filing a cluster of patents that are draining to read and worse to draft. The CI section closes with Matt running a full audit on all active CIs after realizing the board has grown beyond what the team can execute in parallel, and a pointed conversation about ClickUp, Gim Command, and what it actually costs to hire for a homegrown ERP that nobody else in the world has ever touched.

Matt opens the week with something he hasn't felt in a while — his passion back at full volume. Site visits, marketing roadmap sessions with the UK hire, deal structuring on strategic accounts, and a surprise shop tour at Ring Brothers, the company he helped build before Develop. It's energizing, but he knows high-passion Matt can be a lot for people who haven't seen it in a while. Michael's version of the week is quieter but equally significant: he ran an experiment where he just stopped doing EA tasks — didn't sort email, didn't answer the phone — and found four hours he didn't know he had. The in-person EA candidate came in for her trial day and got hired on the spot.On the sales and marketing side, Matt walks through a pattern that's now showing up repeatedly: companies that chose the low bid over Develop two to three years ago are coming back with machines they want to tear out. The email that landed in his inbox that morning flat-out said Develop's was the most thorough proposal, but the board went low bid anyway. His team wants to start a whiteboard over-under bet on how long each one takes to call back. Michael, meanwhile, is deep in value engine mapping — a Ryan Deiss framework for flowcharting the entire customer journey from ad click to money in the bank — and walks through how it helped him identify where Gimbel's funnel is leaky rather than just too small. IMTS prep is intensifying: a RoboDrill is now going into a 10x10 booth, CoolantClear units are getting shipped in a second crate to potentially sell on the floor, and nobody has confirmed whether selling product on the show floor is actually allowed.Engineering covers Develop's five-robot Project 29 machine getting derailed by one supplier on large milled plates, a factory acceptance test that caught issues before anything shipped, and a facility expansion plan that just hit a wall — precast concrete lead times jumped from three months to ten because of data center construction consuming Wisconsin's supply. Michael is filing a cluster of patents he can't talk about yet, has a rotary union collet chuck nearly ready, and is wrestling with product launch priority order across four items that could all ship within three months. The CI section closes with Matt deciding to delay his line of credit renewal rather than forecast through a normalization that keeps moving, Michael turning off Shopify notifications and finding more peace than he expected, and a pointed conversation about EA-as-forcing-function — both hosts landing on the same insight that having someone who will actually make the call, even imperfectly, beats the founder stalling indefinitely in their own inbox.