Growth Unhinged

Durability is an existential question for AI apps. Yet there’s been almost zero objective data on what's happening. To change that, I teamed up with ChartMogul where I’m an Analyst-in-Residence. We scraped the websites of 3,500 software companies, using AI to categorize companies as B2B SaaS, B2C SaaS, or AI-native companies (which might be B2B or B2C). Then we compared gross revenue retention (GRR) and net revenue retention (NRR) rates across these groups using ChartMogul data. What we found: 1. In aggregate AI-native companies had *even worse* retention than even consumer SaaS. It doesn't look like the recurring revenue of B2B. Median GRR was 23pp lower than B2B SaaS. Median NRR was 34pp lower. As Cassie Young recently said, many companies are in for a "gross retention apocalypse" as buyers scrutinize AI experiments, AI tourists move onto the next hot product & big platforms swallow LLM-wrappers. 2. On the optimistic side, AI retention has gotten much better since the beginning of the year. Median GRR jumped from 27% in January to 40% in September. I suspect many of the early AI tourists left; those who remain are more committed. Even still, it's nearly impossible to sustain hyper-growth when 60% of your install base walks out the door. 3. The best protection against churn is shifting from consumers and self-serve purchases (<$50 per month deals) to larger B2B purchases. AI-native products that sell for <$50 per month see just 23% GRR and 32% NRR. This is 20 points worse than either B2B or B2C SaaS. But AI products that sell or >$250 per month see 70% GRR and 85% NRR. This is essentially the same as B2B SaaS. The downside of being easy to buy is being easy to cancel. (Another factor is that AI products often push people to buy while they're still in "trial" mode w/ tactics like 7-day trials & daily usage limits.) As much as I <3 PLG, the winners are quickly moving from PLG > Enterprise. --- See the full report in today's Growth Unhinged newsletter: https://lnkd.in/ewnsj8MD Hope this brings some objective data to the conversation. #ai #churn #retention #customersuccess

Another banger of an article by Kyle Poyar with empirical data (via ChartMogul) on GRR and NRR trends for B2B SaaS vs. B2C SaaS vs. AI-Native products. Shows a more rational view of the whole AI-is-killing-SaaS mania that swept the punditry this past year. (One of the highest compliments we've received for our latest "Martech for 2026" report is that it too presents a BALANCED view of what's actually happening with SaaS and AI — in our report, it's more about architecture and functionality than vendor economics.) Not to say that SaaS has a free pass. Particularly in #martech, where the crazy landscape of 15,000+ SaaS solutions is objectively too bloated in many legacy categories. I believe the top quartile will evolve and thrive, but the bottom half will fade. The data Kyle shared on median GRR and NRR here paints that picture starkly. But the top quartile that do evolve — and many are evolving very rapidly with AI, while leveraging their existing foundational strengths upon which so much of real-world marketing operations is built — will be more formidable competitors to many of the AI-native startups in their space. Definitely going to be an interesting year in 2026... Here's Kyle's latest post: https://lnkd.in/eQPggRp3

What's new in AI pricing: giving up on the idea of value-based pricing? PostHog just announced new AI pricing that's disruptive in a way that might upset pricing experts. They're not trying to tie pricing to outcomes. They're not charging based on seats. The model is simple: it's a pass-through of AI costs w/ a 20% markup. And every customers gets $20 of free usage to try it out. What's to like about this model: (a) it gives everyone access to AI features & lets them choose how they want to use them, (b) it feels fair -- it's pretty hard to argue w/ or negotiate against the model, and (c) it positions PostHog's value as *elsewhere* -- they have 10 other products they can monetize & AI can be used across all of them. I see this as a step toward a potential future direction* that might look something like this: 1. Customers choose between self-managed/BYO LLM and managed AI. Many folks are probably sitting on millions of $$ of unused AI tokens from OpenAI and Anthropic. Meanwhile they're getting overage bills from AI vendors in coding & other categories. Enterprises would (IMO) love to easily plug their unused credits into their favorite 3rd party tools. It would mean fewer AI vendors to manage, fewer tokens to predict & way less waste. The margins on managed AI would likely come down substantially (similar to PostHog's 20% markup) since this would be "competing" with BYO. 2. Vendors would monetize elsewhere, likely in more predictable subscription/SaaS models (although PostHog's products are all usage-based). This fits into how enterprise buyers traditionally like to pay for technology. And it forces vendors to clarify their value-add above & beyond the AI tokens (no room for AI wrappers). 3. This would open the door for 3rd party products to be sold *through* OpenAI & Anthropic, similar to the AWS marketplace model. If 3rd party products help sell more OpenAI tokens, why wouldn't you give OpenAI sellers quota relief on these products? And why wouldn't Google & Anthropic follow suit? Anecdotally, selling via the AWS marketplace can result in 60%+ faster procurement cycles. This could be win-win for the LLMs & the app vendors. --- What's exciting about this direction is that it would *simplify* buying, accelerate AI adoption & align incentives across the ecosystem. Although it does feel like we'd be going back-to-the-future. *I don't see this model applying to every product. It'll likely apply best to developer-focused tools and/or bigger platforms w/ a lot of AI usage. The folks who would be most insulated would be "service-as-a-software" players who deliver tangible work products/outcomes. #ai #aimonetization #pricing

It’s been 60 days since I left VC to become a full-time solopreneur. It feels like it’s been 6 months. An update on how it’s going: For context, I write a newsletter about startup growth and monetization called Growth Unhinged. I also work with tech companies 1:1 as an advisor and consultant. 1. If I had to describe the transition in one word it would be *overstimulated*. I see so much opportunity yet everything falls to me. Getting summoned to three weeks of federal jury duty in January certainly doesn’t help (& might’ve caused a mild panic attack). I’m resolving to make 2026 the year of focus and, hopefully, some delegating. 2. The newsletter became a Business Bestseller on Substack (#47), only two months after introducing the paid tier! It’s now doing >$75k in annualized revenue from subscribers! (Ads are a bigger revenue source, although I’d like the two to be equal by the end of next year.) 3. I’m working on more premium offers to make paid subscriptions a no brainer without alienating free readers. Only 0.6% of readers are on the paid tier… not great compared to tech companies. What’s on my mind for 2026: subscriber-only perks on top GTM tools (coming in January 👀), special in-depth reports, and ways to bring the community together. I’m very open to ideas! 4. My best decision yet has been to team up with CJ and Ben to launch the Mostly Growth podcast. It’s a great excuse to have water cooler conversations with “coworkers”. And we recently started bringing in guests like Brian Balfour, Emily Kramer, Leah Tharin, Dave Kellogg & Roland Ligtenberg. Somehow we’re already 20 (!) episodes in, which I hear is better than 90% of podcasts. Hope folks like it as much as I like recording it! — If you made it this far, consider following along by subscribing to Growth Unhinged: https://lnkd.in/exTbjKaM #solopreneur #startups #writing

Just wrote an article with Kyle Poyar. We hit $1M ARR 3 months after launching and broke down our strategy to double it. I'm covering the 4 main pillars of our outbound playbook: → Cold Calling → Email Campaigns → LinkedIn Campaigns → LinkedIn Manual Prospecting ... plus a complete breakdown of our 5 favorite plays. Growth Unhinged has been my favorite GTM newsletter for some time now, so it's an honor to be featured there. Thanks Kyle for having me!

A modern outbound playbook for 2025: Many predicted the demise of outbound. Well, it's alive & kicking. GTM leaders told me that intent-based outbound is the no. 2 channel where they plan to invest more in 2026, just behind AI discovery. Here's an extremely tactical guide to what's working (and AI workflows you can steal) via Fivos Aresti, co-founder of AI-first growth studio Workflows.io. 👉 Manual cold outreach for ICP target accounts Signals are great, but most target accounts aren't in market (yet). Start by mapping out the TAM & then filter down to a list of target accounts. Fivos found 66,000+ companies, then narrowed it to 5,700 ICP targets. These were split into four tiers: - Dream 150 accounts (highly manual) - Tier 1 (cold calling & more personalization) - Tier 2 (automation via email & LinkedIn) - Tier 3 (automation via email only) 🛠️ Key tools: BuiltWith (identify HubSpot users), GetLatka (identify SaaS companies), Exa (crawl websites w/ AI), Crunchbase (identify funding), Apollo (identify firmographic info), Clay (research & qualification). 👉 Automated, signal-based plays for other ICP accounts These perform 2x-10x better than cold outreach plays. The five signal-based plays they prioritized: - Customer alumni play - Website visitor de-anonymization play - LinkedIn connections of founders play - LinkedIn engagement play - Social listening play LinkedIn connections of founders is performing best. Fivos sees a 25.4% (!) reply rate, although much of this is harvesting demand from 1.5+ years of posting content. --- Get the full playbook including signal-based workflows you can steal in today's Growth Unhinged newsletter: https://lnkd.in/eSsYvxJZ Hope you find it useful 🙏 --- Today's newsletter is supported by Metronome, who's organizing an insanely valuable webinar on AI pricing with THE Elena Verna (link in comments). #outbound #aigtm #gtmengineering

Seemingly everyone is replacing point-and-click with a prompt bar. Design no longer differentiates. We're going to need to differentiate on the *agent* experience instead. One 🔥 example from Replit: The product helps you improve your prompt. Most of us (myself included) aren't prompt engineers. We're riddled with prompt anxiety. And a garbage prompt can often mean a garbage result -- and not much patience to keep going. Small things Replit does that make a big difference: 1. They have a Mad Libs-style prompt tutorial directly in the bar. You don't need to remember the best possible prompt. You just replace the text with what you're trying to build: type of project, who it's for, what it does, and features, style & detail. 2. They provide suggestions, likely pulled from data around Replit's most successful apps. These includes apps, data tools, automations, games, business tools, and websites. I found it far more helpful than a blank box telling me to "build anything". 3. They let you choose your own Agent Experience. Specific areas you can customize: - Speed (choose "Fast" to make lightweight changes quickly) - Level of autonomy (low, medium, high or max) - App testing - High power power - Agent access to web search & media generation These options make Replit usable for coding n00bs (like me) up to pros and for use cases ranging from prototypes (build it fast) to production grade apps (take extra time to make it work). --- The best AI products help us clarify our ideas. And they shorten the distance between curiosity and real value. Yaakov Carno tried out 40 AI products with a prompt bar UX and reported back on how to do it well. See the full piece in Growth Unhinged here: https://lnkd.in/ecJ8TRFJ #ai #ux #genai

If you're searching for GTM tech this Black Friday, start with these: Maja and I asked our communities about their favorite B2B GTM tools in 2025 including which were "most impactful" and which are "must try". 195 software GTM leaders weighed in. (And Gemini sorted through 1,000+ open-ended responses 🤯.) ChatGPT topped the list of most impactful tools with 50 write-ins. They were followed by OG HubSpot (40 write-ins) and breakout unicorn Clay (30 write-ins). This list included tried-and-true tools you're probably familiar with like LinkedIn, Gong, Notion, 6sense, Canva, Figma, Apollo & ZoomInfo. Things got more interesting on the "must try" list. GTM leaders are trying out vibecoding (Lovable - 2nd, Replit - 5th, Magic Patterns) and AI agents (n8n - 3rd, Relay.app, Dust). Folks are also on the lookout for more & better tooling for advanced GTM plays (see: Clay, Keyplay, Unify). If the "must try" list is any indication, we're eager to graduate beyond ChatGPT to more advanced AI use cases. But 2025 is still the year of AI experimentation vs. true impact. Read the full 2025 State of B2B GTM report here: https://lnkd.in/ea_zYUzB --- PS: I'm working on a new Growth Unhinged perks program. Let me know your favorite GTM tools and I'll try to help you save some $$! 🙏 #gtm #ai #b2b

I thought there'd be consensus on how to monetize AI by now. I was (mostly) wrong. There's outcome-based pricing (see: Intercom, Crescendo, Chargeflow), credit pricing (see: Salesforce, Lovable), work-based pricing (see: Imagen, EvenUp), hybrid pricing (see: Monday, Clay), and lots more. The only thing people seem to agree on is that AI credits are bad. But they're also a lifeline to protect against exploding costs. The top 10% of your users might account for 70-80% of your AI costs. (Heck, even MoviePass is selling credits these days.) I'll be unpacking how to (profitably) add AI to your product with Aakash Gupta from Product Growth and Axel Sooriah from Atlassian Jira Product Discovery. On the agenda: why AI products fail to make money, the connection between pricing models & profits, the evolving unit economics of AI, and how product teams need to work differently. Join us for a LinkedIn Live on 12/4. Register here: https://lnkd.in/ebgT38D4 Hope to see you there! #ai #ad #atlassianpartner

We're in a new UX era for software: the era of the prompt bar. That doesn't mean you can just add a prompt bar & move on. What 40+ of the best AI-native companies are doing: I enlisted Yaakov Carno to map out AI prompt journeys across products like Canva, Notion, Vercel, Lovable, Airtable (& lesser known AI companies). He traced every step from homepage to first value. The 7 most interesting observations ⤵️ 1. Consider offering multiple entry points: the prompt bar OR a visual UX. Gamma and Wix are two 🔥 examples. Gamma even has 3 options. 2. Increase relevance with a "hidden" step between the prompt bar & output. Lovable, for example, has a series of onboarding questions to collect more context. 3. Don’t just tell users they can “generate anything.” Show them specific, meaningful examples that connect to their job-to-be-done. Canva, for instance, has a visual use case selector w/ a category & sub-category. 4. Use the prompt bar to educate users about what kind of tasks your product supports. Zapier there are sample prompts for lead management, customer support, marketing & growth, and project mgmt. 5. Distinguish whether the AI is a copilot or autopilot. Notion nails this with a mix of placeholder text, surrounding CTAs, and examples that clearly convey what you can do and why. 6. Integrate context, data, and sources at the beginning. bolt.new has a new minimalistic design that focuses only on importing from Figma and GitHub. Constraints increase the likelihood of relevant, valuable outputs. 7. Add a fun factor to make the experience feel different & maybe a bit magical. Playfulness can lower hesitation. Riff and Leap use “Shuffle” or “Surprise me” icons to encourage exploration. --- Read the full deep dive in today's Growth Unhinged newsletter here: https://lnkd.in/ecJ8TRFJ Then chime in: Do you have a favorite AI prompt journey? And what new tools are folks using to build & optimize these new UX experiences? Hope you find this useful 🙏 #ux #ai #genai #product