Jim Morales, CFA

@ron_ecomm @colindougherty @JonAlbertBlair I'd pull into COGS so inside CM calc.

@The10XCFO @colindougherty @JonAlbertBlair In your definition, is 3PL in OpEx?

THREE WAYS DTC BRANDS QUIETLY DIE @JonAlbertBlair has audited 50+ brands. The patterns: 1. OPEX over 10% of revenue. (Modern ecom is 10% or less. The lean ones? 6-7%.) 2. Scaling broken unit economics "scaling a problem instead of scaling profitability." 3. All the cash trapped in excess inventory. Most brands don't have a marketing problem. They have a math problem.

@oliverbrocato 15% CM Margin? 1.67 MER? Yeah good night Irene

100k/month in ecom means ur poor. 1/ $60k Meta 2/ $25k COGS 3/ $12k team 4/ $6k refunds, apps, random bs U clear like… nothing. But congrats. U bought urself a full-time headache.

@colindougherty @JonAlbertBlair A lot of brands using historicals to forecast from 2020-2024 are making a grave mistake. Many founders have never been thru a business cycle.

My 2c OpEx = everything below COGS: Admin Software (except SMS/Email/fulfillment software if you have in house WH, i include in COGS since variable and drives sales) Labor Costs/Professional Services - no labor in COGS unless you have people involved in manufacturing (rare) General Marketing OpEx - agencies, UGC T&E You could argue UGC is variable but I'd only include if they get cut of sales vs a flat fee. In house WH vs 3PL can be a big driver of OpEx from WH rent/CAMS.

@JonAlbertBlair how is OpEx defined? Is it everything after Contribution margin after marketing? I feel like sometimes people only count people when they say OpEx but don't count sass. Also do you count variable people expenses into OpEx: such as product photography, content creation for ad?

@eoghan I assumed Fin=financial

@Grace_9g9 @SullyBusiness How are they downloading anything if no internet on plane? How are they accessing the AI?

@SullyBusiness Sometimes my clients don’t have internet on the airplane. They need to download the PDF and read it later.

Learning to ask Claude for HTML outputs instead of PDFs or excel files will change your life.

@PaulEnu_ Thanks for heads up I fixed it.

@The10XCFO I was about to dm you, Jim Noticed your inbox is locked. Do you know about this?

The first ones at the carcass eat the best...

I think COVID distorted buying habits with all the free money from PPP/0% credit sloshing around, people at home bored looking for next thing and a new generation more digital native coming of age. Also this is first era a lot of people have operated in normal biz cycle rotation without cheap money providing steroid shots to keep music going. Not surprising marketing is a tad behind the times...

I don’t think we talk enough about how culture has shifted since COVID and how that affects how people need to be marketed to.

@LCSlates @levelsio Recourse. It's the unspoken oil that greases the credit card economy. There's more of "them" (customers) than of "you" (merchants). They have to keep them using the product for the biz model to work so they're higher on priority stack than merchants.

Well, here goes nothing. Building mine this morning thanks to a $2,333 annual chargeback. (thanks for the prompt @levelsio) Reason: Fraudulent. (10.4) Proof (18 pages, incl 1 main summary): User added 3 sites, added 4 team members, added 3 integrations, logged in 49 times in 54 days (Google Auth), generated articles and images for each site, published them through Cuppa into Wordpress, I have the IP address at login and initial checkouts (match), I have site WHOIS info, I have cryptographic info... I just have an overwhelming amount of evidence that this is bullshit* Stripe win chance: Low (1 out of 5) @stripe, why do you already take that money already? Can't you at the very minimum help owners out here - even just a little - and wait for the dispute to even clear or for me to even respond before taking this money? *There is just something very wrong that this has just become the norm. Why?

I’m sure this can all be done via code but I ain’t playing w all that at C level (and I’m former Unix programmer) - automated daily briefing posted to Slack every morning showing everything across calendar, email, Fathom transcripts, Slack channels etc - automated summaries of podcasts - weekly email newsletter summary - automated competitive intelligence on email campaigns - automated research brief on what’s trending in X area on X and Reddit I’m tracking Value to me is scheduling these off hours to avoid token burn and that I’m not in terminal. Would love to do via cloud or better interface to manage everything so open to suggestions.

Anyone actually using Cowork for anything interesting? Between code and chat, haven't really found a use case

@BryanBumgardner @northbeam - CAC trends by vertical (DTC, e-commerce, SaaS, etc) - CTR trends across platforms over the last 12–18 months - CPMs across the major ad networks, same timeframe

I'm sitting on @northbeam's database of aggregated client data. I'm talking about billions of dollars in ad spend. Like $100 mil a week. I can generate historical ROAS, CAC, CTR, whatever you want to see, over time. What do you want to see? Taking orders now 👇

Classic barbell breakpoints on display here In slowing-meh economy, single product or slim catalog biz with no moat and low brand equity get squeezed hard. <10M 10-50M getting benefit from the sub 10M pulling back and residual momentum from breaking out of the sub 10M land. Larger catalogs typically, customer base big enough to monetize off repeat biz vs new, better unit economics from some scale. 50-100M is that no mans land, big enough to be complicated and have run out of steam from 1st growth spurt but not enough brand equity to keep chugging along. 100M+ gets benefit from all of this from brand halo and are more in the staple category that consumers fall back on during downturns. Of course there are brands crushing it/sucking wind across all these bands so this is not applicable to all cases.

💀 May was terrible for SMBs, especially in DTC and ecomm. Looking at @northbeam aggregated May data, compared year-over-year: 🔻 Businesses making under $10m a year spent a lot more than others, but saw horrible decreases in MER and CAC, with paltry revenue returns. 🤔 To nobody's surprise, CAC got more expensive for all types of businesses - but iconic $100m+ businesses mostly dodged the pain. 💸 In tough economic environments, consumers buy what they know. Disrupters can't break in when there's no disposable income. Full report forthcoming. What numbers do you want to see?

@samrags_ Who buys from merchants with cash and gives up recourse?

@The10XCFO Tell me how customer recourse works w cash

Instant settlement is the key. Card networks are optimized for the consumer: the merchant takes all the risk. The merchant delivers the goods now and accepts T+30 settlement. They're extending credit to an agent. They must believe at the end of the month, they'll be made whole In practice, they might want to: - Calculate the agent's historical payments, historical chargebacks - KYC / KYB / KYA - Warm up period for unknown agents - Require an enterprise service agreement - Get a credit report Or... they could just use stablecoins. Instant settlement. No merchant risk. Sell to any agent today.

@clayyroy This just confirms my M&A thesis there are lots of future diamonds out there if you know what to look for.

Four years ago, I listed my brand on Flippa for $1.1M A buyer was close to taking it, but after too much back and forth, I gave up on the deal and just kept going. Last week we did $2.7M in 7 days. The hardest part of building something isn’t the work. It’s staying in it when it feels like nothing is working. When the numbers are flat, when you start questioning every decision you made before. That’s where most people quit. That’s exactly where you have to stay.

@aryanpour Land grab games work great... until they don't...

Most Fintech banks sacrifice long term sustainability over short term growth. And there are still Fintechs following the EXACT same strategy that led Parker to bankruptcy. If an offer sounds too good to be true, it probably is.

@awxjack How does recourse work in a2a?

The next generation of U.S. B2B founders will choose A2A because startups have to go global on day one, and bank-to-bank payments are already the default in the world's fastest-growing markets. Businesses that build on A2A rails from the start won't just reduce costs, they'll have the payment infrastructure to scale across borders without rebuilding from scratch.

@pephufen Depends on what you’re selling Supplements? Cool Stuff for 60 year old women? Ehhhh they’re prob not on candy crush

I believe AppLovin will overtake Google's share of wallet for most D2C brands in 2026/2027. Here’s why: 1. Cold audience at scale AppLovin's consumer platform only rolled out about 18 months ago. For our clients Axon is already the third largest ad channel, behind Meta and Google. Our clients spend roughly 70 percent of their ad dollars on Meta because Meta converts cold audiences with no buying intent. A good product, a strong algorithm, and good creative can persuade someone who was not looking to buy. That means scale is possible. Google does not do this well. It captures demand. It does not create it. So the scale is limited. The only platforms that tap the same cold audience layer are TikTok and AppLovin. That is the gap Axon walks into. 2. The pixel is the moat. And Axon's runs on a different graph than Meta's. For an ad algorithm to predict who will convert, it needs to tie web events back to a known identity. Without that, the model is guessing. Meta does this through its logged in user graph. Every event captured by the Meta pixel gets tied back to a Facebook or Instagram profile with years of behavioral history. That is what makes the model work. The pixel without the user graph is useless. From what I understand, AppLovin built a different kind of graph. They spent years running their SDK inside thousands of mobile apps, mostly in gaming. They know which device played a game last night, downloaded an app last week, made an in app purchase last month. When the Axon pixel fires on an advertiser site, the conversion gets tied back to a device with a real behavioral history. The model trains on that. Two things follow from this. First, the Axon graph is independent from Meta's and resilient to the iOS privacy changes that broke parts of Meta's attribution. Second, it compounds with every new pixel install. Right now, Axon's pixel is on a small fraction of US ecommerce sites. When AppLovin opens the platform to all advertisers, that footprint scales fast. More pixels means more web events tied back to the device graph, which means a more accurate model. That is what compounds. 3. More advertisers means a better model. yrevEbuyer types. This is the same flywheel Meta rode for a decade. We have already seen a meaningful step up in ROAS across our Axon accounts since the latest model update a few weeks ago. The model improves every quarter. I see no reason to believe that stops. 4. The team is the variable that does not get priced in. The AppLovin team is one of the sharpest I have watched operate. Driving incremental revenue at a CPA or ROAS that works for an SMB is the only thing that matters in this business. They get that So the setup is: cold audience platform, expanding pixel footprint on a graph independent from Meta's, sharp team shipping model improvements every quarter. That is the gap that compounds.

@iandmacomber @AnthropicAI @tryramp @Rich_Barton I see request only on my side. This a GA release or beta to select few?

Today, @AnthropicAI added @tryramp Data as a connector in Claude. You can now ask Claude what 50,000+ businesses are actually spending on and get an answer grounded in real spend data. Vendors, categories, growth, switching patterns, and more. For decades, @Rich_Barton companies have focused on "turning the lights on." Homebuyers have @zillow, job seekers have @Glassdoor. But how companies actually spend, operate, and make decisions has stayed dark. Today, that's changing. The people best served by transparent market data aren't the ones who can afford Bloomberg terminals. They're the founder sizing a first contract, the procurement lead negotiating a renewal, the researcher studying AI adoption, the agent booking software on a user's behalf. AI has made those users far more capable but only if the data they need is actually available. Ramp Data is live in Claude today, and accessible via API, MCP, and CLI. Every market gets better with information transparency. B2B software is next.

@willnitze copy product and ads easy... now try to copy sourcing, pricing, measurement, inventory optimization etc...

"Copycat" brands invariably run into a huge issue: They don't know what, specifically, to copy. Because they don't know which specific combo of variables drives value for the brand they're copying. So they guess. And, predictably, their guess is way off.